-1~~m -
DIVERSIFIED
0                                     SECONDARY
F                                     EDUCATION
and
DEVELOPMENT
Evidence fromii Colombia
an1df Tanl zanlitl
George Psacharopoulos
William Loxicy
FI LE  Cu PY
Report  .  1 jiT   I:E
TI l    t VIE jl  I 1E L) DFC(NL(.EY F)l2 AT 1(11
Author ;3A( j.-iAP 'I/R  .-<Ii> LO
Ext.     C1 Roo-Jli:  Do:flpt.
OLD PUDLICATIOIN T   c  1c1t5
JC,HNS
9 9          1 [opUNi_



Diversified Secondary Education
and Development
A World Bank Publication



Diversified Secondary Education
and Development
Evidence from Colombia and Tanzania
George Psacharopoulos and William Loxley
PUBLISHED FOR THE WORLD BANK
The Johns Hopkins University Press
BALTIMORE AND LONDON



Copyright © 1985 by The International Bank
for Reconstruction and Development/The World Bank
1818 H Street, N.W., Washington, D.C. 20433, U.S.A.
All rights reserved
Manufactured in the United States of America
First printing December 1985
The Johns Hopkins University Press
Baltimore, Maryland 21211, U.S.A.
The World Bank does not accept responsibility for the views expressed herein,
which are those of the authors and should not be attributed to the World Bank or
to its affiliated organizations. The findings, interpretations, and conclusions are
the results of research supported by the Bank; they do not necessarily represent
official policy of the Bank. The designations employed, the presentation of mate-
rial, and any maps used in this document are solely for the convenience of the
reader and do not imply the expression of any opinion whatsoever on the part of
the World Bank or its affiliates concerning the legal status of any country, territory,
city, area, or of its authorities, or concerning the delimitation of its boundaries or
national affiliation.
Library of Congress Cataloging-in-Publication Data
Psacharopoulos, George.
Diversified secondary education and development.
Includes bibliographical references and index.
1. Education, Secondary-Colombia-Curricula-Case
studies. 2. Education, Secondary-Tanzania-Curricula-
Case studies. 3. Vocational education-Colombia-Case
studies. 4. Vocational education-Tanzania-Case studies.
1. Loxley, William A.   II. Title.
LB1629.5.C7P73  1985    373.19'09861       85-45103
ISBN 0-8018-3119-9



Contents
Preface                                                      vii
Acknowledgments                                              ix
1. Introduction                                               3
Assumptions Underlying Diversification 3
Scope of the Study 5
2. Historical Background                                      8
The Diversification Debate 8
Vocational Education Worldwide 11
Types of Diversified Education 16
Diversification in Colombia and Tanzania 17
3. Previous Evaluations                                      22
Review of the Literature 23
Vocational Education and Economic Development 31
World Bank Projects 33
4. Research Design and Methodology                           36
Research Design 37
Main Variables 44
Test Reliability 50
List of Variables S1
5. The Colombia Case Study                                   53
Background of the Study 54
Socioeconomic Profile of 1981 Cohort 63
Internal Efficiency of Schools 65
Comparing Costs with Outcomes  85
The Class of 1981 One Year Later 92
Explaining the Decision to Work or Study 105
The Class of 1978 Three Years Later 111
6. The Tanzania Case Study                                  138
Background of the Study 138
Socioeconomic Profile of Form IV Students 145
Internal Efficiency of Schools 154
v



Vi                                                    CONTENTS
Comparing Costs with Outcomes 173
The Class of 1981 One Year Later  179
Predicting Post-Form V Destinations   194
Conclusion 205
7. Summary and Conclusions                                   210
Internal Efficiency: Colombia 210
Internal Efficiency: Tanzania 214
External Efficiency 215
Concluding Remarks 225
References                                                   229
Index                                                        239



Preface
DURING THE PAST DECADE many developing countries started to "diver-
sify" their curricula at the secondary school level by introducing prevoca-
tional subjects in addition to traditional academic courses. The most
frequently stated reason for this change was that prevocational subjects
would widen students' choices of future careers and thus make the school
system more relevant to the world of work.
Given the popularity of such educational innovation, it is surprising that
no comprehensive studies or empirical evaluations have been made of its
impact on national economic development. Policymaking has been based
on an intuitive assumption that the inclusion of prevocational subjects in
the secondary education curriculum offers a more solid basis for career
choices than does traditional academic schooling alone.
The World Bank, which has lent extensively for curriculum diversifica-
tion, therefore undertook an evaluation of secondary school systems in
Colombia and Tanzania that had been diversified with Bank assistance.
This book reports the results of that study. A random sample of over
10,000 students following either diversified or conventional (academic or
vocational) curricula was taken just before graduation in order to obtain
information on their socioeconomic background and educational charac-
teristics. The same students were contacted one to three years later to find
out what further education they had received and what their experience in
the labor market had been.
Four major criteria were used to compare schools offering diversified
curricula with those offering only academic or only vocational subjects.
The principal considerations were:
* Access to the different types of school by students with different
socioeconomic origins (equity)
* Cognitive learning of the students (internal efficiency)
* Labor market experience of the graduates (external efficiency)
* Cost-effectiveness (economic efficiency).
Although there were some differences between the two countries and
particular courses of study, the results can be summarized as follows.
Compared with schools offering only conventional curricula, the diver-
sified schools recruit more students from low-income backgrounds and
vii



viii                                                    PREFACE
provide better cognitive skills to their graduates. However, it costs more to
educate students in diversified schools than those in control (academic or
vocational) schools. Moreover, despite their superior cognitive skills,
graduates of diversified schools do not find jobs more easily and do not
earn more than the graduates of control schools.
The explanation for this negative result may be the short horizon of the
longitudinal tracer study. Hence, another survey of the same sample is
contemplated in future years, when the graduates have become more
settled in their careers.
AKLILU HABTE
Director
Education and Training Department
The World Bank



Acknowledgments
THIS STUDY of Diversified Secondary Education Curricula (better known
under its DiSCuS acronym) was funded by the World Bank's Research
Committee in July 1981 and conducted under the Education Department
of the Operations Policy Staff. George Psacharopoulos served as principal
investigator and project director. William Loxley was a consultant for the
two-year duration of the study. Ana-Maria Arriagada served as research
assistant to the project.
The research design of the two-country case study required a sampling
framework, questionnaires, the collection and processing of base-year
data, and a follow-up survey of the national cohorts of secondary school
graduates. To this end, ministries, research organizations, and scholars in
the host countries provided the skills and staff needed to make this project
possible.
In Colombia, Eduardo B. Velez and Carlos C. Rojas of the Instituto SER
de Investigaci6n in collaboration with the Colombian Ministry of Educa-
tion carried out the enormous task of surveying 8,000 students in public
and private secondary schools nationwide and tracing them one year later.
They also raised data retrospectively on 2,000 graduates of an older
cohort.
In Tanzania, Issa M. Omari, in collaboration with staff at the Education
Department of the University of Dar es Salaam, staff at the Institute of
Education, and officials of the Ministries of Education and Manpower and
the Examination Council carried out the complex tasks of surveying more
than 4,000 secondary school students in 1981 and following up the cohort
in 1982.-
This book has benefited from the input of a number of consultants. Paul
Hurst, Kevin Lillis, and Desmond Hogan reviewed the literature. Keith
Hinchliffe provided the unit costing of the different types of schools and
carried out the first cost-benefit calculations. Antonio Zabalza analyzed
the returns of the 1978 Colombian cohort. Philip Foster made valuable
comments on an early draft of the manuscript. Robin Horn revised several
statistical tables in the second draft. John Smyth, Michael Crossley, and
John Adams also read drafts of the manuscript and offered suggestions for
improvement. Richard Kollodge organized and verified references and
citations.
ix



x                                         ACKNOWLEDGMENTS
Several World Bank staff have been involved in this project from its
inception to conclusion. Wadi Haddad, Jack Mass, Ralph Harbison, and
Lawrence Wolff, among many others, have significantly contributed to
various stages of this research. Liliana Longo has efficiently administered
the various logistical aspects of the study.



Diversified Secondary Education
and Development






1. Introduction
SECONDARY SCHOOLS that offer prevocational courses in conjunction with
traditional ones have flourished over the past decade in many developing
countries. In fact, interest in these "diversified" schools seems to be
growing because more and more governments are seeking assistance to
establish diversified curricula. Although the World Bank has been invest-
ing heavily in these educational programs, there have not been any studies
on the impact of diversified education on economic development.
These innovative secondary schools are neither completely academi-
cally oriented nor are they simply vocational training institutions. Because
the diversified curriculum concept weds academic with some degree of
vocational education, students can develop not only vocational skills in
the field of their choice, but also traditional cognitive skills in classes
designed to prepare them for a university. The stated rationale for such
schools-and the basis for their popularity-is to permit students a wider
set of future career options than is offered by the usual technical or
academic curricula.
The belief that this hybrid form of secondary schooling is also the best
way to balance labor market needs with more equal access to education
has led many governments to invest heavily in diversified schools. The
World Bank has encouraged such innovation as a development strategy
and as a means to make the rapid expansion of secondary education
consistent with a better match between skills learned in school and those
needed in the labor market.
Assumptions Underlying Diversification
As a strategy for revising the instructional content of secondary schools,
curriculum diversification is based on a number of assumptions concern-
ing the value of conventional secondary education, the need for educa-
tional innovation, and the kinds of employment opportunities available.
In most cases, especially in developing countries, it is assumed that there
exists a fundamental mismatch between the type of education and training
offered in traditional academic schools and the skills and other character-
istics required of graduates entering the labor market.
3



4     DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Variants of this argument have been discussed by a number of re-
searchers, among them Ahmed and Coombs (1975), Dore (1975), Van
Rensburg (1974), Court (1972), Orata (1972), Lourie (1978), Benoit
(1974), Nyerere (1967), King (1977), Ruddell (1979), Weeks (1978),
Schiefelbein (1979), Figueroa, Prieto, and Gutierrez (1974), and Unger
(1980). A second assumption is that the content and methods of tradi-
tional secondary schooling unrealistically exaggerate the educational and
occupational aspirations of graduates, creating excessive demand for uni-
versity education and highly valued jobs. (See Klingelhofer 1967; Oxtoby
1977; Evans and Schimmel 1970; Silvey 1969; and Ruddell 1979.)
Diversification of secondary school curricula, it is argued, will ensure
that more students, including those with less advantaged backgrounds,
receive education and training better suited to their own social and eco-
nomic conditions and the needs of their respective nations. Because such
schooling is more closely attuned to the manpower requirements of indus-
trialization and growth in the developing world, fewer graduates will
select less-needed, middle-level jobs, and fewer will elect to continue their
education. A combination of traditional academic education and prevoca-
tional training, it is contended, will give students more accurate informa-
tion about the less prestigious jobs available and will arm them with more
realistic attitudes and aspirations toward the world of work.
Little evidence exists to support or refute these hypotheses; to date,
there has been no large-scale investigation of diversified secondary school
experiences among developing nations. Given that lack of information,
the World Bank embarked on a study of the subject in cooperation with
government authorities. Approximately 20 percent of the costs of educa-
tion projects financed by the Bank in the past twenty years can be directly
or indirectly attributed to components for diversified education. All the
Bank's secondary education projects have supported this kind of curricu-
lum. The pattern is as much a result of World Bank policy as it is of the
demand for such schools by national governments. After more than fifty
such projects had been completed, it was felt that enough time had elapsed
to warrant testing the comparative effectiveness of the diversified school
model against the more traditional types of academic or vocational
education.
An internal review of diversified education components in World Bank
projects over the past two decades (Haddad 1979) identified several
questions. The review indicated a lack of evidence that the new type of
schools improved the quality of education, changed student attitudes
toward the labor market, or had the desired effect of directing graduates to
areas of employment where they allegedly were most needed. Diversified
education projects assisted by the Bank have not contained enough tracer



INTRODUCTION                                                    S
(follow-up) studies or other types of evaluation to show whether diver-
sification has met its objectives.
Scope of the Study
The present study was intended to begin a comparison of the advantages
accruing to students and graduates of diversified schools and those with
more traditional types of formal training. Potential advantages of diver-
sification can be measured in these ways:
* Wider access to secondary schooling by less privileged socioeconomic
groups
* Cognitive attainment, as measured by test scores in both prevoca-
tional and academic subjects
* Occupational choice consistent with a student's interests
* Returns to national investments from workers who have a combina-
tion of academic and prevocational skills, rather than merely one or
the other.
Specific hypotheses about the internal and external efficiency of diver-
sified education need to be tested in order to obtain a clearer picture of its
effect on economic development. Using nationally representative longitu-
dinal samples of graduates who attended either academic or prevocational
secondary schools, the study tests the following hypotheses to measure the
relative effectiveness of diversification:
* Diversification favors the recruitment of students from lower-income
families and thus provides more equitable access to schooling.
* Diversification enhances cognitive achievement in the field of voca-
tional course work and later on the job.
•Diversification creates a desire for further training and instills occupa-
tional aspirations better suited to national economic needs.
* Diversification diminishes the private demand for postsecondary
education, especially at the university level.
- Diversification increases the motivation to seek work at the end of
secondary schooling.
i Diversification leads to employment in the field of vocational spe-
cialization at school.
* Diversification enables graduates to earn more than they would after
other forms of schooling.
* Graduates of diversified schools have a shorter period of unemploy-
ment while seeking work after graduation than do graduates of other
secondary schools.



6      DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Investment in diversification has a higher social rate of return than
that in either vocational or academic education alone.'
Colombia and Tanzania, the countries selected for the empirical inves-
tigation of diversification, met at least two criteria that would allow a
sound evaluation of these postulates. First, their programs have been in
place and well implemented for a relatively long time, and, second, diver-
sification has been introduced on a scale large enough to permit a sufficient
number of students to participate in a meaningful sample.
Secondary school diversification has proceeded furthest in Latin Amer-
ica and East Africa. Colombia was chosen to represent Latin America
because it has implemented public diversified education extensively, along
with purely vocational and academic secondary schools. Tanzania was
selected in East Africa because diversification there is a well-established
policy based on self-reliance and has spread throughout the educational
system.2
A longitudinal tracer study methodology was adopted to assess the
results of diversification. The methodological design required gathering
data on schools and individual student characteristics from a cohort about
to complete its secondary education. Data collected from students in their
final year of studies represent base-year information. The base-year data
set then was used to perform a cross-sectional analysis that compares
students in diversified, vocational, and academic schools on their attitudes
and cognitive skills.
In addition to base-year information collected during the first year of the
study, tracer surveys were taken periodically to learn about the students'
experience after leaving school-for example their employment experi-
ence and earnings-which could be used to conduct a cost-benefit analysis
of different types of secondary education curricula. Finally, drawing on
base-year data and the follow-up information, an assessment
was made of the direct and indirect effects of key policy decisions, such as
the type of curriculum followed, on what students learn and what happens
to them later in the labor market. Such analyses can indicate whether
students' cognitive achievements, attitudes, or graduate earnings differ
(and by how much) by type of curriculum. It is also possible to assess the
extent to which changes in curriculum are cost-effective.
The following chapter of this study provides a brief history of diversified
1. The social rate of return measures the economic payoff of an investment in human
resources relative to its cost. For an elaboration of this concept, see Psacharopoulos (1981b).
2. The concept of education for self-reliance stems from the general movement in Africa in
the 1960s that a country should develop on the basis of its own resources (see Nyerere 1967).
It is thought that schools should teach students to behave collectively and that the curriculum
should be geared toward practical subjects relating to work-mostly agricultural-the
graduates will be doing later in life.



INTRODUCTION                                                     7
education worldwide and examines the role of the World Bank in this
respect. Chapter 3 reviews evaluations of past World Bank projects and
discusses problems of implementation. Chapter 4 describes the research
design of the two case studies in Colombia and Tanzania and discusses
sampling and survey instruments.
Chapters 5 and 6 present the results of the case studies of the two
countries. Beginning with descriptions of their national education systems,
the chapters include base-year analyses of the internal efficiency of diver-
sified schools, followed by an evaluation of the tracer study and an
examination of the external efficiency of diversified education based on the
choice of further study or work, and the determinants of earnings. Also
presented is a cost-benefit analysis of the alternative curricula.
The final chapter synthesizes the results of the study by assessing the
contribution of diversification to economic and social development.



2. Historical Background
IN RESPONSE to economic, social, and political pressures in developing
countries, governments have introduced prevocational training into the
traditional academic curriculum in secondary schools. These pressures
continue to shape schooling policies and often lead to educational reform
of the type evaluated in this study. (For a review of implementing educa-
tional innovations, see Lillis and Hogan 1983; and Heyneman 1984.)
The Diversification Debate
During the 1960s, after many nations had achieved independence from
colonial rule, it was rightly thought that investment in education would
significantly improve the prospects for economic and social development.
Education systems rapidly expanded, but skyrocketing recurrent costs
forced many governments to cut back the pace of expansion. Shortages of
materials, equipment, and teachers widened the qualitative differences
among schools, both private and public, rural and urban. The rapid
expansion was blamed for the diminution of educational quality and high
wastage rates within the education system.
The number of students graduating exceeded the positions available to
them in the job market. In addition, graduates' skills often did not match
the requirements for the jobs that were available. Clerical and sales
positions were filled by academic school graduates, but technical jobs went
begging. Traditional schools had not provided the necessary vocational
training, while less formal, on-the-job apprenticeship programs were too
poorly organized to meet the rapidly increasing demand for special skills.
In the midst of rising unemployment among both educated and unedu-
cated, the governments of developing nations found it difficult to meet the
need to expand education with the limited resources available. Many
parents hoped that education would enable their children to escape pov-
erty and better their lives through instruction in skills that would help
them enter employment in the wage sector of the economy. Parents
believed-and rightly so-that the chances of at least one child success-
fully completing schooling would rise with the numbers of children they
had who attended school.
8



HISTORICAL BACKGROUND                                           9
Given these considerations, there have been at least two major argu-
ments for and against the introduction of a vocational emphasis in the
traditional academic secondary curriculum. One is economic and the
other is political.
Economic Efficiency
The economic argument for diversification is focused on a perceived
need to orient the formal education system to the world of work. (See
Cliffe 1973; Diyasena 1976; Figueroa, Prieto, and Gutierrez 1974; Lema
1978; Price 1973; Sato 1974; Tchen 1977; and Vulliamy 1980.) Itis based
on the assumption that traditional education, apprenticeship, and on-the-
job training cannot train enough workers to meet current and future
demands for skilled labor. Those arguing for an emphasis on vocational
courses in secondary education cite evidence that a system of academically
oriented education predisposes students to enter white-collar jobs and not
ones that require manual labor and skills in short supply. It is often argued
that schools should equip students with the knowledge, skills, and atti-
tudes they need to enter specific vocational fields, such as commerce,
agriculture, and technical trades, whether as employees or self-employed.
Proponents of vocational education agree that if a formal education
system is to be viable for developing nations in the 1980s and beyond, then
more vocational schooling will be necessary.
Some opponents of diversification may accept this diagnosis, but they
emphasize the costliness of the strategy and its inability to attract compe-
tent teachers and suitable equipment. They also warn that the overall
quality of education may decline, so that graduates will be proficient in
neither academic knowledge nor specialized vocational skills; as a result,
they will be less able to acquire skills on the job (Foster 1965; Blaug 1979;
Grubb 1979). At best, it is contended, students will acquire only a vague
general introduction to vocational skills. Finally, there is a fear that a
diversified education system may give rise to implicit and hidden tracking,
locking students into career paths that convey lower status and create
invidious distinctions between social groups. Opponents of diversified
education also argue that curriculum reform is not the most efficient way
to implement education for self-reliance. Of greater importance, especially
in low-quality school systems, is upgrading the caliber of the teachers.
Another factor to be considered is the type of skills to be taught in the
classroom. Usually, a hierarchy of skills can be agreed upon: academic
education prepares graduates for numerous careers that require verbal and
quantitative skills; professional or vocational education prepares the
graduate for a more specific type of job in a single field; work-based
training generally prepares the individual for only a single job. Conse-



10    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
quently, a loosely defined lexicon of formal cognitive skills would include
academic, prevocational, and vocational education, as well as appren-
ticeship training and work experience (Zymelman 1976).1 A number of
experts, however, believe that schools should not train students only for
specifically defined jobs. The most schools should be expected to do,
according to this view, is to provide a broad technical foundation that will
be refined by further off- or on-the-job training. In short, so this argument
goes, the problem of reconciling any mismatch between the world of work
and the world of education cannot be solved merely by reforming the
curriculum (Foster 1965; Blaug 1979; and Grubb 1979).
Sociopolitical Considerations
A second major argument for diversifying secondary education con-
cerns considerations of equity. There are those who believe that an
academically oriented system in a developing country will produce gross
long-term inequalities between a small elite of educated and their unedu-
cated counterparts.
The vast majority of citizens in developing nations are rural and unedu-
cated. If governments wish to retain the support of their principal constit-
uents, they must find means to satisfy the social demand for "free"
education. They may do so by instituting a traditional education curricu-
lum, or programs outside the formal education system, or schooling of
such short duration that many students are given no incentive to go on to
universities. Some governments have attempted to check the growth of
academic enrollments by offering combined academic-vocational pro-
grams along with solely academic ones. But unless the two curricula are
given equal recognition and support, this strategy is doomed to failure
once parents and students realize the implied low status of and low
economic returns to the vocational programs.
Opponents of secondary diversified education programs complain that
because vocational secondary schooling is often more costly than
academic secondary programs (because of the need for specialized equip-
ment), introducing such a curriculum is a less cost-effective way of elimi-
nating enrollment discrepancies between rich and poor and urban and
rural students than is expanding access to traditional academic schooling.
Historically, formal schools in industrial nations have served social class
interests. Because primary and secondary schooling is a prerequisite for
1. Lillis andHogan (1983, p. 92) claim thattechnological developmentin mostruralareas
is still at a level where skill development and preparation for employment occur naturally on
the job rather than in school.



HISTORICAL BACKGROUND                                        11
university enrollment, the middle and upper classes have wanted schools
to have a highly academic content. For working-class students, however,
primary and secondary schooling has been viewed as the end of their
educational careers; hence, the curriculum for these students has usually
been oriented to trade and vocational subjects and rarely to the humanities
and sciences. As a result, dual secondary education systems have evolved
that restrict the choice of occupation for working-class students (William-
son 1979).
Generally, however, ethnic and social cleavages in developing nations
have not been reinforced by the types of curricula offered to different
groups. Far greater inequities have been perpetrated by the scarcity of
school places than by explicit educational policies created "to provide
different structures and types of schooling for distinct social strata" (Fos-
ter 1983, pp. 1-4).
Vocational Education Worldwide
A strong positive association has been well extablished between a
country's level of economic development and the percentage of secondary
students in vocational education. Yet over the past thirty years there has
been a decline in the share of resources devoted to vocational education
within almost all national systems. Whereas in 1950, 25 percent of all
secondary students were in vocational programs, by 1975 the percentage
had fallen to 17. That decline (documented by Benavot 1983) indicates a
trend toward less differentiated and more comprehensive forms of second-
ary schooling of which diversified education is only a part.
Africa
Secondary education played an important role in Africa during the
colonial period. The emphasis in educational policy had vacillated be-
tween training for specific skills and providing merely an overall voca-
tional orientation. Just before independence there was a shift toward a
more academic orientation, in response to the need to educate a competent
cadre of civil servants, and away from a system of mass education designed
to assist in the process of democratization (Lillis and Hogan 1983). At the
time most countries were acquiring independence, it was already widely
believed that the best preparation for work was to learn broad categories
of skills that could be applied in a variety of work situations; that is,
vocational and technical knowledge should prepare graduates for addi-
tional training on the job, while theoretical knowledge needed for more



12    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
highly skilled or professional jobs is best learned in the classroom (Clignet
and Foster 1966; Dodd 1969; Jolly and others 1973; Moris 1972; Omari
1984).
After independence, most African nations shifted from an emphasis on
academic education toward programs that were intended to dampen the
rising expectations of students for higher education and employment in
the prestigious public sector. The dilemma in Africa over the past twenty
years has essentially been to reconcile a growing social demand for all
kinds of education, but especially academic education, with what were
thought to be the demands of the labor market.
In Kenya, as elsewhere in Africa, the fine-tuning required to match
vocational curricula with the actual needs of the labor market was never
adequately achieved. Colonial trade schools, still in existence today, have
found their graduates to be increasingly unacceptable to employers in the
majority of Kenyan industry. Kenyan firms have preferred to recruit from
lower down the educational ladder (King 1977), because younger and less
educated employees could be paid lower wages while learning the neces-
sary skills on the job. Furthermore, by continuously requiring additional
theoretical knowledge, these schools narrowed the marketability of their
graduates' skills. By contrast, Court (1972) and Kipkorir (1975) have
noted the success of village polytechnical schools in Kenya that train
young people for self-employment.
Efforts to diversify secondary education in the region stem largely from
the nature of African economies. In these largely rural societies, farming,
industrial, and commercial sectors are not sufficiently developed to utilize
the advanced technology generally taught in schools; the modern sector
(both public and private) is less and less able to create jobs; and there are
weak links between school and work. As a result, prevocational courses
have been added to traditional academic programs in an attempt to instill
in students the concept of self-reliance.
Since many African nations have a large public sector and a strong
tradition of manpower planning, they have opted for diversified education
programs that dilute traditional academic course work with vocational
subjects. These countries include Lesotho, Botswana, Ethiopia, Somalia,
Uganda, Cameroon, Gabon, Nigeria, Swaziland, Tunisia, Morocco, and
Tanzania.
Latin America
Unlike the situation in many African countries, formal education in
Latin America is closely tied to market economies with large private and
public industrial sectors. These conditions have made it easier to maintain
a strong private academic system with links to universities while support-



HISTORICAL BACKGROUND                                         13
ing vocational schools that feed directly into the job market. When gov-
ernments needed to drastically expand the secondary education system,
they did so by encouraging the formation of multi-track schools that
would offer both vocational and academic programs under one roof.
Assisted by World Bank education projects, diversification has pro-
ceeded far in Colombia, the Dominican Republic, El Salvador, Jamaica,
Guyana, Trinidad and Tobago, Brazil, and Peru. Similar in organization to
the European comprehensive school, diversified education in Latin Amer-
ica has attempted to eradicate the highly rigid organizational structure of
separate secondary vocational and academic schools that had existed. By
introducing comprehensive schools, students were free to combine
academic and vocational programs without foreclosing future training or
career options.
National apprenticeship programs have been most characteristic of
vocational training in Latin America. They include SENAI in Brazil, SENA in
Colombia, IICA in Costa Rica, SENATI in Peru, and INCE in Venezuela.2
These programs graduate small numbers of well-trained students who
have worked in close association with employers who have paid for the
training through a payroll tax deduction (Castro 1976). However, such
programs traditionally have been outside the mainstream of the formal
secondary school vocational programs prevalent in Latin America.
Asia and the Pacific
Diversification of the secondary school curriculum has also occurred in
Asia and the Pacific. In Papua New Guinea a much publicized Secondary
Schools Community Extension Project (SSCEP) was launched in 1978 in
five provincial high schools. It planned to integrate academic and practical
subjects to teach students the skills needed for self-employment in their
home villages. This was to be done without allowing a decline in academic
standards as measured in national examinations required for graduating
from the tenth grade (Lillis and Hogan 1983; Vulliamy 1980; Weeks
1978). Standard syllabuses for the examination would be modified for
those attending the SSCEP course (Crossley 1984a).
A comprehensive evaluation of the project was conducted by the educa-
tional research unit of the University of Papua New Guinea throughout
the pilot phase. Its conclusion was that the SSCEP should be continued and
expanded. Consequently, in 1984 a trial dissemination phase was inaugu-
2. Servicio Nacional de Aprendizaje Industrial (SENAI), Servicio Nacional de Aprendizaje
(SENA), Instituto Interamericano de Capacitaci6n Agropecuaria (IICA), Servicio Nacional
de Adiestramiento en Trabajo Industrial (SENATI), and Instituto Nacional de Cooperaci6n
Educativa (INCE).



14    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
rated. Two new schools joined the project (with more expected during
1985), and a new university diploma course in educational studies was
introduced to improve SSCEP in-service training and professional support
services. There is a strong conviction that such diversified secondary
education will contribute to the development of more appropriate skills
and attitudes for graduates returning to village life. But whether such an
intensively supported pilot project can be successfully replicated through-
out the nation remains to be seen (Crossley 1984b).
Another diversification project has been undertaken by Thailand. With
Canadian and World Bank assistance over the past twenty years, Thailand
has moved to extensively reorganize and reshape secondary education. A
Thai-Alberta Co-operative Assessment Project Report (Thailand Ministry
of Education 1980), assessing thirty-two rural secondary schools, sug-
gested that satisfactory progress had been made toward the objectives of a
flexible diversified curriculum and adequate equipment, facilities, admin-
istration, teacher staffing, and training.
In Sri Lanka, a study by Wanasinghe (1982) reported that reform of the
secondary school curriculum attempted to orient the teaching toward
local opportunities for self-employment. Although no empirical evidence
was presented, the author argued that the reform movement demanded
too much of the available teachers and was poorly supported with training
and materials.
In Indonesia, a tracer study by Clark (1983) looked at the performance
of secondary school graduates in the labor market. Computing rates of
return for graduates of academic and vocational programs, the author
found academic tracks to be the better investment, but also that academic
and vocational graduates had a substantial wage advantage over second-
ary school dropouts and primary school graduates.
Finally, Squire (1981) reported that in Korea, diversified curriculum
programs have led to a better wedding of school skills and the require-
ments of the labor market.
United States
Interesting insights can be acquired by contrasting the experiences of
industrialized nations with those of the developing countries. In the United
States there has been a long debate over the best way to link education and
work. Industrial trade schools flourished in the early 1900s when there
was quickening demand for piecemeal machine work and when lengthy
apprenticeships could not keep pace with demand. But that philosophy
shifted during the depression era of the 1930s, when the costs of voca-
tional education became more burdensome to a government that had to
cope with substantial additional needs. Then, it was suggested that private



HISTORICAL BACKGROUND                                         15
industry pay for-such training (Davis and Lewis 1975). During and after
World War II, demand for skilled workers increased, again favoring a
policy emphasis on formal training. For the past three decades, the United
States has pursued a public school policy that has given fairly equal weight
to academic and vocational subjects (Lecht 1979; Lewin-Epstein 1981).
Apparently, the emphasis and rationale for vocational schooling in the
United States followed trends in the business cycle. In periods of peak
demand for skilled workers, when apprenticeship programs were inade-
quate to meet the needs, formal vocational training programs sponsored
by the government flourished and probably paid for themselves if for no
other reason than that they helped alleviate bottlenecks in the skilled labor
market. More recently, vocational training has been used as a means to
create jobs for groups with special needs (such as the long-term unem-
ployed or those with few marketable skills) rather than to train people for
standard existing occupations.
Western Europe
By World War II most European countries had evolved a highly differ-
entiated dual system of education based on social class (Foster 1983;
Husen 1975; Boudon 1974; Lauglo 1983; Bellaby 1977; Ford 1969).
Because of the strong class differentiation, students were subjected to a
dual tracking system that destined them either to vocational or to
academic pursuits. With postwar economic recovery, demand for a wide
variety of skills created a need for a less rigidly differentiated educational
system. By making secondary schools more comprehensive, students of all
social classes could enroll in the curriculum of their choice if they had the
aptitude, motivation, and, sometimes, the necessary financial means.
Europe's comprehensive schools have served as a model for mixing curric-
ulum programs under one roof, although some countries outside Europe
have continued to impose rigid barriers between academic and vocational
schooling (Husen 1975; Williamson 1979; McPherson, Gray, and Raffe
1983; Eggleston 1983).
Socialist Countries
Much as in Western Europe, the socialist nations responded to the needs
of their industrial economies for a wide range of vocational and academic
training with schooling that emphasized highly technical skills. As a result,
Eastern bloc nations focused their efforts on vocational-technical educa-
tion at the expense of a purely academic curriculum. Because the U.S.S.R.
and its East European allies had repudiated the Western traditions relating
social class to employment, young people were seldom stigmatized by



16    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
following technical and trade, rather than academic, curricula. Many
socialist governments in developing areas have followed the Soviet exam-
ple and have emphasized a vocational orientation even more strongly in
order to meet manpower requirements.
Socialist regimes have often turned to vocational education as a way of
eradicating the differences between mental and physical labor-and hence
class differences-that formal schooling was thought to have created.
Perhaps China and Cuba have carried this distinction further than other
nations (see Dumont 1974; Figueroa, Prieto, and Gutierrez 1974; Gray
1976; Leiner 1975; Lofstedt 1981; Price 1973; Roca 1977; Carnoy and
Wharteim 1976, 1977). While work-study programs in those countries
were intended to narrow the status differentials between occupations, the
students' subsequent economic contributions have helped to offset the cost
of their education, thereby allowing the educational system to accommo-
date greater numbers than would otherwise have been possible. Such
financing of recurrent expenditures by students is rare in primary educa-
tion systems.
In the U.S.S.R., the "factory schools" started in the 1920s have been a
strong feature of Soviet education. Over the past forty years the U.S.S.R.
has pursued a diversified education policy, with special emphasis on
technical-vocational education (Blumenthal and Benson 1978). The
reasons for the approach lie in the industrial needs of Russia's economic,
scientific, and technological development plans since the revolution. The
emphasis on science and technology and the consequent upgrading of
vocational education have helped break the historical link between general
academic school and higher education. Given the nature of Soviet society,
social and occupational mobility-and hence class differentiation-are
highly dependent on schooling. One result has been that students' expecta-
tions have become based on their individual educational records rather
than on the specific courses they take.
The example of vocational education in the U.S.S.R. and China has
influenced not only East European but also many developing countries,
among them Ethiopia, Tanzania, Angola, Mozambique, Guinea Bissau,
and Cuba.
Types of Diversified Education
For our purposes, diversified education refers to secondary schools that
provide a wide range of courses which are typically found either in
all-academic or in all-vocational schools. Furthermore, within the diver-
sified curriculum, many areas of study such as agriculture, commerce,
technical subjects, and home economics are offered under one roof. Stu-



HISTORICAL BACKGROUND                                       17
dents in such schools are encouraged to specialize in one of many voca-
tional subjects and at the same time to continue academic courses. By
encouraging flexibility, diversified schools allow students the options of
pursuing either advanced technical training or continuing with university
education.
The reasons for following a diversified education program have been
varied, but the structure of educational institutions offering vocational
curricula has remained fairly standard. Vocational and academic subjects
have been taught in separate schools, or they have been placed together in
a "comprehensive" school, with only little crossover permitted.
There are two common organizational structures for diversified second-
ary education. The first is a multitrack comprehensive school in which
several vocational fields coexist with an academic course of study. This
structure allows students to mix academic and vocational subjects,
although students usually take vocational course work in only one field.
Having several vocational options under one roof allows a large number
of vocational students to gain exposure to academic courses as well. The
main advantage of such a system is that it permits students to be exposed
to academic courses while simultaneously pursuing a rigorous vocational
program.
A less common diversified system is one that combines only one major
vocational course (rather than two or more) with a high-quality academic
curriculum, again under one roof. These double-stream schools were
created by modifying the curriculum in previously all-academic schools.
As in comprehensive schools, students pursue vocational course work in
conjunction with academic subjects. Frequently, such schools require all
their students to take some vocational courses.
Another way of looking at diversified education is to concentrate on the
proportion of students' time devoted to academic or vocational subjects
rather than on the type of institution in which the instruction is offered.
Figure 2-1 portrays a time continuum, showing diversified schools as a
midway response to the earlier segregated single-track academic or voca-
tional institutions.
Diversification in Colombia and Tanzania
Colombia and Tanzania have experimented extensively with diversified
secondary education over the past decade as part of an attempt to match
schooling more closely with employment prospects.
Colombia's diversified secondary schools, called INEM (Institutos
Nacionales de Educaci6n Media), have a six-year program that combines
academic with prevocational subjects. In addition to these schools, there



18    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
remain private and public secondary schools offering an academic curricu-
lum. It was hoped, however, that the diversified schools would attract
many students away from inferior private academic schools and allow
them to pursue vocational subjects of their choice while continuing a
lighter load of academic course work. Colombia also has vocational
schools confined to subject areas such as agriculture, technical trades, and
commerce, that do not offer academic course work.
During their first two years in INEM schools, students study prevoca-
tionai subjects to familiarize themselves with career options in the various
trades. After two years of study, their work becomes more concentrated
on a particular vocational field, such as agriculture or commerce, though
students continue to take some additional academic courses. During their
final two years, the students specialize. A student who has enrolled in the
general industrial training program in the preceding two-year cycle may
either focus on metalwork, construction, or perhaps electrical mechanics,
or concentrate on an academic subject.
Graduates of INEM schools differ from those of other Colombian educa-
tional institutions. Because they have pursued both academic and voca-
tional curricula, their range of future educational and occupational oppor-
tunities is presumed to be wider. Students who choose not to continue
schooling after graduation still have obtained specialized skills and tech-
nical knowledge that allow them greater employment options. Nonethe-
less, it remains to be see what long-term advantages accrue to INEM
graduates compared with those who have either conventional vocational
or academic schooling in terms of skill in specific trades, academic knowl-
edge, and capacity to benefit from further training or to find and succeed at
jobs.
Tanzania's experience offers an interesting comparison with that of
Colombia. Tanzania's economic system stresses public ownership and
control of most major enterprises. The educational system has been de-
signed to provide the trained manpower needed for the public sector to run
Figure 2-1. Class Time Intensity Continuum by Type of Education
Single track       Multitrack         Single track        On-the-job
academic          diversified         vocational          training
(Lycees,       (Comprehensive       (Vocational          (Shops,
gymnasia)          schools)          institutions)       factories)
0%                                                      100%
Class time spent on vocational subjects



HISTORICAL BACKGROUND                                         19
efficiently. In contrast to Colombia, Tanzania has not introduced diver-
sification merely as a means to match middle-level skills with manpower
requirements. Rather, the main impetus for Tanzania's diversification
efforts stems from a strong commitment to the ideals of work education
similar to that found in the Chinese, Cuban, and Soviet vocational educa-
tional systems.
Because of Tanzania's philosophy of self-reliance and self-sufficiency in
skilled manpower, students are required to gain experience in practical
subjects in addition to academic pursuits by majoring in a vocational
subject of their choice while in secondary school. Very few students
continue their formal education after Form IV. (Secondary education in
Tanzania consists of a four-year cycle-Forms I to IV-followed by a
two-year cycle-Forms V and VI.) Combining practical work experience
with formal academic training in the first four years of secondary school is
intended to orient all secondary school students toward careers in such
vocational subjects as commerce, agriculture, domestic sciences, and the
technical fields. Those who are selected to continue their studies in upper
secondary school and eventually at a university will henceforth be exposed
to a purely academic curriculum, yet they will have acquired some voca-
tional training in the lower secondary cycle.
During their four years in lower secondary school, students pursue one
of four general vocational courses-in commerce, agriculture, or technical
or domestic science-that supplement their regular academic course work.
The amount of time they devote to all vocational and academic subjects is
prescribed by the Ministry of Education. Anywhere from 25 to 40 percent
of course-work periods per week is spent on the vocational course during
the fourth year alone. On completion of the four-year program, students
are considered sufficiently trained in their respective vocational area to
seek on-the-job training, enter government-run training institutes, or seek
employment in their area of specialization.
Tanzania has proceeded far in its effort to institutionalize vocational
subjects along with the academic curriculum. The purpose of the diver-
sified system is not to prepare students for a university, but neither is
it intended to provide vocational skills divorced from a meaningful
academic education.
One major issue regarding the evaluation of diversified schools is
whether such schools should fully prepare students for jobs or whether
they should simply teach prevocational skills. The terms "prevocational"
and "vocational" have been used interchangeably in the literature to
describe such schools, and they have therefore been used interchangeably
in this study, although we prefer the term "prevocational." The rationale
for full vocational preparation may be more debatable than the argument
for prevocational education, however. For this reason, and because some



20    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
of the evaluation criteria used in this study pertain to labor market
considerations, we offer the following quotations to show that in the two
case countries diversification was also meant to prepare students directly
for work (emphasis added throughout).
G. Varner (1965), consultant to Unesco and the World Bank, made the
following observations about Colombia's experience:
Colombia has a large deficit in human resources ... for its economic
development ... this deficit has been caused by inadequate educational
programs . . . The number . .. of students is inadequate to satisfy ...
future needs for growth and development ... (p. 53). Curricula should
differ for different students according to . . . their vocational objectives
(p. 58). Since vocational goals of students vary, it is convenient for
secondary schools to make available different types of curricular experi-
ences to prepare students for different vocational objectives .. . (p. 23).
The law instituting diversified schools is very explicit on their purpose:
Instituto de Educaci6n Media Diversificada is one which . . . offers
several academic and vocational programs (Decree 1962, no. 20, 1969,
art. 4).
... education should respond to the need for human resources required
for development of the country (Decree 1085, no. 195, June 8, 1971).
Modalidad means a set of related subjects aiming at training pupils in
aspects specific to an economic sector (Decree 1085, 1971).
Secondary education will include two cycles ... 1. Basic Cycle . .. and 2.
Vocational Cycle (Decree 080, January 22, 1974, art 3).
We believe in success of the INEM system because . . . thousands of
conventional high school graduates multiply the number of applicants
for college, unfortunately still a traditional one . . . which does not
contribute to ... development (INEM, "Principios Fundamentales de la
Educaci6n Media Diversificada en Colombia," [Medellin, April 1973]).
Similar quotations on the vocational intention of diversified schools can
be found in Tanzanian government documents:
The general aim of education in Tanzania is to . . . prepare the young
people for work . . . Secondary education is terminal and aims at
equipping the pupils with skills ... To achieve this, secondary education
is diversified and vocationalized into commercial, technical, agricultural
and home economics biases. Biasing of schools aims at (a) putting
emphasis on manual work that ties in with economic and social de-
velopment of the country; (b) preparing pupils for work in the villages
and towns. (Tanzania Ministry of National Education 1980, pp. 3, 4.)



HISTORICAL BACKGROUND                                         21
The goal of the commercial bias is to provide skills that enable the Form
IV school leavers to attain the Accountant Assistant, Grade II ... The
goal of the agricultural bias is to provide knowledge and skills equiva-
lent to the level of Agricultural Assistant . . . The goal of the technical
bias is to provide knowledge and skills equivalent to Trade Test, Grade
III. (Tanzania Ministry of National Education 1979b, pp. 63-68.)
The biases can be considered vocational or pre-vocational. In the case of
technical bias, students who do not get a place in form five, do trade test
which upon success qualify them as craftsmen. For them the bias is
vocational because they can be employed directly as craftsmen. The bias
is pre-vocational for those who pursue Full Technician course. The
Form IV examinations for those in Commercial bias are equivalent to
National Business Examinations at stage two level and they can be
employed directly. (Tanzania Ministry of National Education, letter to
the Diversified Secondary Curriculum Study research team, January 25,
1984, p. 6.)
In fact, Tanzania went one step further regarding such curricula and in
the process cleared up one of the semantic ambiguities regarding their
vocational or prevocational character. According to a 1984 report of the
Presidential Commission on Education, "Diversified secondary education
will now be called 'Basic Vocational Education and Training' and will
continue to be provided by emphasizing basic vocational skills" (Tanzania
Ministry of National Education 1984, p. 23; emphasis added).



3. Previous Evaluations
LITTLE EVIDENCE EXISTS to show unequivocally that vocational schooling
has either failed or succeeded in its intended goals. Consequently, it is
difficult to know whether the social and economic benefits of vocational
schooling in general, or diversified education in particular, exceed the
benefits accruing to academic secondary education and how these benefits
compare with the relative costs of such schools.
In evaluating the results of educational reform along vocational-
diversified lines, four broad effects of diversified education might be
examined:
* Access to schooling by various socioeconomic groups
* Degree of cognitive achievement in the subjects studied
* Differences in students' ambitions, motivations, attitudes, aspira-
tions, and other noncognitive expectations concerning occupational
careers and further education
* Post-school experiences in further training and employment of gradu-
ates.
While there is considerable speculation about the overall success or
failure of diversified education, it is more likely that educational reforms
succeed or fail in relation to the particular situation; hence, an assessment
of each national experience should be evaluated on its own merits.
This chapter reviews past attempts to evaluate vocational education
programs in the United States and European nations. Although these
countries are highly industrialized, there are many parallels with develop-
ing nations regarding philosophy, design, implementation, and historical
development of educational programs, as well as evaluation problems
encountered in understanding vocational education. The chapter also
addresses the scant evidence available from studies of developing nations
and compares the two sets of experience. There follows a brief discussion
of the relationship between vocational education and economic growth
and development. Finally, there is a review of the World Bank's experience
in financing diversified education in a number of countries.
22



PREVIOUS EVALUATIONS                                          23
Review of the Literature
Social Influences on Enrollment
Do working-class parents send their children primarily to vocational
schools, while white-collar workers more often enroll their children in
university-bound programs? That is, do the patterns of enrollment cause
or reinforce the perpetuation of social class distinctions by an educational
system that offers different curriculum programs for each group?
Loxley (1983), using data for Europe compiled by the International
Association for the Evaluation of Educational Achievement (IEA), found
that parents of students attending vocational schools differed in their level
of educational attainment from those of students attending academic
schools. Parents of vocational students tended to be engaged in manual
occupations and have less-educated backgrounds than did parents of
students attending academic programs. The secondary vocational school
population drew from families headed by predominantly blue-collar
workers and persons with less than a full secondary school education. In
Italy, for example, fathers of fourteen- and eighteen-year-old students in
the academic track had an average of one more year of schooling than
fathers of students in nonacademic courses. Those findings are similar to
profiles for students in the United States (Silberman 1978; Lecht 1979).
Vocational students are not a homogeneous group, however; some
vocational curriculum programs are likely to attract a greater percentage
of enrollees from families in which the father received more education
than the typical father of a vocational school student. For example, in
Italy, eighteen-year-old students enrolled in technical fields had a mean
parental education two years higher than students in industrial fields,
which implies that more of their fathers were foremen and craftsmen.
Similarly in the Federal Republic of Germany, fathers of fourteen-year-old
vocational students were generally in more highly skilled occupations than
fathers of students in the academic track.
In assessing differences in the social background of both fourteen- and
eighteen-year-old vocational and nonvocational students in eight indus-
trialized nations, there appears to be a typical pattern: the educational
systems are highly differentiated, separating students by social class char-
acteristics fairly early in their school careers. (For documentation of the
variability of social influences on educational access at different levels of
economic development, see Anderson 1983.)



24    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Formation of Expectations
As noted in the previous chapter, secondary education systems in West-
ern Europe have been structurally differentiated by curriculum: one track
is reserved mainly for middle-class and the other for working-class stu-
dents. Given such conditions, it is unlikely that students trained for
white-collar occupations and students trained for blue-collar employment
will perceive their future in quite the same way (Boudon 1974; Passow and
others 1976). If the dual system of secondary education were not rigid,
then all students would probably perceive their future opportunities in
much the same way.
High educational expectations among a large number of vocational
students, however, would imply that perhaps "too many" plan to go on to
college rather than terminate their formal education upon graduation
from secondary vocational school, as envisaged in educational policy. It
would also suggest that many vocational graduates would try to move out
of the trades for which they had been trained. Among nonvocational
students, high educational aspirations would suggest that these students
are likely to continue their training within their chosen field of study.
In 1971, the IEA completed a large-scale study of many European
educational systems covering vocational and academic track students
(Comber and Keeves 1973). Using these data, Loxley (1983) reported
significant differences in the educational and occupational aspirations of
fourteen- and eighteen-year-old students in a number of European coun-
tries. Vocational students trained for specific blue-collar employment
aspired to occupations of a lower status than the white-collar professional
jobs sought by secondary academic students. In sum, in the 1970s
Europe's dual system of education continued to be a powerful influence on
shaping the expectations of its students in conformity with the intention of
the curriculum.
Little is reported in the literature on developing countries about the
attitudes instilled by vocational and nonvocational courses of study.
(Evans and Schimmel 1970; Moris 1966). Pioneering work by Foster
(1965) in Ghana and later in Cameroon (Clignet and Foster 1966) showed
that all students uniformly aspired to higher education and occupations of
high status. Similar work was carried out by Olson (1974) in Kenya over a
period of several years. He, too, revealed that aspirations were consistently
and often unrealistically high, considering that so many vocational as well
as nonvocational students aspired to white-collar positions. However,
because secondary school students in East and West Africa during the
1960s and early 1970s represented a very small proportion of their age
group, they might not have been acting irrationally in expecting high



PREVIOUS EVALUATIONS                                        25
occupational rewards. In the early 1970s acute shortages of upper- and
middle-level manpower in the public sector were widespread, and stu-
dents' expectations reflected this situation.
Gouveia (1972) examined preferences for types of secondary schooling
by social class origin in Sao Paulo, Brazil. Because Sao Paulo was a
dynamic industrial setting at the time of the study, the author found no
difference in the choice of vocational subjects by either middle- or work-
ing-class youth; all showed a predilection for industrial courses. A princi-
pal cause of this was the structure of the Sao Paulo economy at the time:
the wages paid to skilled labor were higher than those paid to clerical and
sales employees.
Studies on Cuba (Bowles 1971) and Tanzania (Mbilinyi 1974a) have
indicated that ideological indoctrination dampens personal ambitions in
socialist systems. Little is known about how pursuing a vocational curricu-
lum affects the aspirations of those destined to attend such institutions in
lieu of going on to a university. But Zachariah and Hoffman (1984) report
that the introduction of manual labor into schools failed in two very
different societies: India under Gandhi and China under Mao.
Finally, there have been numerous qualitative assessments of the impact
of innovations in vocational education worldwide. Such assessments do
not offer data to support either failure or success but, rather, provide
subjective conclusions reached by experts in the field. Wanasinghe's
(1982) assessment of Sri Lanka's problems in implementing its experiment
in a compulsory prevocational education curriculum is an example. After
examining the relationship between the school curriculum and community
work prospects, teachers' receptivity to the prevocational curriculum, and
its costs, the author concluded that prevocational studies should be made
optional rather than mandatory. This assessment was based not only on
the slight prospects for successful implementation, given teachers' atti-
tudes and the program's costs, but also on the fact that vocational educa-
tion did not change student attitudes about work or further education.
Differences in Cognitive Achievement
Within the United States and Europe there are substantial differences in
vocational and academic students' achievement, with nonacademic-track
students performing less well than college preparatory students (Plowden
1967; Breneman and Nelson 1981; Husen 1975). These variations often
are attributed to differences in ability rather than in social class. The
consensus of educators is that offering either a vocational or an academic
curriculum seems to produce gains for some students at the expense of
others, but that these gains do not always occur for the same socioeco-
nomic groups. (See Rosenbaum 1976 for a review.)



26    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Few researchers have studied vocational test results because they have
been more interested in knowing how well students perform on universally
accepted standard tests of verbal and quantitative skills, rather than on
narrower, vocationally oriented subjects. Thus, while it is known that
vocational-track students perform less well in academic pursuits than do
students in college preparatory courses (Woods and Haney 1981; Wiley
and Harnischfeger 1980; McPherson, Gray, and Raffe 1983)-and this
might be expected because the vocational curriculum is less tied to those
pursuits than the academic curriculum-it is not known whether students
in vocational programs could outscore academic students on tests of
vocational content. Again, no studies have ever attempted to measure
these differences because tests of commercial or industrial content are not
given routinely to students studying mathematics, science, and the human-
ities.
Data extracted from a 1971 IEA multinational study of science achieve-
ment, made available to the Bank for analysis, offer some indication of the
pattern found in industrial countries. Although no published accounts
other than Loxley's use IEA data to compare the achievements of students
in different curriculum tracks, data collected on large national samples of
students in 1971 can be divided by curriculum placement into academic,
vocational, and comprehensive tracks. While this distinction is not univer-
sal, essentially the tripartite system is intended to distinguish between (1)
college-bound students; (2) those who are being trained in specific trades,
whether highly technical, industrial, agricultural, or commercial; and (3)
those following a "comprehensive" course that offers neither a specific
vocational nor a purely academic curriculum (that is, students are offered
fewer academic subjects than those in college preparatory schools but have
a smattering of vocational course work).
In every European country studied (England, France, Sweden, Ger-
many, Italy, Hungary, and Finland) fourteen-year-old students taking
academic courses were well ahead of vocational-track students in
academic performance. These differences in mean achievement scores are
maintained even after a number of out-of-school factors influencing
achievement (including sex, social class, and innate ability) are taken into
account. In addition, mean differences in vocabulary scores measuring
verbal facility favored academic-track students. If vocabulary scores are
taken as a measure of aptitude acquired out of school (because verbal
facility is acquired primarily in the home), the implication is that the
tracking system in industrialized nations has clearly separated students by
innate ability and socioeconomic origin at a fairly early age.
Similar results were obtained for Italy, Hungary, Sweden, France, and
the United States, using the same international achievement survey in
science but conducted for eighteen-year-olds rather than fourteen-year



PREVIOUS EVALUATIONS                                          27
olds. By the time students reach age eighteen, their number has been
reduced by the selection process. One might therefore expect to find no
difference in achievement or verbal ability in the sciences across vocational
and nonvocational tracks. Despite this greater selectivity, however, college
preparatory students in all five countries had higher scores than did
vocational students.
The conclusions reached from the IEA multicountry study of both
fourteen- and eighteen-year-old groups strongly imply that significant
differences in science and verbal achievement are based on the curriculum
program a student pursues. Vocational students in both age groups do less
well than their acadeniic school counterparts on both types of tests. It is
clear that industrial societies in the early 1970s tended to channel their less
academically able students into trade programs. Why and how vocational
programs came to include a substantial concentration of students of
below-average academic ability (as measured by tests of verbal reasoning)
in less important than the fact that cognitive differences exist across
curriculum tracks at ages fourteen and eighteen.
First-generation achievement studies have recently been undertaken in
many developing nations, but they have seldom been directed at determin-
ing differences between vocational and nonvocational education. What
little evidence exists suggests that students in a vocational course do
poorly in academic subjects, compared with students enrolled in academic
programs (Comber and Keeves 1973 [Thailand, India, Iran, and Chile];
Fuller 1976 [India]). Again, however, the tests measured achievement in
math or science, to which all students had had some exposure, rather than
achievement in vocational subjects, with which the academic track stu-
dents had little familiarity. Further, it is not known whether the lower test
scores might be the result of students' taking courses for which they are not
well suited, preselectivity based on ability grouping, or social class differ-
ences across curriculum programs.
Experiences after Schooling
Although vocational education in the United States served only about
12 percent of the students of secondary school age, recent debate has
focused on increasing its relevance as a means of reducing the imbalances
expected in the labor market. After years of falling economic productivity,
rising unemployment, and decreasing occupational mobility, the idea of
giving every high school graduate a marketable skill is appealing. Advo-
cates of vocational schooling want to provide students with such skills,
and they also want schools to simulate the work environment for future
job satisfaction and higher productivity.
Those who have opposed the vocational school concept as a viable



28    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
alternative or complement to an academic course have forcefully argued
that vocational education reshapes curriculum and teaching philosophy to
fit the socioeconomic background, ability, and aspirations of lower-class
youths, thereby placing a stigma on vocational programs, separating
manual from mental skills, and offering a false hope that vocational
education will guarantee a specific job or enhance employment mobility.
This argument is quite similar to the one mentioned in chapter 2 and is
widely cited in developing countries.
Opponents of vocational education (among them Grubb 1979) argue
that employers should be the judges of its effectiveness. Because most
employers, according to these observers, do not prefer vocational over
academic school graduates, and because most students seldom find jobs in
areas for which they were specifically trained, the economic value of
vocational education as a panacea for the ills of structural unemployment
can be challenged.
Cost-benefit studies assessing the rate of return to vocational education
in the United States have produced mixed evidence, although they seem to
point to few or no differences in economic returns that would support
vocational education. In general, private rates of return to U.S. secondary
schools over time have ranged between 14 and 20 percent. Social rates of
return have averaged between 10 and 12 percent (Weisberg 1983; Meyer
1980; Cohn 1979). Higher rates of return to secondary technical educa-
tion (relative to general) have been reported for Colombia, but lower ones
for the Philippines, Turkey, and Thailand (Psacharopoulos 1973). In
Thailand, the social rate of return to secondary technical education is
negative. Also Clark (1983) reports higher rates of return to secondary
general relative to vocational education in Indonesia (26 and 14 percent,
respectively).
Graduates of post-high school training programs would have to earn
more than graduates of high school vocational programs if the investment
in post-high school training is to be economically rational. Corazzini
(1968) contrasted a four-year high school vocational program with a
two-year post-high school vocational training program. After comparing
benefits and costs for each program, he concluded that graduates of
post-high school programs open only to high school graduates found
better-paying jobs-usually in a slightly higher technical category-than
high school graduates who went immediately to work. However, gradu-
ates of the post-high school vocational program represented a poor
investment if, once trained, they chose to compete at the same skill level in
trades open to vocational high school graduates with job experience only.
Hu and others (1971) compared private returns to both academic and
vocational education in the United States and found very high rates of
return for vocational schooling (more than 50 percent). But different



PREVIOUS EVALUATIONS                                          29
findings reported by Corazzini (1968) and Taussig (1968) indicate much
lower rates of return for vocational education (about 10 percent) and
suggest that the extra costs of such education are generally not fully
recovered by additional benefits.
Studies have also been conducted for postsecondary vocational educa-
tion (Watson 1977; Carroll and Ihnen 1967), revealing that the dis-
counted benefits exceed the costs. Freeman (1974) has examined returns to
proprietary school training (such as in data processing, auto maintenance,
cosmetology, and barbering). Because the length of training is relatively
shorter, the total cost of proprietary schooling is similar to that of colleges
and universities despite higher tuition. And the returns to the national
economy appear to exceed those of higher education.
Rate of return studies for secondary schools in developing countries
(Psacharopoulos 198 lb) show that private returns average about 15 to 20
percent (the social rates are between 12 and 15 percent). Low returns to
secondary education in middle-income countries, such as Israel and
Greece, reflect specific socioeconomic conditions prevailing in those coun-
tries. For example, under Israel's highly egalitarian system, high school
diplomas do not carry much greater weight in the labor market than do
primary school certificates. For this reason, apprenticeship programs in
Israel are more cost-effective than formal vocational education programs
because, expenditures for formal education are also higher than those for
apprenticeship programs (Cohn 1979, p. 127).
Large-scale tracer studies showing rates of return for vocational school-
ing and training programs are lacking for most developing countries. One
reason for this is that the programs were started too recently to gain a clear
picture of their graduates' earnings (Van Rensburg 1974; Perry 1981).
Benefits resulting from vocational schooling have been reported by
Schiefelbein (1979) in Chile and Colombia. Vocational education gradu-
ates in those countries obtain higher salaries than those from academic
tracks with the same number of years of education. According to Schiefel-
bein, students are increasingly aware of the labor market payoff from
opting to follow a particular curriculum. Puryear (1979) reports the same
results in Colombia for those graduating from SENA programs when
compared with academic students.
Two studies by Bukhari (1968a, 1968b) looked at short- and long-
range benefits derived from specialized secondary vocational programs in
Jordan and Tunisia. These studies concluded that (1) the more specific the
vocational training, the less likely it was to be relevant to the actual needs
of the job market; (2) vocational secondary schooling in Jordan cost ten
times more per pupil than academic secondary schooling; and (3) the
benefit-cost ratio for vocational graduates was 2:1, while for academic
secondary graduates the ratio was 3: 1. Bukhari concluded that vocational



30    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
schools should be integrated into the general academic educational system
to create comprehensive schools because few students were using their
specialized skills in the industries in which they were later employed and
the costs of educating vocational students exceeded those of general
secondary students.
Other studies of vocational schocl programs-in Sri Lanka (Wijemanna
and Welkala 1975), India (Fuller 1976), Barbados (Oxtoby 1977), and
Swaziland (Sullivan 1981)-have pointed to similar conclusions. In gen-
eral, (1) if they have a choice students attend an academic school; (2)
teachers who possess both theoretical and practical ability in vocational
education are difficult to find; (3) industry is reluctant to pay the salaries
expected by vocational graduates; (4) there is a realization that maintain-
ing up-to-date shop equipment similar to that used in modern industry is
too costly for most school budgets; and (5) students often do not take jobs
in the field for which they were trained.
In 1971 El Salvador implemented a diversified secondary school pro-
gram that provided academic and commercial subjects as well as industrial
arts, agriculture, pedagogy, health, navigation, fishery, and hotel adminis-
tration. The objective of the program was to improve the employability of
secondary school leavers and to prepare them for further specialized
studies in secondary schools. McGinn and Balart (1980) report an eval-
uation of this program, carried out by ODEPOR (Oficina de Planificaci6n y
Desarollo Organizacional). The evaluation was a cross-sectional study of
a weighted sample of 613 members of the first class (5 percent of those
graduating in 1974), to determine their success in the labor market one
year after graduation. The basic hypotheses tested referred to the rela-
tionships between employment and field of specialization, earnings and
academic success, and secondary school stream and higher education
career. The findings show no significant differences among graduates of
different fields of study with respec: to labor market success (in terms of
employment opportunities or earnings), even though unit costs of different
educational streams vary considerably. In terms of further education, no
functional relationship was found between the course taken in secondary
school and the field of specialization in higher education. The evaluation
concluded that the diversification reform in El Salvador was a failure
because (1) the results of target schools were no different from those of
traditional academic secondary schools, and (2) diversified schools were
on average more expensive than traditional secondary schools.
In Brazil, a 1970 law required all secondary schools to introduce voca-
tional subjects in addition to academic curriculum, with the ultimate aim
of imparting occupational skills to all high school graduates. According to
Castro (1983, p. 9), this "country-wide school reform that was to offer
practical content to all students turned out to be a failure."



PREVIOUS EVALUATIONS                                        31
A 1984 paper by Urevbu provides a critical appraisal of Nigerian
vocational education. (The term "vocational education" is used explicitly
in government documents to refer to the lower secondary school curricu-
lum, which is terminal for those who do not have the aptitude for
academic subjects.) The author concludes that it is uncertain whether
vocational education in Nigeria promotes economic development any
more than does traditional schooling.
A study of schools in Kenya, however, produced negative preliminary
results. A tracer study of industrial school graduates did not reveal any
employment advantage over a control group of academic secondary
school graduates (Narman and others 1984, table 3). Moreover, the per
student cost of industrial schools is about double that of control schools
(Cumming 1984, para. 6.1.2). In a similar vein, a comparative study of the
introduction of manual training schools in China and India gives a series
of reasons for the reversal of such educational policy in both countries
(Zachariah and Hoffman 1984).
Summarizing the attributes and drawbacks of secondary vocational
education, one may conclude that, especially in the advanced countries,
students have traditionally been guided into different curriculum streams
in accordance with their fathers' educational levels and the students' own
abilities; that academic students expect jobs with higher pay and more
status and perform better in academic subjects than do vocational
students; and that the evidence on the relative earnings of vocational and
academic school graduates is very mixed. This study addresses the ques-
tion whether that pattern is also typical of developing nations, which have
drawn on Europe's experience in the past, or whether vocational schooling
is too recent an innovation to have formed the basis for a stratified
educational system in those countries. The empirical studies on Colombia
and Tanzania throw some light on these questions.
Vocational Education and Economic Development
To determine the way in which the level or type of education influences a
nation's economic output, one must first assume that an increase in
economic productivity requires enhancement of the quantity or quality of
resources or the efficiency with which they are used. Education can make
its most direct contribution to economic growth by improving the quality
of labor-that is, by making the work force more productive, more
disciplined and reliable, healthier, and more mobile. In turn, higher quality
labor may alter the amount of available physical capital by increasing the
rates of savings and investment, and it may boost the efficiency with which
resources are used. The two most important factors affecting the quality of



32    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
labor in any economy are (1) formal education and training and (2) work
experience (on-the-job learning), both of which improve labor produc-
tivity.
There is considerable evidence attesting to the contribution of education
in general (that is, undifferentiated by level and type of schooling) to the
economic growth of many countries. Psacharopoulos (1984) reviewed the
existing data relating economic growth to advancements in education.
Although no universal pattern was found, education's contribution to
growth tended to be higher in Asia, Africa, and Latin America and lower in
Europe and North America. The results of this survey lead to the conclu-
sion that education has been one of the dominant contributors to develop-
ing nations' economic growth.
Biede (1970) has argued that investment in vocational education is a
major factor in economic development. Walters and Rubinson (1981)
estimated the effect of U.S. educational expansion on aggregate economic
activity over the past century, after first controlling for the inputs of
increased labor and capital in the production process. Looking at the
period since the depression of the 1930s, and distinguishing educational
levels (primary through the Ph.D. level), they found that advancements in
secondary schooling had contributed more to the growth of output than
had advancements in higher education. In addition, however, increases in
the number of doctorates granted over this forty-year period had a sig-
nificant positive effect on gross national product (GNP) per capita.
Meyer (1979), using data for more than 100 countries, reported that
secondary school enrollment was a principal factor affecting economic
development in 1950-75. Extending Meyer's analysis, Benavot (1983)
related secondary enrollment rates to the GNP by distinguishing vocational
from traditional academic secondary education. Benavot reasoned that if
vocational education was the most productive form of schooling, there
should be a relationship between a country's level of economic develop-
ment and the vocational share of formal education at the secondary level.
The author distinguished the economic effects of vocational education
from those of general academic secondary education using data on more
than 100 countries over three fifteen-year periods (1950-65, 1955-70,
1960-75). His hypothesis, that countries with relatively stronger voca-
tional systems would achieve faster economic growth, was only partly
substantiated. Apparently, while both vocational and general education
contribute to economic growth, the positive impact of general secondary
education is far more pronounced than is that of vocational schooling
alone. How can these results be explained?
Formal vocational training is thought to fulfill its purpose when it
provides needed skills for the economy that cannot be offered elsewhere
more efficiently. In a country with a very limited industrial base, for



PREVIOUS EVALUATIONS                                         33
example, on-the-job training will not supply all the skilled labor needed
for growth. Moreover, beyond a certain level of sophistication trade skills
demand a theoretical foundation in order to be understood and properly
applied. The abstract knowledge that provides a sound basis for training
in sophisticated labor skills is best imparted in the classroom. In develop-
ing countries, as long as the average level of education is very low,
vocational schools must undertake a substantial amount of general educa-
tion. This function overlaps with that of the general academic secondary
school, and its economic effects are difficult to measure.
In a country where the literacy rate is sufficiently high, vocational
educational institutions can begin to train students in highly specialized
and technical skills requiring a strong theoretical base. Here, too, how-
ever, there is a danger of overspecializing to the point where it is difficult
for graduates to use their expertise because there are not enough highly
skilled jobs available. Very little more is known about training for employ-
ment in general than is known for employment in a particular field of
work. Do certain kinds of curriculum and institutional arrangements
promote the ability to learn from work experience? Blaug (1973, p. 38)
states that "education is economically valuable not because of what
students know but because of how they approach the problem of
knowing."
In sum, there is considerable debate over and conflicting evidence about
whether vocational education imparts knowledge and skills unlike those
found in secondary academic settings. Because the consequences for em-
ployment and for national economic development are considerable, the
questions should therefore be more fully examined.
World Bank Projects
One hundred seventeen secondary education projects assisted by the
World Bank between 1963 and 1982 contained provisions for a diversified
curriculum. (For the evolution of Bank policy on diversified schools, see
Heyneman 1984.) Because of problems involved in distinguishing the
costs of individual components, it is difficult to arrive at exact expendi-
tures for diversification in these projects. A conservative estimate is that
the Bank has devoted 20 percent of its education sector lending to diver-
sified education components over the past twenty years, at a cost of about
$800 million (1983 prices).
Except for the occasional use of tracer studies, the Bank's most common
method of judging the success of diversified secondary curriculum projects
has been to determine whether the original objectives stated in appraisal
reports were being implemented. The five most commonly stated objec-



34    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
tives for diversifying curricula are listed below (with the percentage of the
117 projects that had each item as an objective):
1. To meet qualitative manpower needs (57 percent)
2. To meet quantitative manpower needs by expanding secondary
school enrollment (44 percent)
3. To offer broader curricula to rectify the bias toward academic
education and to make the curricula more relevant to the needs of the
economy (38 percent)
4. To upgrade overall secondary education (38 percent)
5. To improve the employment prospects of graduates (35 percent).
Some of the stated objectives are too broadly conceived to be measur-
able. Ten of fifteen projects that cited enrollment expansion as a goal met
their target. But of these ten, four faced severe problems in placing their
graduates (because of slower than anticipated workforce growth or more
rapid than planned expansion of secondary school enrollments).
With respect to the fourth goal, improving the quality of education,
reasons for failing to meet this stated objective included shortages of
qualified staff, teachers' reluctance to grapple with new orientations, and
inadequate provision of essential instructional material. The lack of these
resources perpetuated conventional teaching practices. Failure to imple-
ment curriculum reform was due most often to lack of materials as well as
lack of teacher commitment and support.
"Hardware" items-buildings, laboratories, workshops, machinery,
and furniture-are the major cost items in most education projects, and
diversified education programs are no exception. Nonetheless, to ensure
the success of diversified education programs once buildings, laboratories,
workshops, and machinery are in place, it is essential to prepare teachers
and school principals for the new curriculum and to provide them with
instructional materials. The World Bank's early loans provided too little
funding for such "software" relative to hardware, since the Bank expected
borrowers to finance this task. For example, funding for the preparation of
teaching materials was included in only 30 percent of the 117 projects, and
the training of school principals to implement diversification was funded
in only 25 percent of the projects. Eighty-five percent of all projects
included proposals for some form of curriculum development such as the
preparation of syllabi for prevocational subjects. Yet no evaluation
scheme was devised to rate the quality or appropriateness of these teacher
guides. Bank analysis revealed further that the major problem with diver-
sified curriculum development arose during implementation: in 75 percent
of projects including software funding, implementation of curriculum
changes needed to meet the needs of the new programs was marginal. This



PREVIOUS EVALUATIONS                                             35
was due either to a lack of instructional materials or to shortages of
qualified teachers.
Another consideration in the discussion of curriculum development
concerns whether the examination system should be revamped to
accommodate changes in the increasing variations in fields of study. In
examinations to qualify for promotion to the next schooling cycles the
persistent dominance of academic subjects tends to foster an attitude on
the part of both teachers and students that these subjects are most essen-
tial. The result in many cases is that the time and interests of teachers and
students are diverted from practical academic subjects.
These observations point to the conclusion that specialized teachers are
needed for diversified secondary education. Teachers trained in standard
lecture procedures for academic subjects are not appropriate for the more
manual material taught in shops and laboratories. Eighty-five percent of
Bank-supported projects providing for teacher training as part of curricu-
lum diversification registered a chronic lack of competent teachers, partic-
ularly in science and practical subjects. Because of these shortages, it is not
surprising that most of the teachers continued to provide conventional
lecturing and exposition. Some projects did report an improvement in the
educational environment at the project schools, with teachers using new
teaching methods extensively.
No evaluation has previously been made of the overall efficiency of
diversified curricula relative to traditional academic or vocational school-
ing, as measured by unit costs and rates of repetition of grades among
students. In fact, because of three major limitations, it was not possible to
make a full scientific assessment of whether the completed projects
achieved their stated or implied objectives or of the effectiveness of diver-
sification as an educational model. First, the Bank completion reports are
written as soon as a project has been implemented and therefore are not
able to provide any inferential data about the effects of diversification on
students' training, expectations, or employment experience. Second, very
few projects have been in operation long enough to permit the examina-
tion of their effects on general economic efficiency. Third, the efficacy of a
system or model cannot be evaluated fairly if the prerequisites-buildings
and workshops ("hardware") or teaching materials ("software")-are not
all in place.
This situation has now been partially remedied. In Colombia and Tan-
zania diversification took place long ago and has been reasonably well
implemented. It has therefore been possible for the World Bank, in coop-
eration of the governments of the two countries, to undertake detailed case
studies of their experience with diversified education.



4. Research Design
and Methodology
THE PRINCIPAL POLICY ISSUE addressed in the study of Colombia's and
Tanzania's school systems is whether the outcomes of diversified educa-
tion differ substantially from those of conventional academic and purely
vocational secondary schooling. The findings are based on two main
dependent variables: what is learned in school, and what graduates
accomplish in the world of work.
This study also investigates the qluestion whether selection into various
curriculum streams affects students' postsecondary school opportunities.
The construction of student profiles can show how students differ across
streams, both intellectually and socially, and the way their career patterns
are affected by exposure to different curricula. This in turn can provide the
basis for an assessment of the impact of students' home and school
resources on their learning and earning capacity. The extent to which the
schools studied nullify or reinforce social stratification processes also can
be documented.
As mentioned previously, whether a diversified educational system has
significant advantages over a nondiversified one is a hotly debated issue.
Intuition suggests that teaching specialized knowledge in schools would be
beneficial for an expanding modernizing economy, but there exists no
definitive evidence to support this position. Opponents of diversified
education have argued that vocational knowledge provides more re-
stricted employment and earnings opportunities than does general
academic knowledge (Blaug 1979; Conroy 1979; Dore 1975; Evans and
Schimmel 1970; Godfrey 1977). If that is true, perhaps school systems
should leave highly differentiated training to firms in the private sector and
specialized vocational training organizations, and instead concentrate on
teaching students to use established knowledge and to use analytical
methods for problem solving.
Future earnings and employment prospects are not the only measures of
the value associated with a given course of study. One can also look at the
economic returns to public and private investment. Differences between
the achievements of graduates of different curriculum programs reflect the
particular skills taught in each program. In the case of competitive labor
36



RESEARCH DESIGN AND METHODOLOGY                                  37
markets, it is assumed that the value of these skills will be reflected in the
going wage rate for a particular job (Lewin-Epstein 1981). Comparing the
returns to investment in specialized training with those accruing from
general academic education can point to situations in which investments
pay off and thereby can indirectly suggest education strategies for enhanc-
ing national economic development. The findings could be used to offer
policy advice regarding priorities in educational investment for the various
types of schools.
In this chapter, we discuss the methodology adopted for testing differ-
ences in characteristics of students and graduates participating in alterna-
tive curriculum programs. We also describe the way in which information
on schools, students, and graduates was obtained and list the main vari-
ables.
Research Design
Chapter 1 gave a short list of hypotheses used to test whether diversifica-
tion leads students to pursue careers that are different from what might be
expected if they followed traditional academic or purely vocational
courses. For example, do prevocational secondary school students obtain
higher test scores in their specialization subject than nonvocational stu-
dents? Do prevocational graduates have lower expectations of attending
university than students in an academic setting? Does diversification lead
to differences in postsecondary activities? How do the benefits of diver-
sification compare with the costs?
Ideally, the benefits to the economy of a diversified curriculum program
could be isolated by using an aggregate production function such as X =
f(D, Z), where, in summary form, X is a measure of outcome, such as
productivity; D is a set of variables denoting the number of graduates who
followed academic, vocational, or diversified curricula; and Z is a vector
of standardizing factors, such as family characteristics and school inputs.'
In fact, this method was ruled out because diversified schools have so far
produced only a small number of graduates; hence, the impact of diver-
sified education is unlikely to be adequately revealed in such an aggregate,
macro production function.
We therefore designed our research to concentrate on three possible
effects of a diversified curriculum on students and graduates: (1) changes
in cognitive achievement in school, (2) changes in student attitudes about
1. That is, the methodology would be similar to the one used to examine the effect of
education on farmer productivity (Jamison and Lau 1982).



38     DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
future career opportunities, and (3) changes in occupational attainment
and employment prospects after students leave school.
These effects can best be described by using the example of graduates of
a diversified school who have specialized in agriculture. In the case of
cognitive achievement (1), the question is whether, other things being
equal, these students have a greater knowledge of agricultural practices
and less knowledge in other fields than do students who followed nonagri-
cultural curricula. With regard to attitudes (2), we need to determine
whether agricultural graduates, regardless of their background, express a
preference for work in rural areas. To show the effect on employment (3),
we must know whether agricultural graduates work in the farming sector
of the economy immediately after leaving school. If there are negative
results in knowledge (1) and attitude (2), there is a strong presumption
against economic benefits (3) from the diversified programs. However,
positive results in (1) and (2) do not establish that there are benefits to the
economy, although they do indicate it is worthwhile to pursue the issue.
Testing Points
We have defined three points in time for assessing the possible effects of
diversified curricula: while the student is still in school; about one year
after graduation, when the graduate might be in his first employment or
continuing his education; and several years after graduation. A longitudi-
nal tracer study framework was adopted. Cognitive and attitudinal tests
were given to students in their last year of secondary school, and a
questionnaire asked about their backgrounds and family characteristics at
the same time. A second questionnaire was administered to the same
students one year after leaving school.
A third assessment, several years later, is not part of this study because
of cost and timing, and because the project was designed to stimulate local
research. The foundations have been laid for tracing the graduates further,
however, so that the governments and local research institutions partici-
pating in this study may collect more information in the future.
Longitudinal Information
It was decided not to collect retrospective data on the graduates' past
school and occupational records, but to undertake a longitudinal study of
current students so as to avoid issues of selectivity bias, which might have
made it difficult to test our hypotheses (Psacharopoulos 1981a).2
2. Data were retrospectively collected on a "pseudo-panel" cohort of 1978 graduates in
Colombia to permit an assessment of some of the longer-term effects of diversification,



RESEARCH DESIGN AND METHODOLOGY                                   39
The starting point in the Colombia and Tanzania case studies was the
schools in a district where diversified curricula have been taught for a
number of years, along with nondiversified schools. To test whether
diversification leads to different educational and employment experiences,
research included several components:
* First, nationwide random sampling was used to obtain a representa-
tive group of high school students in proportion to their enrollment in
the different curricula.
* Second, a survey administered to high school seniors just prior to
graduation was used to collect base-line information for comparing
diversified, vocational, and academic students.
* Third, a follow-up survey was administered to the same cohort one
year after graduation to assess students' experiences immediately after
they left school.
In this study, the subjects (high school seniors) were not assigned to
conditions (the various curricula); instead, observations were made of
students who were already enrolled in the different tracks. Because of the
nonexperimental design, selection bias may well have occurred if some
parents had placed their children in a certain program because they
expected it to lead to better economic rewards than another. Schools
themselves may have restricted entry into their programs on the basis of
the students' past school performances. These choices involve ability
screening or possession of prior information and represent the two most
likely sources of bias in a study of this kind. The effects of these biases can
usually be adjusted for by taking into account the student's background
and some measure of previous school performance or natural ability.
Thus, the longitudinal character of the study makes it superior to a
cross-sectional study of graduates because of the implicit standardization
for attributes that are not easily observed.
In a sense, the design of the study was formulated retroactively, because
there are no pre-test observations on the same or equivalent scales as the
post-test observations. However, a tradition has developed in the social
sciences of seeking out pre-test measures which correlate with the post-test
results within experimental groups, even though they are not measured on
the same scales as the post-test. More easily retrieved measures, such as the
respondent's age, sex, social background, mental ability, and place of
residence, can often be obtained from school records or from respondents
themselves and substituted for the lack of pre-test observations.
although the limitations of this technique were recognized (see Psacharopoulos and Hinch-
liffe 1983).



40    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
To assess the performance of students from diversified schools (the
target group), the research design had to include an adequate number of
observations of students who did not attend such schools (the control
group). Thus, the testing of hypotheses depends on the statistical signifi-
cance of differences between the mean performance of the target and
control groups on a set of indicators, standardized for a set of factors
outside the school setting.
For example, in Tanzania where each school has its own special field of
study, the control and target gro-up comparisons would be across the
vector shown in table 4-1. In Colombia, where several subjects are offered
under one roof or in a separate institution, the comparison is more
complicated, corresponding to the matrix shown in table 4-2. Both vertical
and horizontal comparisons are possible, to address such questions as
where a given subject is best taught (horizontal) or what kind of subject is
more conducive to further employment prospects (vertical).
In addition to determining the direct and indirect effects of home and
school variables on students' cognitive and noncognitive achievements, it
is also possible to use longitudinal data to examine the influence of home
and school on post-secondary school outcomes, such as students' first jobs
and earnings. For example, if the curriculum can be shown to affect
earnings, it may be concluded that a particular educational program
contributes more to individual or social gain, and hence to the economic
development of the nation, than do others. The magnitude of this con-
tribution will suggest where rates of investment might be increased or
reduced.
No data on earnings are available for the many secondary school
graduates in the sample who were pursuing further education and train-
ing. One can, however, analyze these students' academic career paths and
note their decisions to continue training in their field of specialization.
To predict discrete events such as the probability of working, not
working, studying, or not studying, logit analysis has been applied. Be-
cause the dependent variable in this case is dichotomous, an ordinary least
Table 4-1. Target-Control Group Comparisons
in a Single-Stream Diversified School System (Tanzania)
Curriculum      School type
Academic       [Xl] Nondiversified
Commercial      X2  Diversified
Technical       X3  Diversified
Agricultural    X4  Diversified
Home economics  X5  Diversified
Note: Brackets denote the control group.



RESEARCH DESIGN AND METHODOLOGY                                      41
squares (OLS) regression would skew the results (Hanushek and Jackson
1977; Rumberger 1983). The effects of independent variables from logit
estimates can be interpreted in the same way as estimates obtained from
ordinary regression techniques: The sign of the coefficient indicates a
positive or negative effect of the corresponding independent variable, and
the asymptotic t-values indicate whether the effect is significant. The major
difference between logit and ordinary regression techniques lies in the fact
that because the logit model is nonlinear, the magnitude of each effect
depends on the value of the other independent variables.
Finally, rates of return are calculated to determine the social profitabil-
ity of various school curricula.
Hypothesis Testing
OLS regression or logit analysis has been used to search for significant
effects that particular curricula have on students' attitudes or graduates'
performance. The regressions are of the general type: Y = f (curriculum,
family characteristics, individual student characteristics, school character-
istics, regional characteristics), where Y could be any of the outcomes
discussed earlier.
The first test is whether the coefficient referring to curriculum is signifi-
cantly different from zero and, if that is the case, whether it has the
expected sign.
Assuming that the coefficient is significant and has the expected sign, the
next question is: how does its size compare with the cost of obtaining this
effect? For example, studying subject X may raise a graduate's earnings
somewhat more than studying subject Z, but if it costs considerably more
to teach subject X, then subject Z is more cost-effective and hence should
be promoted.
Table 4-2. Target-Control Group Comparisons
in a Multi-Track Diversified School System (Colombia)
School type
Diversified,  Nondiversified,
multi-stream   single stream
Curriculum    (target group) (control group)
Academic (control)   [X.1]         [XI2
Commercial            X21          X22
Industrial            X31          X32
Agricultural          X41          X42
Social services       X51          X52
Note: Brackets denote possible control groups.



42    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Beyond single equation regression analysis one may fit recursive path
models of the type illustrated in figure 4-1, in which five major variables
are interrelated in an initial-to-eventual-outcome time sequence. The vari-
ables include:
1. Family background, one measurement of which is father's occupa-
tion (OCCFATH)
2. Student's ability, as measured by an aptitude test (ABILITY)
3. School type or curriculum the student actually followed (BIAS)
4. Measure of cognitive achievement on subject matter taught in school
(COGN)
5. Eventual labor market performance measured, for example, by labor
earnings (Y).
The recursive model would be
(1) ABILITY = Il OCCFATH
(2) BIAS = 02 OCCFATH + P3 ABILITY
Figure 4-1. Path Model for Assessing the Effect of School Type
on Cognitive Achievement and Graduate Performance
el'z Scool 
type1
,;=e ~~ P
(ackgroun/              \            \~~~~~~p
micieee
'at n                                    A 



RESEARCH DESIGN AND METHODOLOGY                                 43
(3) COGN = 13N OCCFATH + P15 ABILITY + [16] BIAS
(4) Y= 17 OCCFATH + 18 ABILITY + [139] BIAS + [13lo] COGN
The ,(-coefficients in the above equations are standardized betas, and
those of greatest interest to this study are 16, 19, and 1lo. That is, having
standardized for antecedent factors, we want to find out the extent to
which the diversified curriculum affects students' cognitive skills and
eventual labor market outcomes. In other words, one of the main objec-
tives of the inquiry is to quantify the triangle shown in figure 4-2.
The central hypothesis to be tested is that there exist both direct (posi-
tive) effects (arrow A) and indirect effects (arrows B and C) of curriculum
type. To put it differently, the total effect of curriculum on eventual labor
market outcome (and, by extension, the contributions of school type to
economic development) is measured by the sum of direct and indirect
effects.
Total effects of diversified curriculum = 1 + (16) ($10)-
Survey Instruments
Three sets of questionnaires have been used in the Colombia and
Tanzania studies to obtain the information necessary to test the hypoth-
esis:
Figure 4-2. Direct versus Indirect Effect of Curriculum Type
on Student Attitudes and Graduate Performance
Crrium
A
Ieconomic
perfrac
*"B\
achievement
orattudes



44    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
* School questionnaires filled in by the headmasters of the schools in the
survey
* Individual questionnaires admninistered to secondary students still in
the last year of target and control schools
* Follow-up questionnaires administered to the same students one year
after graduating.
Data were obtained from several sources, including students, school
principals, Ministry of Education officials, and graduates. Questionnaires
administered to students were filled out in school during the last quarter of
the final year of secondary school. They provided information on students'
attitudes and behavior, social background, ability, future career plans, and
aspirations. In addition, achievement tests in the various academic and
vocational subjects were administered to all students irrespective of the
curriculum stream they were enrolled in. (For the questionnaires and tests,
see Psacharopoulos and Loxley 1984.)
Questionnaires given to school principals yielded a wealth of informa-
tion on school expenditures, resources such as textbooks and equipment,
the amount of class time devoted to the curriculum, and teacher character-
istics such as work experience and degrees held. Data on school size and
class time were obtained from Ministry of Education officials. From
outside bodies (such as the Exarmination Council in Tanzania and the
Instituto Colombiano para el Fomento de la Educaci6n Superior [ICFES
1979, 1980] in Colombia), student scores were obtained on a battery of
other tests.
Follow-up information on students was collected one year later by
means of a mail survey and homre visits. The questionnaire asked the
graduates what decisions they had made regarding further education and
employment after leaving secondary school.
Main Variables
Tlhese surveys made it possible to categorize each student by sex, age,
social background, curriculum program studied, and type of school
(general, prevocational, public, private). Information was obtained on the
number of times a student repeated a year (whether in high school or
primary school) or changed curriculum programs. It was also possible to
learn about secondary school students' educationally related aspirations,
plans, and attitudes.
The information from all sources has been merged into a single master
data file for each country, using the spss and SAS computer statistical
packages. The data, which are available at the World Bank and the local



RESEARCH DESIGN AND METHODOLOGY                               45
country research institutions, contain a wealth of information that can be
used to analyze questions beyond the immediate scope of this study. It is
hoped that the data base will continue to be analyzed and interpreted for
many years to come.
What follows is a description of the main types of information obtained
about the students and the schools they attended.
Social Background Indicators
Questions to the students asked about parental education, occupation,
and income level, as well as about possessions such as the number of books
in the house. These measures are believed to be a proxy of the social
well-being or status of a family and to reflect the degree to which the home
environment is able to support students in their school endeavors. In both
industrialized and developing societies, parents with more education,
higher incomes, and higher occupational ranking are thought to provide a
better learning environment for their offspring than less privileged fami-
lies. Although education in years and income in the local currency are
relatively straightforward measures of family resources, the rankings of
particular jobs in different societies are more arbitrary.
Future Plans and Aspirations
In addition to data on the objective variables such as age, sex, mental
ability, and socioeconomic status, information was also obtained about
students' educational and occupational expectations. These more subjec-
tive measures can serve as either independent or intermediate variables to
isolate the different effects of curriculum programs. Because the occupa-
tional choices listed in the student questionnaire can be grouped into
categories corresponding to the curricula offered (for example, technical,
agricultural, or commercial), it was possible to ascertain whether students
chose to follow occupations inside or outside their field of study.
In addition, students were asked to state how much they expected to
earn on their first job. These predictions can serve not only to measure
expectations and aspirations but also to assess how realistically students
view the conditions in the labor market.
Exposure to Curricula
To determine their amount of exposure to given fields of study, students
were asked to list the courses they took and the number of class periods
attended in each subject per week. Information was also obtained from the
headmaster on the class time allotted to each subject. In both countries the



46     DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
data allowed us to construct an index of class time devoted to academic
and to individual vocational subjects. As a result, it was possible to group
individuals by degree of exposure to learning and test for the impact of
exposure time on achievement in the various curricula.
School Quality
In both Tanzania and Colombia, questionnaires given to principals and
headmasters of each secondary school in the sample were designed to elicit
quantitative information that would suggest the overall quality of each
school's learning environment. Such data are helpful in obtaining an
average measure of school quality against which to rank individual
schools. The information also makes it possible to control for the effects
on learning of variable school resources and to gauge how strongly the
quality of the school affects students' achievement in various programs.
For example, if the educational attainment of teachers and the number
of textbooks in a school are measured in their respective units and re-
gressed on the achievement of the school's commerce students, the result-
ing standardized partial regression coefficient would indicate the extent to
which the two factors influence student achievement. These coefficients
provide the necessary weights for determining the benefits associated with
the two measures of school quality. The results of the regression approach
could also be supplemented with cost accounting to determine the mone-
tary value of each indicator of quality, provided that the monetary value of
a unit of achievement (output) can be estimated. If there are accurate
measures of the incremental income of graduates strictly attributable to
each curriculum program, that estimate can be made.
Cognitive Achievement Tests
Special tests were administered to each student in the various core
curriculum subjects (such as commerce, agriculture, technical, and
academic). Scores obtained for each student made it possible to compare
strengths and weaknesses in knowledge across all curriculum subjects,
regardless of the stream in which the student was enrolled.3
3. One could argue that paper and pencil tests do not tap the vocational skills learned in
school simply because such tests do not measure manual dexterity or evaluate procedures and
products. Because of the logistical complexity of measuring individual performance, how-
ever, a test of manual skills would have been difficult to administer, especially since we
wished to test all students in the sample. Agriculture schools would not necessarily have had
typewriters, for example, and commercial schools would undoubtedly not have had plowing
equipment. There would also have been difficulties in rating welding, cabinet making, or
physics experiments.                                             (continued)



RESEARCH DESIGN AND METHODOLOGY                                    47
The tests were constructed by the local research teams and were based
on examinations that had previously been given in the various curriculum
programs. The tests are therefore likely to be a valid measure of the
knowledge instilled by the various curricula. For example, an item de-
signed to test commercial knowledge reads, "When a variation of over 1%
in the quantity demanded results from a variation of 1% in price, demand
is said to be: (A) completely inelastic, (B) completely elastic, (C) of unitary
elasticity, (D) relatively inelastic, (E) relatively elastic." Presumably those
exposed to economics in the commercial stream should find this question
easier to answer than those without such exposure. In the same way, test
items in each subject area measured factual knowledge, although higher
levels of reasoning (such as synthesis and evaluation) were required to
answer some of the items.
General Ability Tests
In addition to measuring cognitive achievement in specific curriculum
subjects, we recorded information about students' general abilities. Gen-
eral ability or aptitude tests measure innate mental proficiency and hence
predict how well students may learn in whatever field they pursue. In
contrast, achievement tests (both academic and vocational) measure the
effect of formal instruction on a student's level of knowledge in a specific
area and contain more questions on knowledge and mental skills taught in
the school than do aptitude tests (both verbal and nonverbal). As mea-
sured by achievement tests, performance in the academic and more ab-
stract subjects, such as math and science, is likely to be more closely
correlated with general ability than is performance in subjects that require
special abilities (such as shorthand and manual skills). Differences in the
degree to which the results of general ability tests can be correlated with
the results of achievement tests arise because some subjects require more
ability in abstraction and deduction than others.
We supplemented the achievement test data with measures of students'
general ability for at least two reasons. First, we needed to control for
innate ability in case some students had been selected to follow a given
curriculum stream on the basis of their perceived ability to learn. Second,
because we did not conduct a pre- and post-test experiment, testing and
retesting students to see whether their scores improved, it is difficult to
Although it is not feasible to test manual dexterity in a study of this kind, one can assume
that paper and pencil tests at least measure how well students learn what is taught in the
classroom. Because our paper and pencil tests were based on curriculum content, they
measure cognitive learning and provide at least an indirect measure of the vocational skills
acquired.



48    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
determine whether those who scored highly in curriculum tests of achieve-
ment did so because of what they had learned in school or because they
were more clever than other students to begin with. Distinguishing the
influence of students' innate ability to learn from the influence of what is
taught in school allows one to obtain a "purer" measure of the effect of
school-related factors on cognitive learning.
A good measure of general ability would be a student's test scores on
leaving primary school because such a test would indicate what a pupil
knew before entering secondary school. In Tanzania and Colombia,
however, these early attainment scores could not be secured for various
logistical reasons. Consequently, in Colombia individual scores on the
1981 ICFES aptitude test administered to students in secondary school
were used as measures of ability or proxies for intelligence. (Each student's
identifying code in the sample was matched with the ICFES code.) These
tests were found to be highly reliable and were designed to elicit informa-
tion on general aptitude rather than achievement in a specific subject. In
Tanzania both verbal and quantitative tests of general ability were con-
structed by the local research team and administered to students in the
sample when the base-year data were being collected. The reliability of
these tests was acceptably high.
In both Colombia and Tanzania achievement tests were strongly
directed toward subjects in the schools' curricula, while the aptitude
measures were not; hence the aptitude-achievement distinction is pre-
served in this study.
Noncognitive Outcomes (Modernity)
It is thought that in the absence of a milieu conducive to change,
risk-taking, and innovative speculation, traditional patterns and values
will persist and thereby hinder social change and economic development.
McClelland (1961) has reasoned that if schools can make students less
traditional in their beliefs and attitudes, they will serve as agents of social
change and ultimately influence economic development. In both Colombia
and Tanzania, therefore, noncognitive tests of the degree of students'
psychological "modernity" were constructed and administered to test the
hypothesis that students acquire different attitudes according to the curric-
ulum in which they are enrolled. Students with more traditional attitudes
were expected to be less inclined to enter a university, for example, and
more likely to go directly into the labor market than others.
Presumably, the more education one has, the more modern are one's
ideas. But it remains uncertain wlhy students enrolled in agriculture, for
example, might be more or less bound by tradition than those enrolled in a
commercial program. What aspecrs of secondary school in general and of



RESEARCH DESIGN AND METHODOLOGY                             49
particular curriculum programs can be expected to modernize the atti-
tudes of students? Presumably, traditional customs and attitudes cannot
be significantly altered until a large segment of the community is exposed
at a fairly young age to new ideas and ways of doing things. If this is true,
then primary schools are the obvious place to begin inculcating modern
values because formal education at an early age plays a key role in modern
society (Inkeles 1983).
Tests of modernity seek to measure differences in student perceptions
along a hypothesized traditional-modern axis of personality traits. Move-
ment from the traditional to the modern end of the spectrum is called
modernization. Inkeles and Smith (1974) worldwide, Kahl (1968) in
Mexico, Schnaiberg (1971) in Turkey, Armer and Youtz (1971) in Ni-
geria, and Holsinger (1973) in Brazil are among those pursuing this tradi-
tional-modern concept. Holsinger showed that primary school students in
Brazil acquired more modern values each year they remained in school.
Although Holsinger never was able to isolate specific factors in the school
setting that contributed directly to modernity, he argued that students pick
up modern values from the social structure, patterns of interpersonal
relations, organizational arrangements, and the hidden network of institu-
tionalized rules in the classroom. Learning to respect the dignity of others
may be only incidental to the business of formal instruction, but it may be
powerfully instilled by daily interaction among pupils and between pupils
and teachers. In vocational programs with different degrees of emphasis
on learning and instruction, the extent to which students acquire a more
modern outlook may vary with the curriculum.
Inkeles and Smith have identified a number of characteristics that can be
used to measure degree of modernity. They include commitment to
women's rights, recognition of the value of time, optimism with regard to
national development, use of mass media, identification with nationalism,
active public participation, few kinship obligations, and the ability to plan
ahead. Some of these dimensions are tapped in this study with the use of
items taken from the Inkeles scale. In the Colombian case study, for
example, a typical question asked of respondents was, "How often do you
read a newspaper? (never, seldom, a few days a week, or every day)."
Another question asked was, "Do you think that a person can be good
without having any religious belief? (yes, no)." In the first example, the
''modern" person is more likely to be literate and read outside sources of
news and information than is the traditional student. In the second in-
stance, the traditionalist would view the criterion of morality as tied
almost exclusively to religion, while the "modern" student would be more
open to other interpretations.
The modernity test used in Tanzania was based on analyses prepared by
Inkeles and Smith (1974), Dodd (1969), and Grant (1964). Although the



50    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Tanzanian modernity scale uses many of the Inkeles dimensions of tradi-
tion-modernity, the format calls for agree-disagree responses, not as "yes"
or "no," but in terms of degree (agree strongly, agree, disagree, disagree
strongly). The scale is based on previous studies of modernity that address
dimensions such as openness to change, orientation to the future, concern
with planning, individual competitiveness, and achievement orientation.
The scale can be validated by relating the modernity scores obtained to
indices such as education that are assumed to have some connection with
modernity. The reliability coefficient on the twenty-one-item scale used in
Tanzania is 0.76, which suggests a fairly accurate measure of modern
attitudes.
Further Training and Employment
The follow-up survey solicited information on what the members of the
base-year cohort were actually doing one year after graduation. For those
who were in training or continuing their education, the type of institution
and field of study were recorded. For those in employment, we surveyed
graduates' sector of economic activity, occupation, earnings, hours of
work, and how long they had to wait before getting a job. For those still
seeking employment, information was obtained on their interim financial
resources. Job changes within the one-year span were noted as well.
Test Reliability
In psychometrics, it is essential that measurements be reliable. The
devices used for measurement, such as achievement tests, must be applied
consistently to ensure that the results are repeatable and that error is kept
to a minimum
When a well-anchored rifle is fired, if the shots are widely scattered
about the target the rifle is said to be unreliable. If the shots are closely
concentrated around the target the rifle is considered reliable. Similarly,
when achievement tests are unreliable, it becomes difficult to draw firm
conclusions because the variance among subjects becomes too large.
Highly reliable test instruments display little random error and hence more
consistency. Usually, as random error increases, reliability is reduced until
the point is reached at which all observed variance is random and reliabil-
ity is zero.
Statistical methods have been devised to indicate the reliability of a test
by evaluating the relationship among its components (usually items or
questions). One common procedure is to look at all possible correlations
of test items and compute an average intercorrelation between them by



RESEARCH DESIGN AND METHODOLOGY                               51
correlating all parts of the measure with the total score. When the average
correlation among test items is low, so is the reliability coefficient. In
general, as the number of items and the average correlation among them
increase, reliability takes on a larger vaue as well. Reliability values can
range between 0 and 0.99. Usually 0.60 is the minimum level at which a
test is judged to be reliable.
Reliability coefficients for the various tests administered in Colombia
and Tanzania range above the minimum (from 0.61 to 0.69). The accept-
able minimum coefficient, however, depends on what the tests are to be
used for. If students are to be selected for curriculum programs on the basis
of their examination results, for example, reliabilities from 0.85 to 0.89
would be mandatory because a person's future career should not be
determined by a test that was only moderately reliable. But, if the objective
is to detect differences among groups, then acceptable reliabilities can be
considerably lower than the level obtained in this study.
Psychometric measurements must also be valid in the sense that they in
fact measure what they purport to measure. The validity of tests developed
for this study can be established only by judgment. Do the tests, according
to expert opinion, appear to measure vocational, agricultural, or commer-
cial knowledge? Because all items were extracted from former vocational
achievement tests by teachers familiar with the subject areas in the respec-
tive countries, it is reasonable to assume that the final tests adequately
represent the curriculum of the given vocational field.
List of Variables
The following list summarizes the clusters of the main dependent and
independent variables for which information was collected in this study.
DEPENDENT VARIABLES
For those still in school
Continuous cognitive measures of performance in achievement tests
developed by the local research teams
Attitude measures (such as willingness to enter jobs or to work in
rural areas)
Plans for continuing further education (whether the graduate plans to
seek entry to another educational institution)
Student expectations about future employment with regard to the
sector of economic activity, occupation, earnings, and length of time
before obtaining the first job
For those in first employment
Length of time unemployed since leaving school



52    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Dummy 0-1 variable, the 1 indicating employment in an economic
sector j (for example, in agriculture)
Earnings in first employment
For those who graduated several years previously
Continuous variable measuring the total time unemployed since grad-
uation
Economic sector of first employment (information obtained retro-
spectively)
Sector of current occupation
First employment earnings (information obtained retrospectively)
Current earnings
First occupation (information obtained retrospectively)
Current occupation
Dummy 0-1 variable denoting the graduate's opinion on specific
questions ("Do you feel you use the skills you were taught?")
INDEPENDENT VARIABLES
Student background
Set of 0-1 dummy variables indicating the student's or graduate's
origin (such as farm or nonfarm)
Father's occupation
Father's income
Year of birth
Score on aptitude test
School characteristics
(Common values for all students and graduates of the same school)
Dummy 0-1 variable, the 1 indicating a diversified school (the target
group) and the 0 indicating a school in the control group (private,
technical, or nondiversified)
Dummy variable indicating the specialized field of study (subject bias)
of a particular school
Student-teacher ratio of the particular school or class
School expenditure per pupil
Composite index of school quality, such as the percentage of teaching
personnel having certain qualifications
Regional characteristics
Labor force participation rate of the region
Regional GNP per capita
Region in which the school is located



5. The Colombia Case Study
DURING THE 1970s, Colombia, like many other industrializing nations,
faced the policy dilemma of how to match education and training to
changing economic needs in a way that would expand secondary educa-
tion opportunities on an egalitarian basis. Policymakers concluded that
teaching prevocational courses in secondary schools would help create a
labor force more capable of using advanced agricultural, commercial, and
industrial-manufacturing technology. Thus, a practical bias was intro-
duced into the formal secondary educational curriculum, so as to channel
students' attention toward employment in those sectors.
The Institutos Nacionales de Educaci6n Media (INEMS) introduced in
Colombia for that purpose, are similar to Europe's comprehensive schools
(for a description, see chapter 2). Provision was also made for less formal
training in centers that were closely linked to work within the local
community. Called Centros Auxiliares de Servicios Docentes (CASD), these
schools pull students out of their formal schools for two or three days each
week to give them special "hands-on" training in vocational skills locally
in demand. With the spread of this system, it was expected that the content
of secondary school training would increasingly reflect the prevailing
availability of local jobs. Structural reforms were also undertaken to
decentralize educational administration. (For a more comprehensive ex-
position of the Colombia study, see Psacharopoulos, Velez, and Zabalza
1985.)
The question addressed in this study is whether INEM schools provide
what the policymakers intended. In comparing the effectiveness of INEM
with non-INEM (control) schools, five broad categories of questions were
asked:
1. Do INEM and control schools recruit students equally from all social
backgrounds?
2. Are there significant differences in students' cognitive and noncogni-
tive achievement between INEM and non-INEM schools across the
various curriculum tracks?
3. If there are differences, how can they be accounted for?
4. How strongly do these differences affect postschool outcomes?
5. How cost-effective are the new schools?
53



54    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Included in the second question are differences in graduates' specific
achievements, occupational expectations, earning expectations, and sec-
tor of employment. Included in the third are effects of home environment
and school quality on educational achievement. The fourth area seeks to
determine how strongly the type of school or the curriculum content
influences students' postschool prospects for more education, training,
and access to the labor market.
In this chapter, we first discuss economic conditions in Colombia and
review the Colombian educational system. After describing the character-
istics of the student sample, we report the major differences among the
outcomes of the various secondary school and alternative curricula, based
on the 1981-82 longitudinal data set. The conclusions are checked against
a longer time span in a tracer study of the 1978 cohort.
Background of the Study
With a US$1,400 gross national product (GNP) per capita and a 6
percent average annual growth rate during the 1970s (1.1 percent for
1979-83), Colombia has become a middle-income developing country.
Manufacturing and industry, accounting for 47 percent of gross domestic
product (GDP) and employing 23 percent of the labor force, grew at an
average rate of 7 percent a year during the 1970s. Agriculture, supplying
31 percent of GDP and employing 30 percent of the labor force, increased
by an annual rate of 5 percent during the same period. The service sector,
accounting for 22 percent of GDP and 47 percent of the labor force, grew at
an annual rate of 7 percent.
One-tenth of Colombia's economic growth during the 1970s has been
attributed to improvements in the skill and educational level of the labor
force (McCarthy, Hanson, and Kwon 1983). Employment growth has
been concentrated in the middle-level skilled occupations and in industries
in which education, training, and job experience are important. Labor
productivity has markedly increased in agriculture, whose share of GDP
dropped from 34 to 31 percent in the twenty-year period 1960-80, while
the sector's share of the total labor force dropped from 52 to 30 percent.
Over the same period the share of industry in GDP fell slightly, from 49 to
47 percent, while industrial employment as a share of the total labor force
rose from 19 to 23 percent. The largest growth in employment occurred in
the service sector, which grew from 29 to 47 percent of the labor force.
Output from the service sector as a share of GDP, however, increased only
from 17 to 22 percent.



THE COLOMBIA CASE STUDY                                           55
Educational System
The Colombian educational system comprises a five-year primary
school cycle followed by a six-year program of secondary education (see
figure 5-1). While most children of primary school age were enrolled in
schools in the mid-1980s, only half of the thirteen-year-olds were in
secondary school (see figure 5-2). Of the latter, approximately half gradu-
ate from secondary school.
Until the early 1970s, most secondary schools in Colombia were pri-
vately run and academically oriented. Most graduates of secondary
schools came from an academic stream that prepared them for university
enrollment but little else. These graduates were judged ill-prepared for
most jobs, other than clerical and sales, without further education. Their
orientation, demographic pressures, and a strong social demand for uni-
Figure 5-1. Flow Chart of the Colombian Educational System
Secondary
Vocational
cu (Ariclural, commercial,\
,'industrial, pedagogical,\
social services) 
Primary                  Academic                   University
12345              123456~I                    1234
Vocational Vocational Occupational
exploration orientation specialization



Figure 5-2. Educational Pyramid, Colombia, 1968
Age                                            MALE                              FEMALE
20
18
17_i :::
16                        _       _ _l,,,,                                   
1 4
13                                a
ON~ ~ 1
11                   ~~~~~~~a
_~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.....  .... : .$
300              200              100                0               100              200               300
Thousands
E Population
a Enrollment
a. Portion of pyramid that reflects previous change of minimum age from eight to seven years and related intensive programs to re-enroll dropouts.



THE COLOMBIA CASE STUDY                                       57
versity education resulted in (1) a proliferation of low quality over-
crowded higher education institutions, (2) a surplus of university-trained
graduates (such as engineers) in an economy already experiencing unem-
ployment in excess of 10 percent at all levels of education, and (3) a high
dropout rate among university enrollees (Haddad 1979).
These inefficiencies, combined with perceived changes in the manpower
needs of an expanding economy, prompted the government to diversify
secondary education in public schools. Aligning instruction in secondary
schools with the more vocational-technical skills needed in the labor
market was viewed as an urgent need (Varner 1965; Low-Maus 1971).
In 1973, 20 percent of the cohort of relevant age was enrolled in
secondary schools. Of these, 70 percent followed the academic course of
study, 9 percent studied to become primary school teachers, and 21
percent attended separate vocational programs in agriculture, commerce,
technical, and industrial trades. Under the new program of diversified
education, nineteen special secondary schools (INEMS) were built to sup-
plement the purely vocational and academic schools. It was projected that,
as a result of diversification, by 1980 roughly 48 percent of secondary
school students would be enrolled in vocational-technical programs, and
of these students, 15 percent would be following a diversified curriculum.
However, because of the doubling of the secondary school enrollment base
(from 20 to 40 percent of the secondary-school-age population in less than
ten years), academic schools have maintained their share of enrolled
students, so that INEM students represent only 3 percent of the entire
secondary school cohort (DANE 1976).
The secondary education system in Colombia comprises three types of
schools. First, public and private general academic schools (colegios and
liceos), which teach essentially humanities and science, represent the bulk
of secondary education and prepare students for university entrance. In
1981, close to 1.3 million students (72 percent of secondary enrollment)
were in these schools. Second, trade and teacher training colleges, or
single-track trade schools (univocacionales), represent another 25 percent
of the secondary school population; these specialize in agriculture, com-
merce, industrial-technical subjects, or teacher training. Although their
students take core academic subjects, the schools emphasize vocational
subjects throughout all six years. The third type of school is the INEM,
which corresponds to the comprehensive multi-track secondary schools
found in industrialized countries, offering academic and vocational cur-
ricula under one roof. While accounting for only a small percentage of
secondary school enrollment, these schools are thought to represent the
"wave of the future" since it is hoped that they would ultimately be able to
attract students away from academic programs in the liceos and colegios
and into trade courses.



Figure 5-3. Structure of the Six-Year INEM Program, Colombia
NONFORMAL
EDUCATION         LABOR FORCE
l                                                                                                                                           oInstitutes of higher education 
Years   I         I       I         I          I         I        I          I         I        I        I        I          I              ___               rfsinl
, \ t \ \ I / \ / \, x, . ____-__,_IMiddle
osientat on   3     Acadetwnic    |              Industrial                     Co--meroal     | A-sicollral   |         Skcial services              Vahndo 
First decision n                                                                                                                                                             wor.er_
Vocational  2                                                    Eight exploratory counses
exploration  I1                                                      of nine weeks each
skilled
Grades                         _-A_ _
S                                           ~~~~~~~~~~~~~~~Pri'mary
4                                                           education
131



THE COLOMBIA CASE STUDY                                       S9
The INEMS offer a six-year program, as shown in figure 5-3. During the
first two-year cycle, students are exposed to prevocational subjects to
acquaint them with knowledge and career options in the various trades.
Vocational learning occurs side-by-side with regular academic course
work, with an emphasis on core academic subjects. The second two-year
cycle is oriented to vocational subjects, such as agriculture or commerce,
along with continuing academic study. Those who leave school at this
stage are regarded as prepared for further training, for example in Servicio
Nacional de Aprendizaje (SENA, national apprenticeship service) or in
on-the-job apprenticeships. During the last two-year cycle, further spe-
cialization takes place. For example, a student enrolled in the industrial
training program in the preceding cycle might now focus on metalwork,
construction, or electrical mechanics, although more than half the time is
still devoted to traditional academic course work.
Regardless of track or school placement, all students spend two-thirds
of class time on academic work and one-third on vocational subjects in the
first two years of the six-year secondary school cycle. Specifically, students
take mostly mathematics, languages, social studies, and courses in natural
sciences. In the last four years of the cycle, students enrolled in vocational
programs specialize increasingly in the vocational subjects of their choice.
By the last year of secondary school, agriculture students, for example,
spend on the average sixteen of thirty-eight hours each week on such
agricultural topics as crop production, cattle raising, and cooperative
management. This represents 42 percent of their class time. In certain
other tracks, the ratio is closer to 50 percent. Hence, in examining the
effectiveness of this education, it becomes imperative to assess separately
exposure to course work variables, both academic and vocational, in order
to control for differences in learning outcomes by subject knowledge.
After completing the six-year course, students are issued diplomas,
which give them the option of attending a university or some other form of
higher education, such as a technical institute, provided the holder also
passes the requisite entrance exams. Other graduates of the course seek
employment in highly specialized trades.
A number of observations can be made concerning the quality and
content of Colombia's secondary schools. The academic secondary
schools vary greatly in quality of existing facilities, staff, and equipment.
Some of the private schools were not originally built for educational use
and therefore lack many features required for teaching modern science and
mathematics. Traditional teaching methods with an emphasis on lecturing
and rote learning are still the rule. But a recent shift away from the
humanities and languages to mathematics and science has begun to
broaden the range of teaching styles. Examination practices are geared
closely to university selection.



60    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Industrial education is mostly public, and the retention rate is similar to
that found in the general academic schools. Five- to six-year courses lead
to positions as skilled craftsmen in industry (for example, metal mechan-
ics, electrical mechanics, construction, foundry work, and auto repair).
Standards of training vary widely, but most graduates have been able to
find employment in large manufacturing and other industries.
Commercial education is mostly private and mainly attracts females.
Completion of a four-year program yields a certificate in typing, secre-
tarial services, and bookkeeping; those who attend the full six-year course
usually become accountants and data clerks.
Agricultural schools are 20 percent private and 8 0 percent public. Many
lack facilities and equipment for proper teaching of agricultural technol-
ogy. Of all single-track vocational schools, agricultural schools are
perhaps most criticized for being too costly and inefficient-either
teaching unduly specialized skills at too early an age or not meeting the
demands for skilled farmers or middle-level agricultural technicians.
Curriculum schedules for the various tracks in all types of secondary
schools, both public and private, are established by the Ministry of Edu-
cation. Some variation occurs within schools because some lack the re-
sources for teaching certain subjects.
Sampling Procedures
Sampling proceeded in two stages. In the first stage, twenty-seven cities
(four large and twenty-three medium-size) were identified as having both
INEM and control vocational schools (see figure 5-4). Sixteen INEM schools
were selected. Within each INEM school, the population was established
for all final (sixth) year students enrolled in each of the five curriculum
programs. Then students from each curriculum program were randomly
selected in proportion to their number in the INEM population. Sixty-seven
percent of the students in INEM sc]hools were included in the sample. The
sample size was established to ensure enough observations in each curricu-
lum program cell for a statistically significant test of differences among
curriculum programs, with allowance for a possible 40 percent loss of
respondents during the follow-up survey.
In the second stage of sampling, students in control schools (academic
or straight vocational) were matched to those in the INEM schools already
selected. The control schools were selected on the basis of their geographic
proximity to the INEM schools and their similarity to the INEM schools in
the curriculum programs offered. Five percent of Colombia's students in
non-INEM schools were sampled.
To select academic control schools, the following procedure was used.
Because these schools are either private or public and are located in either



THE COLOMBIA CASE STUDY                                                              61
medium-size or large cities, a determination was first made of the number
of academic schools necessary to represent each substratum (for example,
government schools in large cities, private schools in medium-size cities).
The goal was to include at least one control academic school in each of
the twenty-three medium-size cities, and two or three control academic
schools in each of the four large cities. The schools were then selected on a
completely random basis. The same procedure was used to select single-
track vocational schools teaching pedagogy, industrial trades, and com-
mercial subjects.
The INEM and non-INEM subsamples cover roughly the same proportion
Figure 5-4. Location of the INEM and Control Schools Surveyed,
Colombia
cwwibLeaoc see
-      OCaSagenaW    )
- ->   .   o   Urban centers with dtversifed
p     .        .                                             secondary ad control schoolt
Monter,a   <                             Urban centers with control
/         ,    Cucuta                        schools only
of
ps, \   _  /      k  I t _ -'xPamplon4          - . - Deparirnent boundar-es
J Buc1wansa@7 ~... '    - - lnrernatiosn.l boundaies
\) Mcdsllan         : ',
0 
ftevk c *          Nontrii                        / a
_           ,GP aln         _
,a,r4tt    Mar- 9s''~tal'  B O   'r   
Andalh(cia !    t -stgds'-ago _/-
0 V P         Vrllas,cencsa_
,  kCah 0,   dro  |a  ,- --- 
'   l0._' 5a.. j            ,          -- -_ ,( 
/ Neisap                                        7)d
(.>  \   /    >  r ~     '>                 \. VENEZUELA
p -n   F *Florencta    >  /         -J
PasO, '     >                 txBRAZIL
ECUADOR\- '                      ,
PERU    >                      MILES   S
1                0 o501    S
.  ...... J.0              250
KILOMETERS



62     DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
of the student population in each curriculum program-for example,
27 percent of commerce students in INEM schools and 30 percent of
commerce students in non-INEM schools. Nearly twice as many students
following academic programs were selected from control schools as from
INEM schools (1,471 compared with 871) in order to take account of the
heterogeneity of academic schools and to achieve a proportionate repre-
sentation of public and private schools.
In the small substrata, such as private schools teaching industrial sub-
jects, all schools and students were necessarily selected. For social services
and agriculture, in which enrollments are very small, the selection
embraced all schools and students graduating in the cities chosen for the
sample and in nearby towns. The overall sample comprised a total of
8,051 students. (table 5-1).
Because of the limited time horizon of the follow-up of the 1981 cohort,
the Colombian case study also included a retrospective survey of nearly
2,000 students who left INEM and control schools in 1978. The findings
based on this component of the study are presented later in this chapter,
following the analysis of the 1981 cohort.
In summary, there were two stages of sampling. First, sixteen INEM
schools were selected and students sampled by curriculum modalities in
such a way that 67 percent of all students in these schools were included in
the samples. In a second stage, control schools were sampled in such a way
they could be matched with INEMS already chosen, based on geographic
proximity to the INEM schools. More important, all schools had to have
similarity in course of instruction (that is, commerce with commerce,
agriculture with agriculture, and so forth). Students in the control sample
constitute 5 percent of Colombia's non-INEM students. (Descriptive statis-
tics for the 1978 and 1981 cohorts are given in tables 5-64 through 5-67.)
Table 5-1. Sample Distribution by School Type and Subject,
Colombia, 1981 Cohort
Subject    INEM       Control     Total
Academic      871      1,471 (35)- 2,342
Agricultural    174     370 (11)   544
Commercial    957       973 (24)  1,930
Industrial    869       945 (17)  1,814
Pedagogical    __b      767 (22)   767
Social services  380    274  (4)   654
Total       3,251(16)  4,800 (113)  8,051(129)
a. Figures in parentheses are the number of schools.
b. INEM schools offer all subjects except the teacher training specialization.



THE COLOMBIA CASE STUDY                                                   63
Socioeconomic Profile of 1981 Cohort
Students entering secondary education often have a preferred school
and course of study in mind. A student's choice of secondary school may
be influenced by how well he or she has done in primary school and by
certain family attributes, herein referred to collectively as socioeconomic
status. Such characteristics as parents' income, occupation, education, and
ownership of property indicate the family's resources for assisting children
in further education. A student's age, sex, and innate ability are also
thought to influence the choice of secondary school program.
Table 5-2 compares population and sample distributions by father's
occupation and educational level. As the literature would lead one to
expect, the graduating secondary school cohort comes disproportionately
from families whose household heads are in professional-managerial
occupations or have secondary or tertiary education. Professional classes
are three times overrepresented in the secondary school sample relative to
their number in the overall population. Students from blue-collar families
are slightly overrepresented as well, reflecting the large number of our
Table 5-2. Sample Representativeness by Father's Sector
of Occupation and Educational Level, Colombia, 1981 Cohort
(percent)
Colombian    Base-year   Representation
Father's background    population' sample (1981)     indexb
Economic sector of occupation
Agriculture                   35           19             54
Commercial                    18           19            105
Industrial (construction,
manufacturing, trades)      28           39            140
Professional, technical        6           17           283
Services                      13            8             62
Educational level
No educati6n                  25            3             12
Primary education             55          62             113
Secondary education           16          28             175
University education           2            6            300
a. Based on census data, DANE (1976, p. 26).
b. The base-year sample as a percentage of the Colombian population. A value below 100
indicates underrepresentation of the particular socioeconomic group in the sample. A value
over 100 denotes overrepresentation.



64   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
sample in vocational schools and curriculum programs. Farmers are
underrepresented proportionate to their number in the population.
This comparison of the characteristics of the sample with those of the
Colombian population as a whole gives an idea of the sample's selectivity.
Comparing fathers' occupations with students' curriculum programs
(table 5-3) shows that proportionately more INEM students come from
blue-collar worker families than do control group students. INEMS were
built in the poorest areas of the cities and were intended to provide
educational opportunities to students from poorer socioeconomic back-
grounds. Presumably, some of these students might not have attended high
school if INEMS had not been introduced.
Table 5-3. Father's Occupation by Students' School Type
and Curriculum, Colombia, 1981 Cohort
(percent)
Curriculum and
father's occupation    INEM   Control
Academic
Nonmanual employee   40        37
Business owner         10      12
Blue-collar worker     41      29
Agricultural worker     9      15
Agricultural
Nonmanual employee   34        16
Business owner          7       6
Blue-collar worker     59      36
Agricultural worlker   21      60
Commercial
Nonmanual employee    37       38
Business owner         11      15
Blue-collar worker     42      37
Agricultural worker     8       9
Industrial
Nonmanual employee    35       39
Business owner          8      12
Blue-collar worker     51      46
Agricultural worker     7       9
Social services
Nonmanual emplayee    37       36
Business owner         12      13
Blue-collar worker     56      39
Agricultural worker    11      23
Note: For each group, column percentages do not necessarily add up to 100 because of
overlapping classifications.



THE COLOMBIA CASE STUDY                                                    65
There are significant differences in family characteristics among stu-
dents following different curriculum programs (table 5-4). Essentially,
more affluent families prefer academic and commercial subjects for their
children and prefer control to INEM schools. Regardless of the type of
school, they deem agriculture least attractive for study, followed closely by
industrial training. Thus, the social demand for secondary education
among wealthier students favors non-INEM schools and academic and
commercial subjects.
Internal Efficiency of Schools
Because students in all curriculum programs and types of secondary
schools had been exposed to some academic course work, it seemed
reasonable to test their academic achievement. The questions used ad-
Table 5-4. Students' Family Characteristics by School Type
and Curriculum, Colombia, 1981 Cohort
Curriculum and family characteristic   IVEM   Control
Academic
Family income (thousands of monthly pesos)  24.60   37.56
Father's education (years)                  6.9     7.8
Number of books at home                     79       91
Agricultural
Family income (thousands of monthly pesos)  17.23   19.34
Father's education (years)                  5.4     4.7
Number of books at home                      60      57
Commercial
Family income (thousands of monthly pesos)  22.86   31.08
Father's education (years)                  5.6     6.6
Number of books at home                     71       82
Industrial
Family income (thousands of monthly pesos)  21.37   25.52
Father's education (years)                  5.6     6.3
Number of books at home                     67       84
Social services
Family income (thousands of monthly pesos)  19.59   29.56
Father's education (years)                  5.5     7.4
Number of books at home                     76       84
Overall
Family income (thousands of monthly pesos)  22.22   30.28
Father's education (years)                  5.9     6.8
Number of books at home                     66       73



66    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
dressed problems frequently found in biology, chemistry, math, and social
science classes, where achievement scores could be aggregated or used
separately. Vocational achievement tests were based on the nationally
recognized curriculum content in various vocational subjects. The tests
were developed by the local research team and exhibit coefficients of
reliability above 0.60 (Velez and Rojas 1983).
Cognitive Achievement
Differences in achievement scores by curriculum program and school
type also were studied (table 5-5). lTo help in the comparison of students in
different programs and schools, the tests were standardized to a mean of
50 and. a standard deviation of 10 (see figure 5-5). On the academic test,
students in the INEM academic program scored highest (54), and students
of agriculture in control schools scored 45, or about one full standard
deviation, below the top group. (A. ten-point difference in this context is
equivalent to the test score differences often found between minority and
other students in the United States.)
The mean scores of students following different programs in INEM and
control schools, shown in table 5-5 (not yet adjusted for out-of-school
factors), reveal that those who specialized in a particular subject scored
higher than others in the respective test. Figure 5-6 also shows that INEM
schools often outperform control schools in other subjects, including
academic knowledge. For example, agriculture students in INEM schools
scored three points, or one-third standard deviation, above agriculture
Figure 5-5. Relationship between Standard Deviation Units and the Per-
centage of Students Scoring above the Mean in a Normal Distribution
19% 
34% -
0    10    20    30    40    50    60    70    80    90   100
Mean 50
Standard deviation = 10



THE COLOMBIA CASE STUDY                                       67
students in control schools. In tests of knowledge of academic subjects,
INEM students often perform better than their control counterparts. This
may be explained by the fact that in INEM schools commercial, agricul-
tural, and industrial students are taught academic subjects alongside INEM
academic students. Furthermore, the INEM vocational programs usually
require an additional rigorous program of applied math and science
courses aimed at reinforcing the vocational specialization. The results just
described suggest that INEM schools have their intended effect of exposing
all their students to a quality academic program.
Mean achievement scores by selected characteristics of the entire sample
also were considered in the study (table 5-6). One notable feature is that
students with a farming background score substantially higher than the
rest in the agricultural test.
Because the comparison of mean differences in INEM and control test
scores described so far has not taken account of out-of-school influences,
such as sex, age, innate ability, aild social origins, it is uncertain whether
the differences result from what is learned in school. For example, students
in control schools might be initially better prepared when entering second-
ary school than are INEM students. To try to distinguish the effects of these
influences, we first looked at students' scores in general aptitude tests.
General aptitude tests can be used to control for reasoning power, or
personal ability to think, that might influence one's score in tests of
specialized knowledge. Scores were taken from school records based on
the national examination administered by the Instituto Colombiano para
el Fomento de la Educaci6n Superior (ICFES 1979, 1980) to students in
secondary school. These tests, similar to the American Scholastic Aptitude
Tests (verbal and quantitative), tapped reasoning power and were not
related to particular subjects taught in school.
Regression analysis was then employed to control for the influence of
general reasoning power and other out-of-school factors on students'
scores in tests on the curricula they had followed. Students of pedagogy
were treated as the reference group (omitted dummy variable) because that
subject is not taught in INEM schools. Eleven variables were included as
proxies for out-of-school influences on learning. From table 5-7 it is clear
that verbal aptitude is a prime out-of-school determinant of the general
academic achievement measures. For the total achievement score of INEM
students following an academic curriculum, the regression coefficient is
4.603, with a t-value of 9.30. This is interpreted as follows. The size of the
coefficient denotes that in the academic achievement test, INEM academic
students scored an average of 4.6 points above the reference group of
students following pedagogy programs. Since the t-value is higher than
two, students attending academic programs in INEM schools score much
higher than the reference group.



Table 5-5. Mean Scores by School Type and Curriculum, Colombia, 1981 Cohort
Achievement score
School type
and curriculum   Academic  Agricultural Commercial Industrial Pedagogical Social services
INEM
Academic        54         51           49         51         51           54
Agricultural    50         61           47         51         50           49
Commercial      48         51           60         49         49           49
Industrial      50         51           48         58         50           49
Social services    47      51           47         49         53           56
ON                               Overall          50.5       51.4        51.4       51.6        50.2         51.0
Control
Academic        50         49           49         49         49           50
Agricultural    45         58           46         49         49           46
Commercial      50         48           57         48         47           49
Industrial      51         48           46         53         48           48
Pedagogical     49         48           45         47         58           50
Social services    49      51           47         47         49           54
Overall         49.5       48.6         48.6       48.4       49.8         49.0
Note: Raw scores (x) on all tests are standardized (normalized) to a mean of 50 and a standard deviation of 10, according to the formula
T = [10 (x - x)/S,] + 50.
Scores in boldface refer to those taking the test in the area of their curriculum specialization.



Figure 5-6. Academic and Vocational Achievement by School Type, Colombia, 1981 Cohort
60~~~~~~16
60 so777
0~~~~~~~~~85
C) ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~5
50- 40
Ca~~~~~~~~~~~~~~~~~~~~~4
-o~~~~~~~~~4
XC~ ~ ~ ~ ~~4
Cu~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~x
Qt                       .. . ....                                                                            .   .. .
3!  0
Test: Academic    Vocational  Academic    Vocational  Academic    Vocational  Academic    Vocational........
Ln ~ ~ ---Arclua-..]                j.---  omril--.-             -s----Inutil- .             ---.oilsrie
CURRICULUM.
Stnar Tevitio: 10dei  Voaioa ANEcNadIemic        Vctoa           cdmc       Vctoacdmc                   Vctoa



Table 5-6. Mean Achievement Scores by Selected Student Characteristics, INEM and Control Schools,
Colombia, 1981 Cohort
Achievement score
Student characteristics   Academic Agricultural Commercial Industrial Pedagogical Socialservices
Sex
Male                         51          51          48         51          49           49
Female                       49          49          52         48          51           51
Father's occupation
Farmer                       48          52          48         49          49           49
°                         Blue-collar                   49         49           50         50         49            49
Nonmanual employee           51          50          50         50         50            50
Business owner               51          50          51         49          50           51
Self-employed                50          49          50         50          50           50
Type of school
Public                       51          51          50         50          49           48
Private                      48          47          49         47          50           51
Other
Urban born                   50          50          50         50          50           50
Rural born                   48         50           49         49          49           48
Repeater in primary school   49         50           50         50         49            50
Did not repeat               50         50           50         49         50            50



THE COLOMBIA CASE STUDY                                       71
Of the ten dummy variables that were used to denote school placement
and curriculum program followed, INEM-academic has the strongest effect
on total academic score. Judging from the signs and magnitudes of the
regression coefficients, INEM schools on balance seem to teach academic
skills better than do the control schools.
The results of regressing achievement scores in industrial, commercial,
agriculture, and social services tests on variables for school type and
out-of-school influences are shown in table 5-8. Here, unlike in the pre-
vious table, we tried to isolate the effect of INEM versus control schools on
vocational achievement. For example, the industrial achievement of in-
dustrial students enrolled in INEM and in control schools is compared.
(The reference group in this case consists of all students, whether INEM or
control, who take no industrial course work whatsoever.) With the excep-
tion of agriculture, the results for the vocational tests show that INEM
students significantly outperform those in control schools in these tests,
even after allowance for the effects of a series of out-of-school variables.
Influence of School Resources
Students' achievement may be influenced by specific school resources.
Table 5-9 shows several quality indicators by school type and curriculum
program. For the various INEM programs, total expenditure per student is
very similar and averages 1,615 pesos a month. In non-INEM schools,
academic and commercial curricula cost less per student (44 percent and
31 percent, respectively) than the average cost per student across all
programs, while the agricultural and industrial specializations cost sub-
stantially more than the average. According to these figures, INEM schools
are not much more costly to operate than the smaller public and private
single-track schools. (These figures do not cover the capital costs of
equipment or buildings and do not separate public from private control
schools, a distinction that will be made later.)
Pupil-teacher ratios are similar among INEM tracks, and classes in INEMS
are on average smaller than those in control schools. INEM schools have
larger enrollments than control schools and have done remarkably well
not only in adequately staffing their programs but also in attracting
qualified teachers. INEM students clearly spend more of their class time on
vocational subjects, and they also have better access to laboratories for
practical work (see table 5-9).
Do the variables that affect school quality also influence cognitive
achievement scores in academic and vocational subjects and, if so, are
there differences between INEM and control schools? To test for such
influences, one can again apply a multiple regression procedure, but with
an expanded set of control variables.
The influence of in-school and out-of-school variables on two achieve-



Table 5-7. Academic Achievement as a Function of School Type and Out-of-School Influences,
Colombia, 1981 Cohort
Total academic           Verbal            Mathematical
achievemnen!         achievement          achievement
Independent variable  Coefficient  t-value   Coefficient  t-value   Coefficient  t-value
Out-of-school factors
Male                    0.264      0.87       2.254'    9.20      - 1.047     3.52
Age                    -0.090      1.20      -0.554'    9.13        0.109      1.49
Family owns businessb    -0.0951'   2.68      0.275      0.9S       0.105      0.30
Household books         0.010'    4.07        0.004'    2.09      - 0.010     4.29
Father's education      0.087a     2.36       0.034      1.14     -0.115a    3.17
Log of family income    0.268'    3.27        0.244'    3.68        0.059     0.72
Father is employerb     0.802      1.85       0.456      1.29       0.146      0.34
Father blue-collar    - 0.441      1.15       0.325      1.04       0.297     0.79
Father nonmanual
employee            -0.419       1.49       0.320      1.40     -0.058       0.21
Verbal aptitude         0.416a   28.24        0.416'   34.86        0.051'    3.50
Math aptitude           0.067a    4.79        0.250a   22.06      -0.013      0.97



School type and curriculum
INEM
Academic               4.603a    9.30        1.274a    3.17      - 3,976a     8.17
Agricultural           1.191     1.27        1.359      1.78       5.519'    5.97
Commercial             0.329     0.67        0.193     0.49        5.823a    12.13
Industrial             2.626'    4.89        2.040a    4.68      -7.685      14.53
Social services      -0.168      0.26      -0.816       1.58     -6.832a    10.92
Control
Academic               0.732     1.58      - 0.329     0.87      - 3.916a     8.58
Agricultural         - 0.474     0.71      - 0.456     0.85        4.343a     6.65
Commercial             0.277     0.56        1.720a    4.27        2.646'    5.42
Industrial            1L541'    2.88         2.505     5.77       -5.345a    10.15
Social services      - 0.567     0.60      -0.325      0.43      -0.055'    0.06
Constant             22.062'   10.90        27.753a   16.91       49.009a    24.63
R 2                         0.22                  0.43                  0.23
Sample size                5,838                 5,838                 5,838
Note: Omitted occupational categories are farmer and self-employed; omitted school subject is pedagogy.
a. Statistically significant at the 5 percent level or better.
b. Despite the apparent similarity between "Family owns business" and "Father is employer," these variables behave quite differently from one another
statistically.



Table 5-8. Vocational Achievement Scores as a Function of School Type and Out-of-School Characteristics,
Colombia, 1981 Cohort
Agricultural          Commercial             Industrial          Social services
achievement           achievement           achievement           achievement
Independent variable     Coefficient   t-value   Coefficient  t-value   Coefficient  t-value   Coefficient   t-value
Out-of-school factors
Male                             0.950'    3.66       -0.804'    3.20         2.337'    8.07      -1.495       5.72
Age                              0.141      1.81        0.153'    2.10        0.023     0.30        0.069      0.89
Family owns business            -0.497      1.34      -0.497      1.43      -0.551      1.48      -0.310       0.85
Household books                  0.007a    2.59         0.003     1.38        0.009'    3.31        0.011a     4.24
Father's education               0.006      0.15        0.089a    2.50        0.026     0.69        0.153a    4.06
Log of family income             0.120      1.40        0.255a    3.18        0.125     1.45        0.351'    4.15
Father is employer             -0.593       1.30        0.346     0.81      -0.635      1.40        0.005      0.01
Father blue-collar              -1.427'    3.55         0.012     0.03      -0.972'    2.42       -0.280       0.71
Father nonmanual employee       -0.803'    2.73         0.041     0.15      -0.531      1.81      -0.701'    2.42
Verbal aptitude                  0.233'   15.20         0.132a    9.17        0.112'    7.27        0.356'   23.50
Math aptitude                    0.001      0.05        0.051a    3.74        0.029     1.96        0.001      0.04
Curriculum
INEM specialty                  10.970'   12.12        12.921a   35.86        6.385a   14.98        5.855a    10.20
Control specialty               11.500a   19.53         8.851a   23.13        1.070a    2.53        1.084      1.17
Constant                        33.235a   15.95       32.249a   16.42       38.231'   18.30        25.996a   12.64
R2                                    0.13                  0.25                  0.09                   0.15
Sample size                          5,838                 5,838                 5,838                 5,838
Note: Each dependent variable is the score on a specialization test. The INEM specialty dummy refers to students enrolled in the INEM track
corresponding to the particular test in that column. The control specialty variable represents students in the non-INEM specialty track. Regression
coefficients are to be interpreted with reference to the omitted category-that is, all students not following that particular subject.
a. Statistically significant at the 5 percent level or better.



THE COLOMBIA CASE STUDY                                                 75
ment scores-one in the academic and one in the commercial curriculum
program-has been studied (table 5-10). The results indicate that school
quality has a considerable impact on both scores. Per pupil expenditure, a
major proxy of school quality, significantly influences academic, but not
commercial, achievement.
Another way to assess the impact of the INEM curriculum is depicted in
table 5-11. For this table, the entire sample is divided into five groups on
the basis of curriculum. For example, the first group under consideration
consists of all students following an academic curriculum (1,545). Of
these, roughly half (730) were in INEM schools and the remainder were in
control schools. After statistically controlling for the influences of back-
ground and ability factors listed in table 5-10, as well as for teachers' mean
salary, in a multiple regression equation, it is possible to determine the
effect of curriculum on students' test scores. In the first example, INEM
program students earned 3.28 points more on average than control stu-
dents in the academic achievement test, after adjustments for differences in
background, ability, and teachers' mean salary. Students in INEM indus-
Table 5-9. Mean Indicators of School Quality by School Type
and Curriculum, Colombia, 1981 Cohort
Per pupil                                         Pupill
School type      school       Teacher's     Vocational  Practical  teacher
and curriculum    expenditure'   qualificationsb   intensity'    labsd  ratio
INEM
Academic          1,609          37            28         50        30
Agricultural      1,653          50            42         74        36
Commercial        1,612          47            50         47        28
Industrial        1,619          45            53         45        30
Social services   1,609          63            49         36        32
Average           1,615          46            46         47        30
Control'
Academic          1,029          24            15         27        28
Agricultural     2,465           44            40          15       30
Commercial        1,270          55            20         50        36
Industrial       2,621           25            36         12        20
Pedagogical       1,062          70            30         45        36
Social services   1,783          38            30         31        18
Average           1,845          43            25         40        34
a. Monthly recurrent per pupil expenditure (in pesos).
b. Percentage of teachers with in-service training.
c. Percentage of time spent on vocational subjects.
d. Percentage of schools with science laboratories rated as having an "above average"
supply of equipment and materials for biology, physics, and chemistry courses.
e. Control category includes private schools.



76   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-10. Academic and Commercial Achievement Scores
as a Function of Out-of-School and School Characteristics,
Colombia, 1981 Cohort
Academic          Commercial
achievement        achievement
Independent variable   Coefficient  t-value  Coefficient  t-value
Intercept                   23.748'   10.46    30.910'   13.62
Out-of-school factors
Male                         0.531     1.56   -0.598       1.75
Age                        -0.070      0.82     0.484      0.57
Family owns business       -1.229'    3.07    -0.711'    1.78
Books owned by household     0.012'    4.25     0.003      1.06
Father's education           0.106'    2.49     0.345      0.81
Log of family income         0.118     1.31     0.235'    2.63
Father blue-collar         -(.880'    2.10    -0.390       0.93
Father nonmanual employee   - 0.448    1.42     0.236      0.75
Father is employer           1.119'    2.27      1.071'    2.17
Verbal aptitude              0.387'   23.38     0.135'    8.17
Math aptitude                0.064a    4.08     0.062a    3.97
School type and curriculum
INEM
Academic                   4.325    7.71     3.832'    6.83
Agricultural               0.123     0.13      1.862a    1.97
Commercial               -0.007      0.01    15.289a   28.12
Industrial                 1.833'    3.10     3.478a    5.88
Social services            0.291     0.43     2.551'    3.80
Control
Academic                   0.261     0.48     4.154'    7.58
Agricultural             -1.636'    2.24       1.554'    2.12
Commercial                 0.763     1.43    11,043'   20.75
Industrial                (C.372     0.53     0.412      0.58
Social services          -1.573      1.36      1.956     1.70
Other school characteristics
Private school             -4.057'    9.59    - 1.418'    3.36
Mean teacher salary        -C.62E7'  15.96    -0.32E7'  10.08
Per student expenditure      C.002'   18.91     0.001'    9.70
R2                                0.30                0.31
Sample size                      4,233               4,233
Note: Curriculum categories are evaluated in relation to pedagogy (the omitted category).
a. Statistically significant at the 5 percent level or better.



THE COLOMBIA CASE STUDY                                              77
Table 5-11. Gains in Achievement Test Scores Associated
with the INEM Program, Colombia, 1981 Cohort
Achievement            Number of
Curriculum       test     Score gain  students
Academic      Academic        3.28      1,545
Agricultural   Agriculture     -a        394
Academic        _ 
Commercial   Commercial       5.20      1,271
Academic        0.43
Industrial   Industrial      15.71       297
Academic        8.01
Social services  Social services  4.47   377
Academic        -a
Note: Background and ability factors are controlled for as in table 5-10. All score
gains are statistically significant at the 5 percent level or better.
a. No significant difference.
trial programs performed substantially better (by 15.71 points) than con-
trol students in the industrial achievement test and in the academic
achievement test (by 8.01 points). This implies that to do well in industrial
subjects by national standards, INEM students did not have to compromise
their achievement in academic subjects. The same is true of commercial
and social services students as well, though the advantage of the INEM
program is less marked. There were no significant differences between the
adjusted means of INEM and non-INEM agricultural students.
One may conclude that from a policy standpoint, the INEM vocational
program is clearly superior to non-INEM programs in imparting vocational
knowledge.
Effect of Course Time on Achievement
One would suppose that greater exposure to vocational curricula would
enhance students' vocational knowledge. INEM schools have offered more
hours of vocational course work per vocational student than have the
control schools. But within either the control or the INEM group, what
little variation could be found in number of hours spent on vocational
subjects seldom proved of much consequence for test performance in the
subdiscipline. Differences between control and INEM schools are far larger
than these differences within groups. As an example, the percentage of
course hours industrial arts students spent on their specialization in INEM
and non-INEM schools is tabulated in table 5-12.
In summary, the substantial differences between INEM and control



78    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
students in the proportion of course hours spent on vocational subjects
contributed heavily to differences in achievement scores on those subjects.
Within INEM or control schools, however, variation in course hours
by itself did not contribute significantly to variations in students' per-
formance.
Modernity Attitudes
The influence of schooling on the psychological disposition to adopt
modern attitudes has been examined by Inkeles and Smith (1974), Hol-
singer (1973), and others. If schooling effectively modernizes values and
roles, then it can aid in developing entrepreneurship and other talents
considered to be prerequisites for economic growth and development. It is
thought that investment in human capital can help to overcome many
characteristics of the labor force that impede greater productivity, such as
poor health, illiteracy, resistance to new knowledge, fear of change, and
immobility. It is through education that a society allows individuals to
fulfill their capacities and change their values, thus allowing them to
participate productively in the economic process. For example, Lewis
(1966, p. 109) remarked that "a developing economy needs to have at
least 50 percent of its children in primary school if traditional customs and
attitudes are to be changed significantly. The idea is simply that until a
large proportion of the community at a fairly young age is exposed to new
ideas and ways of doing things, few old ideas will be challenged, and
economic development will be retarded." Secondary schooling should
certainly enhance this process. Does diversification of secondary educa-
tion influence the acquisition of modern attitudes?
In the Colombian base-year survey, a series of questions taken from
Inkeles and Smith (1974) asked students to identify their attitudes and
Table 5-12. Time Spent on Industrial Courses and Achievement,
Colombia, 1981 Cohort
Mean industrial test score
Course hours spent
on industrial arts  Control    INEM
Less than 35 percent  52.4
35 to 40 percent     53.6        -
41 to 50 percent      -         56.6
More than 50 percent  -         58.8
- Not applicable.



THE COLOMBIA CASE STUDY                                                  79
values. With reference to hypothetical cases, students were asked how they
viewed various problems endemic to economic development. When we
compare response frequencies by item across INEM and non-INEM schools,
we find little difference across nine items hypothesized to capture disposi-
tions to modernity (table 5-13). However, when these nine items are coded
from low to high, based on orientation to modernity, and when the
responses are added to form a composite scale, we find a small but
significant difference in favor of the INEM group mean. This suggests that
INEM students are slightly more disposed to adopt modern attitudes than
are control students.
As was done for the previous achievement functions, the authors re-
gressed psychological modernity scores on dummy variables representing
curriculum programs, school type, family background, age, sex, and natu-
ral ability. The objective in this case was to observe which programs
within INEM and control schools and which characteristics of students'
backgrounds were associated with more modern attitudes. Again, peda-
gogy students were the omitted category and served as the base-line
comparator.
Inspection of table 5-14 shows several out-of-school characteristics to
Table 5-13. Responses to Selected Questions
on Modernity Attitudes, Colombia, 1981 Cohort
Percentage responding
affirmatively
Item                         INEM  Control
Education prepares one for work                      51     51
Lack of social organization responsible for a country's
underdevelopment                                   87      88
Reads newspaper regularly                            20     22
Personal economic situation will be much better
in five years                                      26     27
Hard work of Colombian people necessary for nation's
future success                                     79     77
Degree or specialization more important asset for job
performance than personal contacts                 66      65
Persons can be morally good without
having religious beliefs                           81     76
More interested in international news
than in local sports news                          31     33
Agrees with looking for alternatives to traditional
production methods                                 86      86



80   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
be associated with a modern outlook. Younger students, those with higher
verbal facility, females, those coming from homes with many books, those
with highly educated mothers, those with relatively higher family incomes,
and those whose families owned a business are more likely to be modern in
outlook.
When we observe the curriculum programs of students attending INEM
and control schools, only two tracks appear positively and significantly to
predict modernity scores-the commercial INEM track and the agricultural
Table 5-14. Modernity as a Function of Background
and School-Related Characteristics, Colombia, 1981 Cohort
Independent variable     Coefficient t-value
Intercept                        17.86
Out-of-school factors
Male                            - 0.465a    9.68
Age                             - 0.061a    4.35
Number of siblings                0.005    0.71
Attended private primary school    -0.022  0.39
Father's log income              0.178a    2.09
Family owns business             0.191a    3.67
Dictionary in home               0.195a    3.25
Father's education in yeaLrs     0.007     0.88
Mother's education in years      0.026a    3.25
Father clerical                 -0.057     1.10
Father professional              0.037     0.58
Books in home                    0.198a    8.25
Verbal aptitude                   0.035a   17.50
Math aptitude                    0.003      1.50
School type and curriculum
INEM
Academic                      -0.063     0.83
Agricultural                  -0.109      0.74
Commercial                     0.235a    3.13
Industrial                    -0.163      1.40
Social services                0.103     0.95
Control
Academic                      -0.103      1.58
Agricultural                    0.040     0.52
Commercial                     0.291a    2.64
Industrial                     0.025     0.20
Social services                0.141      1.09
R 2                                  0.075
Sample size                          7,813
a. Statistically significant at the 5 percent level or better.



THE COLOMBIA CASE STUDY                                         81
control track. The former group constitutes close to 10 percent of the
sample and explains why INEM students appear overall to be more modern
than control students. Although one might expect modernity to corre-
spond somewhat with entrepreneurial skills, which are produced by both
commercial and agricultural curricula, it is not clear why all students in
these curricula, regardless of school type, do not score equally well. Given
the above results and the fact that the internal reliability for the modernity
scale was low (of the order of 0.40), we cannot reach a firm conclusion that
diversification of the curriculum affects modernity attitudes one way or
another.
School Experience and Expectations
Students were asked about their experiences in school-for example,
how many times they repeated a grade in primary and secondary school,
whether they had completed all their secondary schooling at one institu-
tion, or whether they were satisfied with their current curriculum place-
ment (see table 5-15).
When we compare students attending INEM and control schools, we find
that those who have repeated primary school are just as likely to be found
in either type of school, more INEM students than control students had
graduated from a public primary school, and students from the two types
of school were equally satisfied with their course of study. Students who
had attended more than one secondary school were more likely to be
found in non-INEM schools, where some 48 percent of students had
switched schools at least once. Few INEM students had attended another
secondary school prior to enrolling in their current INEM school. (There is
a no-transfer policy in INEM schools.)
To measure their ambitions on completing secondary school, students
were asked to indicate the economic sector they expected to work in and
whether they expected to undertake further studies or training. Seventy
percent expect to study and work part-time; very few intend to be working
Table 5-15. Selected Student Characteristics by Type of School,
Colombia, 1981 Cohort
(percent)
Characteristic      INEM  Control
Attended a private primary school    11    25
Repeated a grade in primary school   28    25
Repeated a grade in secondary school  28    33
Satisfied with present course of study  86    86



82     DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
full-time; while 24 percent of students hope to be enrolled full-time in
university study (table 5-16). This suggests that the social demand for
somne form of continued schooling is extremely high. While INEM students
perceive themselves as significantly less likely to undertake full-time higher
education, three-fourths of them expect to be taking some course work
while working part-time.
More INEM than non-INEM students see themselves as likely to work
full-time than part-time immediately after graduating. This trend prob-
ably reflects the INEM students less privileged socioeconomic back-
grounds, which may lead to family pressure to earn outside income before
or during their postsecondary studies. Although small in absolute terms,
the difference between school types in this respect is statistically signifi-
cant.
Students clearly expected to be engaged in the sector for which they had
been trained in secondary school (table 5-17). Social services students were
an exception; only one-fifth of them expected to be working in the public
and personal service sector. Overall, 15 percent of the students were
undecided. INEM commercial stuclents are the most likely to continue in
their sector of study, and INEM social services and agricultural students are
the least likely to do so.
Table 5-16 indicated that 94 percent of all students wanted to continue
their education beyond secondary school. When asked what kind of
further studies they wished to purs ae, 93 percent stated that they wished to
continue to the university, with no overall difference between INEM and
control schools (table 5-18). Those studying social services in control
Table 5-16. Expected Student Destination by School Type
and Subject, Colombia, 1981 Cohort
(percent)
Work full-    Study full-    Study and
time          time           work           Other
Curriculum  INEM  Control  INEM  Control  INEM  Control  INEM  Control
Academic     2       2      35     44      62     53      1      1
Agricultural    9   10      20     16      70     74      1      0
Commercial    5      4       8     30      86     65      1      1
Industrial   5       4      19    25       74     70      2      1
Pedagogical   -      3      -      11      -      85     -       1
Social services  8   7      15     27      77     66      0      0
Overall      5       4      20     29      74     66      1      1
- Not applicable; INEM schools do not offer teacher training.



THE COLOMBIA CASE STUDY                                                  83
Table 5-17. Students Planning to Seek Employment Inside
or Outside Their Curriculum Specialization, by Type of School,
Colombia, 1981 Cohort
(percent)
Work            Work
inside field   outside field    Undecided
Curriculum   INEM   Control  INEM   Control  INEM   Control
Academic      61      60       14      15      25      25
Agricultural    48    52       31      24      21      24
Commercial    87      82        5       8       8      10
Industrial    67      63       25      24       8      13
Pedagogical    -      91       -        5      -        4
Social services  20   24       49      46      31      30
Average'      57      56       25      24      18      20
- Not applicable.
a. Average excludes pedagogy.
schools, however, were the least inclined to go to university.
Students were also asked to indicate which benefits they most wanted
from an occupation they might engage in. The most commonly cited were
social prestige and possibilities for further training and job advancement.
Most students, regardless of school or curriculum program, sought jobs
that offered the chance to obtain further general or specific training.
A question in the base-year survey asked respondents to indicate what
Table 5-18. Percentage of Students Planning to Enter University
by School Type and Curriculum, Colombia, 1981 Cohort
(percent)
Total
Curriculum   INEM  Control sample
Academic       95     94      95
Agricultural    93    90      91
Commercial    90      94      92
Industrial     92     94      93
Social services  92   79      87
Pedagogical    -      94      94
Overall       93      93      93
- Not applicable.
Note: Percentages refer to those who planned to continue their education beyond sec-
ondary school.



84     DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
earnings they expected to receive upon graduating from secondary school.
The response, a hypothetical statistic, reflects a student's perception of the
labor market conditions. Students were also asked what they thought they
would earn if they did not have a secondary school diploma. This hypo-
thetical statistic has been used in many studies to measure the self-assessed
opportunity cost of studying; the difference between the two estimates
represents the perceived monetary value of a secondary school diploma.
Comparisons by school type and curriculum program (table 5-19) show
that, in almost all instances, control students expect to earn much more
than their INEM counterparts. Generally, students in vocational subjects
expect to earn more than those pursuing an academic course. With no
secondary school diploma, all stuclents expect to earn less, but again, INEM
students expect to earn even less relative to the control group. Across
curriculum program and school type, the perceived difference in earnings
attributable to a secondary school diploma averages 28 percent, with INEM
commercial and control industrial students perceiving the largest differ-
ence (36 and 35 percent, respectively) and control commercial students
perceiving the smallest (21 percent).
When anticipated earnings five years after graduation are examined,
Table 5-19. Expected Earnings by School Type
and Subject, Colombia, 1981 Cohort
(monthly pesos)
Difference in
Expected    Expected Expected  earnings with
Type of        earnings    earnings,  earnings   diploma
school and   with secondary    no     in five
curriculum      diploma     diploma    years   Amount Percent
INEM
Academic          10,235      7,990    34,939    2,245     28
Agricultural      10,275      7,924    32,204    2,351     30
Commercial        10,686      7,868    34,114    2,818     36
Industrial        10,941      8,439    38,019    2,502     25
Social services    9,161      7,362    30,531    1,799     24
Average           10,440      7,974    34,666   2,466      31
Control
Academic          10,718      8,652    42,080   2,066      24
Agricultural      11,587      9,147    32,530    2,440     27
Commercial        10,681      8,830    36,953    1,851     21
Industrial        11,846      8,779    39,900    3,067     35
Pedagogical        9,546      7,616    31,706    1,930     25
Social services   10,972      8,164    36,592    2,808     34
Average           10,877      8,645    38,337   2,232      26
Overall average     10,700      8,374    36,854    2,326    28



THE COLOMBIA CASE STUDY                                     85
general academic students expected earnings slightly above the group
average, while agricultural, commercial, and social services students con-
tinued to view their earnings as considerably below the average figure.
Industrial trade students expected to earn 5 percent above the average
figure.
Comparing Costs with Outcomes
To make any judgments concerning the relationship between differ-
ences in cognitive achievements and the costs of producing them, one must
first look at the costs of the different curriculum programs (academic,
industrial, commercial, social services, and agricultural) offered by INEMS
and then compare these with the costs of the equivalent programs offered
by non-INEM schools. The data sources and assumptions for the cost
calculations are briefly described below. More details are to be found in
Hinchliffe (1983).
Unit Costs by Curriculum
The cost of schooling can be looked at from three viewpoints-costs to
society as a whole, to government, and to the individual (or household).
The cost components that may be relevant to one or more of the three are
(1) forgone output or earnings; (2) capital costs of buildings, furniture,
and equipment; and (3) direct recurrent costs. For calculations of social
costs, it is immaterial how (2) and (3) are financed. Costs to the govern-
ment or individual, however, obviously depend on who pays for these
expenditures.
Opportunity costs, as measured by the earnings of workers with pri-
mary education, were based on a study by Psacharopoulos (1983) that
used earnings data from a 1975 urban labor market survey in Colombia.
Earnings of a worker aged twenty-two with a primary education were
calculated to be 5,813 pesos a month at 1981 prices. Analysis of per-pupil
capital costs have been based on World Bank appraisal reports for the
INEM and other secondary education projects in Colombia. Recurrent
expenditures financed by households are available in some detail from the
student questionnaire. Data on school expenditures-for salaries, mainte-
nance, utilities, materials, and equipment-were based on the school
questionnaire and used to compare the costs of INEM and non-INEM
schools and curriculum programs.
Each INEM school offers five curriculum programs, and in many non-
INEM schools there are at least two. The costs per student for each
curriculum had to be estimated because INEM schools did not keep a



Table 5-20. Annual Cost per Student by School Type and Curriculum, Colombia, 1981 Cohort
(pesos)
Curriculum
School type and                                                Social  Overall
cost component     Academic  Agricultural Commercial Industrial services  average
INEM
Recurrent cost borne by
Household            3,300       3,300      2,800      2,800    2,600    2,900
Rest of socie.;    19,300      19,800     19,300      i9,400   19,300   19,400
°°\                             Total direct costa     25,700     26,200     25,200      25,300   25,000   25,400
Total social costb    95,400      95,900     94,900     95,000   94,700   95,100
Control'
Recurrent cost borne by
Household            4,300       4,000      3,800      4,400    3,600    4,300
Rest of society     15,100      25,200     16,600     23,000  21,400   22,900
Total direct costa    22,200      33,700     23,200     31,900  27,800   30,600
Total social cost"    91,900    103,400      92,900    101,600   97,500  100,300
a. Including capital costs and recurrent costs.
b. Including forgone earnings.
c. Public schools only.



THE COLOMBIA CASE STUDY                                       87
separate account for each program. To do this, each student in the sample
was hypothetically assigned to a curriculum program and given the aver-
age per-student cost of the school he was enrolled in. Then the average of
these costs for all students within each program was calculated. The
rationale for using this method was that any variations in average cost
among schools resulted from different combinations of curriculum pro-
grams within schools. For example, if one school had a higher proportion
of commercial students and a lower proportion of agricultural students
than another school and its average expenditure per pupil was lower, the
sole reason for the difference was assumed to be that expenditures for
commerce were less than those for agriculture.
This method led to an underestimation of the costs in agricultural and
industrial curricula and an overestimation of the costs in the other curric-
ula, especially within the INEM group. This is demonstrated by differentiat-
ing costs in the control schools, which teach fewer subjects than do INEM
schools and hence separate expense accounts can be distinguished. The
unit cost data appear in table 5-20, while selected cost items are converted
into indices by using the INEM overall average cost as the base in table
5-21.
The conclusions reached by examining recurrent costs were:
* Variations in costs between curriculum programs in INEMS are very
small.
* INEM academic and commercial programs are about 20 percent and
14 percent more expensive, respectively, than the control counter-
parts.
* INEM agricultural and industrial programs are significantly less expen-
sive than their control counterparts (28 percent and 25 percent,
respectively); the same is true for the social services specialty, but to a
lesser extent (11 percent).
When capital and recurrent costs are taken together, variations among
INEMS are very slight. There are, however, wide differences between the
INEM average cost and costs of single- and multi-track control schools,
with the latters' agricultural and industrial specialties being one-third and
one-quarter more expensive, respectively, than the average INEM cost.
The total social costs of students in all programs in both types of school
(INEM and non-INEM) are dominated by earnings forgone, which are
assumed to be equal for all students. Differences in total costs lie within a
range of only 12 percentage points.
Differences in total social costs between school types and between
academic and prevocational programs are small. For capital and recurrent
costs alone, however, there are significant differences among non-INEM
programs and between these and similar programs in INEM schools.



Table 5-21. Unit Cost Index by School Type and Curriculum, Colombia, 1981 Cohort
(INEM average= 100)
Social
Cost component    Academic  Agricultural Commercial Industrial services  Overall
INEM
Capital and
recurrent costs      100        103           99         100        98      100
°°                                 Social cost
(including forgone
earnings)            100        101          100         100        99      100
Control
Capital and
recurrent costs       87        133           91         126       109      120
Social cost
(including forgone
earnings)             97        109           98         105       102      105



THE COLOMBIA CASE STUDY                                       89
Cost-Effectiveness of Cognitive Achievement
To relate costs to achievement performance, three pieces of information
were examined: annual unit costs (capital and recurrent) of keeping a
student in secondary school, group test-score means in academic achieve-
ment, and the test-score mean in each vocational specialization (table
5-22). Scores were adjusted for innate ability and out-of-school character-
istics.
There are two relevant sets of cost and achievement comparisons. The
first compares students studying the same subject but attending different
types of school; the second compares the academic control group with
students who have taken prevocational subjects.
First, when the costs of academic programs are compared, INEM schools
require expenditures of 25,700 pesos per student year, while control
schools require only 22,200 pesos per student year. The difference of
3,500 pesos (16 percent) can be interpreted as the average amount of
additional expenditure per student year required to raise the academic
achievement score five points, or 0.3 of one standard deviation, for INEM
academic students. This five-point difference, applied to the percentage of
INEM students who outscore control students on academic tests of achieve-
ment, is 12 percent. (Note that because test scores are standardized to a
mean of 50 and a standard deviation of 10, a ten-point difference would
encompass approximately one-third of the students in a particular group;
see figure 5-5.) In short, 12 percent more INEM students than non-INEM
students score above the mean, but 16 percent more resources are required
per student.
Second, the net cost and achievement differences between INEM and
control schools for the various curriculum programs are shown in table
5-23. The INEM industrial program costs 21 percent less per student year
than its control group counterpart; yet INEM students score about 1.5
standard deviations, or 43 percent, above control students in industrial
tests. On academic tests, INEM students score 0.8 standard deviations, or
29 percent, higher than non-INEM students. Thus, although it costs less,
INEM industrial education raises both academic and vocational achieve-
ment levels above those in industrial education in control schools. Like-
wise, INEM agricultural students cost 22 percent less per year to educate,
yet 19 percent of them score higher than control agricultural students in
the agricultural tests. Only 14 percent, however, score higher than the
mean on academic testing.
Social service education costs 10 percent less in INEM than in non-INEM
schools, while 16 percent more INEM students than non-INEM students
have higher than average scores the mean on the social service specializa-



Table 5-22. Annual Cost per Student and Achievement by School Type and Curriculum,
Colombia, 1981 Cohort
Adjusted
Adjusted             vocational achievement
academic
School type      Unit cost'  achieve-                                      Social
and curriculum     (pesos)    mentb   Agricultural Commercial Industrial services
Academic INEM         25,700       53
Acadevmc con=.rol     22,200       50          -
Agricultural INEM     26,200       49          62          -           -         -
Agricultural control    33,700     48          58          -           -         -
Commercial INEM       25,200       49          -           61
Commercial control    23,200       49          -           56          -
Industrial INEM       25,300       55          -           -           62        -
Industrial control    31,900       47          -           -           46        -
Social services INEM  25,000       48          -           -           -         57
Social services control   27,800   48          -           -           -         53
-Not applicable.
a. Cost refers to public schools only and includes school-related expenses by students, government-financed recurrent costs, and annualized capital
costs. It excludes forgone earnings.
b. Achievement scores are adjusted for out-of-school characteristics and ability (see table 5-10).



THE COLOMBIA CASE STUDY                                            91
tion test. Social service students in INEM do not, however, score better on
academic tests than social service students in control schools. The pattern
is similar for the INEM commercial program. It costs 9 percent more pesos
per year to educate INEM students, and for that amount 19 percent more
INEM students than control students score above the mean on the commer-
cial specialization test.
In summary, INEM students come from slightly less favored socioeco-
nomic backgrounds and score lower than other students on tests of general
ability, but once corrections for out-of-school factors have been made,
their vocational test achievement scores are higher on average than those
of the non-INEM control group. Moreover, INEM students in academic,
industrial, and agricultural programs also score higher on academic tests
than control school students.
Summary of Base-Year Results
Having searched for both cognitive and noncognitive school-related
differences across types of school and curriculum programs, we have come
to a guarded conclusion about which schools are doing better than others.
In academic achievement, INEM academic, industrial, and agricultural
students do better than their control counterparts. In their vocational
achievement scores, students in a given specialization, regardless of the
type of school they attend, always dramatically outperform students not in
that specialization.
INEM schools have more modern equipment than non-INEM schools.
They appear to be achieving their goal of raising the level of students'
vocational ability, while slightly increasing the level of their academic
knowledge relative to non-INEM students in most curriculum streams.
Table 5-23. Differences between INEM and Control Group
in Cost and Achievement, Colombia, 1981 Cohort
iNEM incremental
INEM        acbievement
incremental   (test score points)
cost
Curriculum    (percent)   Academic  Vocational
Academic       +16        +3          -
Agricultural   - 22       + 1       + 4
Commercial      + 9         0       + 5
Industrial     -21        +8        +16
Social services    -10      0       + 4
- Not applicable.
Source: Based on table 5-22.



92    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Thus, INEM schools have the perhaps unintended, but welcome, effect of
fostering academic knowledge more than their academic counterparts.
With regard to socioeconomic background and noncognitive outcomes,
the average INEM student comes from a home where the father has less
education and earns less than the fathers of students attending control
schools. INEM students are, on the average, older than their counterparts,
less likely to have attended a private primary school, more likely to have
repeated a grade in primary school, and less likely to want to study
full-time after secondary school, although they may wish to attend a
university just as much as the non-INEM students. The INEM student is as
satisfied with his or her course of secondary instruction as his fellow
students from other schools. But INEM students anticipate earning less
than their counterparts in every field. This appears to be true even after
differences in ability and socioeconomic background have been taken into
account through statistical controls.
Based on comparisons of costs and achievement gains in academic and
vocational knowledge between INEM and control schools, INEM industrial,
social service, and agricultural streams are substantially less expensive
than their control counterparts. Combined with the fact that these pro-
grams substantially boost achievement scores, they are unquestionably
successful. And although the INEMI academic and commercial programs
cost more than their control counterparts, they also substantially boost
achievement.
Given the above findings, we conclude that INEM schools are certainly
no less effective than non-INEM schools in imparting desired academic and
vocational skills. The fact that INEiv schools seem to teach vocational skills
more effectively than academic skills is certainly consistent with their
supposed function and mandate. Several questions remain, however, a
major one being how INEM graduates fare in the labor market. We now
turn our attention to the results of the follow-up survey in order to answer
that question.
The Class of 1981 One Year Later
In 1982, approximately one year after receiving their diploma, the 1981
graduating class was again surveyed. The questionnaire elicited informa-
tion on the graduates' current status; that is, whether they were studying
full-time, working full-time, studying and working part-time, or other.
Information was also obtained on postsecondary school experiences in the
labor market, further schooling, training, and job search. Sector of em-
ployment, hours worked, and earnings were obtained from among those
working. Those searching for a job were asked the minimum salary they



THE COLOMBIA CASE STUDY                                                93
would accept, while those in full-time education were asked what kind of
curriculum they were pursuing. The follow-up survey aimed to assess
whether postschool outcomes were different for students coming from the
various curriculum programs pursued in INEM or control schools.
Before charting the entry and flow of students through postsecondary
school activities, it is necessary to note the proportion of the original
sample of 8,051 students who responded to the later questionnaire. The
results reported below are based on a 62 percent response rate-slightly
above the 60 percent minimum required by the sample design to yield
school type curriculum cells containing enough observations for differ-
ences in outcomes to be statistically significant. (Efforts were continuing in
1983 to trace additional graduates of the original cohort; any responses
from them will eventually be merged with the existing data base.) As
shown in table 5-24, the respondents closely reflect the proportions of
students by school type and curriculum track appearing in the base-year
sample.
Graduate Destinations
Of those students who responded, 37 percent were studying full-time,
30 percent were working full-time, 11 percent were both studying and
Table 5-24. Response to Activity Survey One Year
after Graduation by School Type and Curriculum,
Colombia, 1981 Cohort
1981 base year    1982 follow-up
Response
School type    Number          Number             rate
and curriculum  surveyed  Percent surveyed  Percent (percent)
INEM
Academic        871     10.8     514      10.3     S9
Agricultural    174      2.2      78       1.6    45
Commercial      957     11.9     588     11.8     61
Industrial      869     10.8     474      9.5     55
Social services    380   4.7     225      4.5     59
Control
Academic      1,471     18.3     934     18.7     60
Agricultural    370      4.6     243      4.9      65
Commercial      973     12.1     557     11.2     59
Industrial      945     11.7     695      13.9     74
Pedagogical     767      9.5     559     11.2     72
Social services    274   3.4     124      2.5     56
Overall         8,051    100.0   4,991    100.0     62



94     DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
working part-time, and 22 percent were in other activities (including
nonparticipants staying at home or searching for a job or an opportunity
for further training). The latter group includes many women and gradu-
ates who had entered the military. Table 5-25 gives the main activity
destinations by school type and curriculum program. Thus, one year later,
more than one-third of the graduates were in school, just under one-third
were working full-time, and aside from 11 percent who were studying and
working part-time simultaneously, about one-quarter of the sample was
neither studying nor working.
There appear to be no major differences in postschool activity between
the groups of INEM and non-INEM respondents, meaning that the type of
school attended does not alter the initial decision of graduates concerning
their academic or vocational pursuits (figure 5-7).
Table 5-26 compares students' expectations for postschool activities
before leaving school in 1981 with their actual destination one year later.
There was an increase of 14 percentage points in those who actually
attended school full-time; this increase was greatest among INEM students
(19 points compared with 11 points for control students). The number of
graduates in full-time employment is higher by 26 percentage points than
Table 5-25. 1982 Destination by School Type and Curriculum,
Colombia, 1981 Cohort
(percent)
Postschool activity, 1982
School type and             Study
curriculum, 1981  Study Work and work  Other
INEM
Academic       34    30       9      27
Agricultural   29    27      15      29
Commercial    39    29       10      22
Industrial     36    31       8      25
Social services   43    26    9      23
Average        37    29      10      24
Control
Academic       38    29      13      20
Agricultural   34    36      11      19
Commercial     36    32       9      23
Industrial     34    32      11      23
Pedagogical    41    29      10      20
Social services   46    26   12      16
Average        37    30      11      22
Overall         37    30       11      22



THE COLOMBIA CASE STUDY                                                95
Figure 5-7. Distribution of Respondents by 1982 Activity,
Colombia, 1981 Cohort
INEM                                 Control
Wor        and workWfk                             andwr
full-time   10%/      \                full-time    an%
29%              Job                   30%      /ob
search                                  search
11%                                     9%
Study         13%                       Study         13%
\   full-time    \        /       \         ftill-time
the number who one year earlier had envisaged themselves as working. On
this latter statistic, no significant difference between INEM and non-INEM
graduates is apparent. Thus, although when polled in 1981 INEM students
had felt less likely than others to continue with full-time education, in
practice they continued in roughly the same proportion as non-INEM
respondents.
To what extent does socioeconomic background relate to postsecond-
ary school activities? As table 5-27 shows, graduates who study full-time
come from families with significantly higher incomes than do graduates
staying at home, studying part-time, or working full-time. This suggests
that the family's well-being determines in part whether a student will study
Table 5-26. Planned (1981) versus Actual (1982) Destinations
of Graduates, Colombia, 1981 Cohort
(percent)
INEM                Control               Overall
Plan  Actual Differ-  Plan  Actual Differ-  Plan  Actual Differ-
Activity   1981  1982   ence   1981  1982   ence   1981  1982   ence
Study      18.0  37.0   +19.0  26.0  37.0  +11.0  23.0  37.0  +14.0
Study and
work     77.0  10.0  -67.0  70.0  11.0  -59.0   73.0  11.0  -62.0
Work        4.5  29.0   +24.5   3.5  31.0  +28.0   3.7  30.0   +26.3
Other       0.5  24.0   + 23.5   0.5  20.5   +20.0   0.3  22.0   +21.4



96     DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-27. Father's Mean Incomte and Educational Level
by 1982 Destination and School Type, Colombia, 1981 Cohort
Father's      Father's
Destination and monthly income educational level
school type     (pesos)        (Years)
Study
INEM             23,198          7.0
Control          23,526          7.7
Study and work
INEM             ]8,745          7.2
Control          20,738          7.8
Work
INEM             18,226          7.0
Control          18,778          7.4
None of the above
INEM             18,242          6.8
Control          17,957          7.4
Overall            20,346          7.4
full-time. In the following section we divide the 1982 respondents accord-
ing to their postschool destinations in order to assess possible differences
in experience by type of school or curriculum program.
Full-Time Students
One-third of the sample (1,665 respondents) were studying full-time in
1982. In the 1981 base-year survey, more than 90 percent of the secondary
school students planning to continue their education wished to take
university courses. In 1982, however, only 58 percent of those who were
enrolled full-time actually were taking university-accredited courses.
Table 5-28 gives a breakdown of the type of institution attended in 1982
by type of secondary school attended and type of curriculum (academic or
vocational) followed in 1981. When all INEM graduates are combined,
regardless of subject studied, no clifference between control and INEM
university attendance rates can be discerned. Of the INEM graduates
enrolled in university courses, however, more had followed an academic
than a vocational program.
A number of questions asked graduates' opinions regarding their cur-
rent studies. Seventy-five percent of zhe INEM graduates reported they were
satisfied, compared with 77 percent of the control group. From both types
of schools, 62 percent of the respondents believed their secondary school
course of study was extremely helpful in their present field of study, and
only 18 percent claimed to be dissatisfied with their preparatory studies.



THE COLOMBIA CASE STUDY                                           97
Table 5-28. Type of Postsecondary Institution Attended
by Those Studying in 1982, Colombia, 1981 Cohort
(percent)
Postsecondary institution, 1982
School type and
curriculum, 1981  University  Nonuniversity  Nonaccredited
INEM
Academic       67         13           20
Vocational     49         23           28
Average        55         20           25
Control
Academic       61         19           20
Vocational     59         21           20
Average        60         20           20
Overall average    58       20           22
On this subject, there were no significant differences in the opinions of the
INEM and control groups.
Postsecondary training programs then were grouped into humanities-
related subjects, commercial and technical areas, and a residual category.
The distribution of graduates by school curriculum program across these
postsecondary training programs is shown in table 5-29. Commercial
courses dominate postsecondary training, and the proportion of agricul-
ture graduates switching to commercial and technical courses is surpris-
ingly high.
Graduates were asked what they could expect to earn if they were
currently employed rather than studying, and what they hoped to earn
upon completing their present education. The difference between the two
values gives an indication of the perceived value of postsecondary school
studies, based on revised expectations. The results for INEM and control
groups in academic and vocational studies are presented in table 5-30. The
self-assessed forgone earnings are generally consistent with students' ex-
pectations one year earlier (see table 5-19). Both INEM and other sec-
ondary graduates expect to more than triple their monthly earnings by
undertaking further study. INEM graduates expect a slightly smaller gain
than do control-school graduates.
One-fourth of the follow-up respondents were currently attending or
had previously been in some sort of training program; there were no major
differences in this respect between the INEM and control groups. Of those
who opted for some type of training, 67 percent had taken at least one
training course lasting an average of twenty-one weeks. Forty-three per-
cent of these training courses were in business administration, which



98     DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-29. Distribution of Students in 1982 Training Course,
Colombia, 1981 Cohort
(percent)
1981                 1982 training course
School type
and curriculum Humanities Commercial Technical Other
INEM
Academic        37          37         16      10
Agricultural    26          32        32       10
Commercial      28          39         18      15
Industrial      25          41        21       13
Social services  31         39         15      15
Control
Academic        26          42         18      14
Agricultural     17         34         28      21
Commercial       18         39         20      23
Industrial      27          46         12      15
Pedagogical     31          30        21       18
Social services  25         40        20       15
Overall           27          40        18       15
attracted INEM and non-INEM graduates equally. Fifty percent of all train-
ing occurred in technical institutes, another 25 percent in Servicio Na-
cional de Aprendizaje (SENA, the national apprenticeship service), 10
percent in firms, and only 5 percent in a university-sponsored environ-
ment. Again, equal proportions of INEM and control-group respondents
were engaged in these kinds of training.
One-quarter of control graduates and 20 percent of INEM graduates fit
into the general academic training category that combines arts, humani-
ties, social sciences, and science education. The greater percentage of
control-group graduates training in academic pursuits is the only signifi-
cant difference found between INEM and non-INEM respondents who
received training after secondary school graduation.
Full-Time Workers
The 1982 follow-up survey reccrded 30 percent of the 1981 cohort
(1,397 respondents) as working full-time. Do these graduates differ in
their attitudes, experiences, and socioeconomic characteristics according
to the type of secondary school they attended?
As shown in table 5-3 1, the profile for INEM and control respondents is
very similar in two respects: their fathers' employment status and the
proportion who cited poor academic records as the reason for not con-



THE COLOMBIA CASE STUDY                                                  99
Table 5-30. Expected Earnings With and Without Postsecondary
Studies, 1982 Follow-up Response, Colombia, 1981 Cohort
Pesos per month
Increment
With        With        due to post-
School type   secondary  postsecondary  secondary study
and curriculum    diploma    study        (percent)
INEM
Academic       10,282      35,814          248
Agricultural    11,916     29,475          147
Commercial    10,505       31,014          195
Industrial     10,934      33,971          211
Social services   11,754   32,594          211
Average        10,760      33,198          213
Control
Academic       10,950      37,697          244
Agricultural    10,065     32,653          244
Commercial    10,322       38,454          272
Industrial     11,043      35,203          219
Pedagogical    10,790      35,033          247
Social services   12,965   35,195          171
Average        10,841      36,303          241
Overall average   10,818     35,106          245
tinuing education. Those in full-time employment cited lack of financial
resources as the reason for not continuing education more often than did
students staying at home or those working and studying part-time.
One question asked students how they acquired their first job. One out
of four found their job through independent means, such as a newspaper
or the employment exchange. But the majority (75 percent) obtained their
Table 5-31. Selected Characteristics of Full-Time Workers
in the 1982 Follow-up, Colombia, 1981 Cohort
(percent)
Characteristic                 INEM  Control
Father is employer                             28     29
Found job through friends, acquaintances       69     82
Working in the private sector                  86     88
Financial reasons for not continuing study     37     23
Poor exams reason for not continuing study    26      23
Secondary school study very useful for current job    53  55
Should have studied more mathematics           15     20
Dissatisfied with current work                30      32



100   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
current jobs through relatives, friends, and personal acquaintances, which
suggests that gainful employment in Colombia is still heavily influenced by
family connections. Because more non-INEM than INEM graduates
obtained employment in this fashion, it is possible that higher socioeco-
nomic status and private school links help graduates land a job.
Table 5-32 indicates the probability of students' being employed one
year after graduation. Those coming from control vocational schools are
the most likely to be working one year later. Another way of identifying
employment differences among students is to consider sectors of employ-
ment, as shown in table 5-33. Again, there are no major differences
between types of school and curriculum, but INEM graduates are more
likely to work in services.
The wide range of occupations entered into by graduates has been
narrowed into six groupings. The distribution of students by school type
and academic or vocational program across these groupings is shown in
table 5-34. More than 60 percent of those with a full-time job in 1982 were
engaged as secretaries, cashiers, bookkeepers, and salespersons. Inter-
estingly, graduates who had followed an academic curriculum, regardless
of INEM or non-INEM distinction, were just as likely to be employed in
these categories as were those trained in vocational programs.
Table 5-35 gives the mean earnings of INEM and control graduates
working full-time. Although differences in earnings are very modest,
academic graduates from both types of schools earn more than the others.
With the exception of INEM acactemic graduates, actual earnings are
clearly less than what graduates were hoping to earn before entering the
labor market (see, for example, tables 5-19 and 5-30).
Do those coming from INEM or vocational curricula experience have a
shorter period of unemployment before landing a job? The data show, if
anything, that those who have followed vocational courses in either INEM
or control schools have significantly longer periods of unemployment
Table 5-32. Percentage Working One Year after Graduation,
by School Type and 1981 Program, Colombia, 1981 Cohort
(percent)
School type    Academic Vocational
INEM
Full-time work   29.6  28.1
Work and study   39.1  39.3
Control
Full-time work   28.7  33.7
Work and study   42.2  43.9



THE COLOMBIA CASE STUDY                                                 101
Table 5-33. Employment of Secondary School Graduates
by Economic Sector, Colombia, 1981 Cohort
(percent)
INEM                Non-INEM
1982
economic sector  Academic  Vocational  Academic  Vocational Overall
Agricultural        2.6        2.3         1.5        2.1      2.0
Commercialb        40.5       40.8       41.8       46.0      43.0
Industrial'        19.8       17.6       18.6        15.9     17.0
Public utilities'   8.6       11.1       16.0        13.4     13.0
Servicesd          28.4       28.1       22.2       22.6      25.0
Total             100.0      100.0      100.0       100.0    100.0
a. The industrial sector includes construction, mining, and manufacturing.
b. The commercial sector includes finance.
c. The public utilities sector includes transportation.
d. Services include teaching.
(twenty-six weeks, as opposed to about twenty-one weeks for academic
graduates of both types of school).
Part-Time Students and Workers
Ten percent of the 1982 respondents (500 individuals) identified them-
selves as working and studying at the time of the survey. Table 5-36 shows
selected characteristics of those students who work and attend school
simultaneously. No significant differences in the factors surveyed are
visible between graduates from different types of school or curriculum.
Essentially, average current earnings (8,655 pesos) and hours per week
Table 5-34. Occupational Distribution of Graduates
by School Type and Program, 1982 Follow-up,
Colombia, 1981 Cohort
(percent)
INEM                Non-iNEM
1982
occupation    Academic  Vocational  Academic  Vocational
Agricultural          4.8        4.8        7.5        3.7
Blue-collar          23.8       23.0       24.7       20.3
Sales                17.1       23.0       25.8       21.8
Secretarial          15.2       11.0       10.8       18.9
Teaching             18.1       18.2        8.1       12.7
Other white-collar  21.0        20.0       23.1       22.6
Total               100.0      100.0      100.0      100.0



102   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-35. Monthly Earnings by School Type and Curriculum,
1982 Follow-up, Colombia, 1981 Cohort
(pesos)
1982 earnings
Curriculum   INEM   Control  Sample
Academic    10,539  10,070   10,207
Agricultural    9,520   9,596    9,556
Commercial    9,664   9,322    9,493
Industrial   9,408  10,291    9,902
Pedagogical    --    10,426a  10,426a
Social services  9,556   9,577    9,568
Average       9,354   9,980    9,887
- Not applicable.
a. Not included in the average.
worked (thirty-six) indicate these graduates work close to full-time. Forty-
eight percent on average had some training (23 percent in SENA, 60 percent
with technical institutes, and 7 percent with firms). Almost one-fourth
were looking for another type of work.
Graduates Looking for Work
Thirty-one percent of all graduates reported in the 1982 follow-up that
they were currently looking for work. They included 23 percent of those
studying, 26 percent of those working, 18 percent of those studying and
working, and 58 percent of those neither studying nor working.
Clearly, those in full- or part-time study or employment cannot be
classified as unemployed. Twenty-six percent of those already employed
full-time were looking for other employment, and of those neither study-
ing nor working 58 percent were actively seeking work and the rest were
not participants in the labor force. Looking at the schools from which the
graduates come, we note that 30 percent of the control graduates and
34 percent of INEM graduates were looking for work. The percentage
difference in this case is statistically significant (P=0.001) and suggests
that INEM graduates are more likely to be actively looking for work than
their control-school counterparts.
Of those graduates seeking jobs, 85 percent were still financially depen-
dent on their parents or relatives, with no difference between INEM and
non-INEM affiliation. On average, the job-seekers had been looking unsuc-
cessfully for work for the previous twenty-two weeks, and had a reserva-
tion wage (minimum acceptable sa lary) of 12,000 pesos per month, much
higher than the wages of those already employed (table 5-37).



THE COLOMBIA CASE STUDY                                               103
Table 5-36. Selected Characteristics of Part-Time Students
and Workers in 1982, Colombia, 1981 Cohort
With    Looking
Hours           post-      for
Current worked  Weeks secondary   other
School type   earnings   per   to find  training    work
and curriculum   (pesos)   week   job   (percent)  (percent)
INEM
Academic      9,355     34     6        39        15
Agricultural    9,488   44     4        40        20
Commercial    8,350     39     6        46        19
Industrial    8,338     39     5        55        22
Social services  8,144  30     5        56        33
Control
Academic      9,503     37     6        43        21
Agricultural    6,595   32     7        40         5
Commercial    8,398     36     6        56        22
Industrial    8,164     35     5        62        30
Social services  9,425  39     2        27        27
Overall         8,655     36     5.4      48        23
The kinds of jobs the respondents sought varied substantially, but the
top five occupations are shown in table 5-3 8, further broken down by type
of school and program. Major differences in job preferences of INEM and
non-INEM job-seekers arise in secretarial, sales, and other office occupa-
tions. More control-group graduates than INEM graduates are looking for
secretarial work, and these job-seekers come just as often from agricul-
tural and industrial programs as from commercial ones. Those seeking
jobs in sales are more likely to be INEM graduates (64 percent versus 36
percent) coming mostly from vocational programs.
In summary, while the proportions of INEM and control graduates
looking for work are much the same, and while both groups expect to earn
similar salaries once a job is found, a few differences can be discerned
regarding the kind of employment sought by individuals from the two
groups. Non-INEM job-seekers are relatively more interested in secretarial
employment, while INEM graduates tend to gravitate toward sales and
office management positions.
Summary of Findings
Descriptive profiles for full-time students, workers, job-seekers, and
those enrolled in training programs reveal few major differences between
INEM and control-school graduates.



104  DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-37. Incidence and Duration of Job Search
and Reservation Wage among Job-Seekers in 1982,
Colombia, 1981 Cohort
Duration of search   Reservation wage'
Percentage searching       (weeks)          (pesos a month)
Including   Exclud-   Including   Exclud-   Including   Exclud-
part-time    ing     part-time    ing      part-time    ing
School     workers   workers   workers   workers    workers   workers
type and      and        and       and        and        and        and
curriculum    students'  studentsb   students   students   students   students
INEM
Academic        31        13        21.0       25.0      12,343     11,683
Agricultural    41        16        20.1       23.0      11,552      9,932
Commercial      33        10        18.8       25.1      11,691     10,936
Industrial      34        12        20.1       24.1      12,358     11,414
Social
services      32         9        22.5       32.4      12,197     11,647
Average         32        11        20.0       25.2      12,121     11,308
Control
Academic       28          9        19.7       24.2      12,370     11,735
Agricultural    30         9        18.6       29.1      12,170     10,790
Commercial     30          7        19.1       28.4      11,998     12,502
Industrial     31         10        18.7       23.0      12,725     11,226
Pedagogical    30          9        16.2       24.0      13,243     12,572
Social
services      27         9        17.0       28.1      11,759     12,800
Average        29          9        19.3       25.0      12,517     11,791
Overall          30         10        19.7       25.1      12,331     11,575
a.    (Number of graduates answering "Looking now")    x 100
(Total number in 1982 cell for school type and curriculum)
b. Same as above, but numerator excludes those working or studying.
c. Minimum acceptable salary.
Because both groups responded in proportion to their membership in
the original base-year sample, the findings imply that among secondary
school leavers in Colombia one year after graduation, roughly one-third
go on to advanced schooling, one-t:hird opt for the labor market, and the
remaining third either work and study part-time (10 percent of that group)
are seeking work (11 percent), do not participate in the labor market
(7 percent), or are not interested in further training (5 percent). This
breakdown applies to both INEM and control graduates.
It thus appears that the pattern Cf experiences within the first year after
graduation is fairly universal for Colombian high school graduates, re-
gardless of type of school attended. The implication is that differences in



THE COLOMBIA CASE STUDY                                          105
secondary schools have little effect on postsecondary school opportunities
within the time frame studied. If students entering INEM and non-INEM
schools had radically different capabilities and backgrounds to begin with,
one might conclude that INEMS play a significant role in equalizing oppor-
tunities for very different groups of youth. However, the analyses of
base-year data reported earlier in this chapter revealed no major distinc-
tions between INEM and control students, except that the former had
somewhat lower socioeconomic status and ability.
Explaining the Decision to Work or Study
To predict whether a respondent will be employed full-time (or, alterna-
tively, study full-time), we construct a dummy variable having a value of
one if working full-time (or studying full-time), and zero if not. Thus the
dependent variable is interpreted as the probability of working (or
studying).
Because the dependent variable is dichotomous, probit analysis is used
instead of multiple linear regression analysis. The estimates can be inter-
preted in the same way as ordinary regression estimates: the sign of the
coefficient indicates a positive or negative effect of the corresponding
independent variable, and the t-value indicates whether the effect is statis-
tically significant. But unlike in multiple regression, the magnitude of the
effect depends on the values of all the other independent variables, because
the model assumes there is a nonlinear functional relationship between the
dependent and independent variables. Hence the "marginal effect" of one
independent variable-that is, the effect on the dependent variable of a
small change in the independent variable-can be estimated by setting all
other variables at selected values or, as in this study, at their means (see
Moock and Leslie, forthcoming).
Findings from this type of nonlinear regression analysis can be used to
Table 5-38. Main Target Occupations for 1982 Job-Seekers,
Colombia, 1981 Cohort
(percent)
INEM                Control
Occupation    Academic  Vocational  Academic  Vocational  Total
Accountant        12         37         12        39       100
Office messenger  18         40         13        29       100
Salesperson       24         40         10        26       100
Secretary          9         28         15        48       100
Teacher           12         39         11        38       100



Table 5-39. Marginal Effect of Variables on the Probability
of Studying Full-Time One Year after Secondary School,
Colombia, 1981 Cohort
Independent            Marginal
variable              effect   t-value
Out-of-school factors
Male                          0.004     0.41
Age                          - 0.042     0.55
Family owns business         -0.001     0.25
Books owned by household      0.004     0.45
Father's education            0.017      1.21
Family income                -0.001     0.16
Father blue-collar            0.002     0.89
Father nonmanual employee    - 0.008     1.23
Father is employer           -0.002      0.60
Verbal aptitude              -0.046      1.38
Math aptitude                - 0.024    0.73
School factorsa
INEM
Academic                   -0.001      0.30
Agricultural               - 0.001     0.67
Commercial                  0.002      0.61
Industrial                  0.0002    0.06
Social services             0.002      1.39
Control
Academic                    0.007      1.39
Agricultural               - 0.0002    0.61
Commercial                  0.003      0.67
Industrial                 - 0.003     0.67
Social services             0.002b    2.06
Test score
Academic                    0.103b    2.15
Agricultural                0.058      1.37
Commercial                  0.019     0.41
Pedagogical'               -0.016      0.41
Social services            -0.100"    2.09
Labor market factors
Labor force participation rate  0.252b    1.97
Per capita income            -0.156b    3.08
Constant                     -0.242      1.49
D-value                            0.020
Sample size                        2,666
Chi square                         43.69
a. Curriculum program categories are evaluated in relation to the pedagogical program
(the omitted category).
b. Statistically significant at the S percen: level or better. Maximum likelihood estima-
tion by the probit procedure of the Choice I'rogram, courtesy of Eric Swanson, the World
Bank.
c. Pedagogical score relates only to contrna students; INEM schools do not offer teacher
training.
106



THE COLOMBIA CASE STUDY                                     107
determine whether or not secondary school students' vocational and
academic achievement performance or their curriculum programs actually
influence their ability to secure employment over and above the impact of
family background, sex, age, innate ability, and other control factors.
Independent variables in the probit analysis included many of the out-of-
school influences discussed previously. In addition, several variables tap-
ping the regional level of per capita income and participation in the labor
force were included, so as to control for geographic differences in eco-
nomic activity that might affect an individual's ability to find employment.
Finally, measures of test scores, both vocational and academic, were
included in the analysis. Cognitive scores are important because they
provide an indication of the degree to which what is learned in school
influences ability to find a job or to continue with education.
Table 5-39 shows that only a few of the variables significantly affect the
probability that a student will continue with full-time study. The age and
sex of the respondent have no effect on the decision to study full-time.
Academic test score was the single most powerful school-related predictor
of full-time study. According to the "marginal effect" of the academic test
score, other things being equal, each additional point scored on the
academic achievement test (near the mean value) is associated with a 10
percent increase in the probability that an individual will study full-time
after secondary school. This should not be surprising, because postsecond-
ary enrollments depend primarily on evidence of academic merit.
The survey also found that graduates who have done well in social
studies tend to continue with full-time study. Labor market factors are
also good predictors of students' decision to study full-time. In particular,
the higher the local per capita income, the lower the likelihood students
will continue education beyond secondary school. This is not unexpected
because relatively high postsecondary earnings would imply higher for-
gone earnings for those who decide to continue studying, thus providing a
disincentive to enter the university.
The results of a similar probit procedure attempting to isolate the
factors associated with whether a student will be employed full-time one
year after graduation are reported in table 5-40. No single variable is
statistically related to the probability of full-time employment after grad-
uation.
Relation between School Type, Cognitive Skills, and Earnings
A path model was fitted to postsecondary students working full-time in
1982. Father's education in years and a composite score representing
verbal and nonverbal aptitude were treated as independent of school type
(INEM or control); two alternative achievement scores (academic and



Table 5-40. Marginal Effect of Variables on the Probability
of Working Full-Time One Year after Secondary Graduation,
Colombia, 1981 Cohort
Independent           Marginal
variable              effect   t-value
Out-of-school factors
Male                          0.002    0.19
Age                           0.105    1.22
Family owns business         -0.004    0.98
Books owned by household     -0.021    1.83
Father's education           - 0.005    0.32
Family income                 0.007    0.94
Father blue-collar            0.002    0.68
Father nonmanual employee     0.009     1.20
Father is employer            0.003     1.02
Verbal aptitude              - 0.042    1.09
Math aptitude                - 0.032    0.86
School factorsa
INEM
Academic                     0.001    0.04
Agricultural               - 0.001    0.50
Commercial                 - 0.002    0.53
Industrial                   0.001    0.29
Social services            -0.002     1.07
Control
Academic                   - 0.006    1.08
Agricultural                 0.003    1.01
Commercial                   0.001    0.14
Industrial                  0.004    0.73
Social services             0.001    0.40
Test score
Academic                   - 0.064    1.18
Agricultural                 0.006    0.13
Commercial                 - 0.012    0.26
Pedagogical'                0.037    0.83
Social services             0.037    0.68
Labor market factors
Labor force participation rate    - 0.089    0.61
Per capita income             0.046    0.79
Constant                     - 0.205    1.10
D-value                           0.009
Sample size                       2,666
Chi square                        25.71
a. Curriculum program categories are evaluated in relation to the pedagogical program
(the omitted category).
b. Pedagogical score relates only to conmrol students; INEM schools do not offer teacher
training.
Source: Maximum likelihood estimation by the probit procedure of the Choice program,
courtesy of Eric Swanson, the World Bank.
108



THE COLOMBIA CASE STUDY                                         109
industrial) and earnings one year after graduation from secondary school
also were part of the model.
Estimates of the model using academic achievement as an intermediate
variable are provided in figure 5-8. The model is fitted to the data for all
graduates with earnings in 1982, because all of them had been exposed to
academic subjects. No one factor, even academic achievement, was shown
to have a significant effect on earnings. Neither did the link between school
type and earnings reveal any advantage of INEM over control schools in
this context. This reinforces the previous finding that earnings differences
between INEM and control graduates are slight.
Figure 5-9 shows the results of the same model using scores on industrial
instead of academic achievement tests as an intermediate variable. In
contrast to the first model, industrial achievement has a significant impact
on earnings. Scores on commercial, social services, and agricultural
achievement tests, however, have no such effects (not shown in figure).
Figure 5-8. Path Model with Academic Achievement Scores Used
as Intermediate Variable, Colombia, 1982 Survey of 1981 Cohort
Diersifed
school
- 0.073*
Family    >     /           \0                 0.054
bac und                               0.147*
0.013
0.033                          Earning
\    /      ;086#  <0.086/
/  >                    / ~~~~~~~Academic
\              z  ~~~~~~~achieve-
AbilitY)        0.404               et        R2= 0.007
N =51o
*Statistically significant path at the 10 percent level.



110   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Figure 5-9. Path Model with Industrial Achievement Scores Used
as Intermediate Variable, Colombia, 1982 Survey of 1981 Cohort
school 
in e  in t o0.033* dl                              Eamings)
thsewo  er   orin  ul-tm  in 192 Itshexplanator  oeri o
Abihey  0.404*               /  ~~~~~~R2 =0.0012
iStatistically significant path at the 10 percent level or better.
The overall pattern is not changed when part-time wage earners are
included in the model.
Table f -uprsue e            results cf an earnings function fitted to data for
those who  were working  full-time in 1982. Its explanatory power is low
(13 percent of the variance n     i torthngs). Difference in local per capita
income is the only variable whose coefficient is statistically significant at
the S percent level of probability.
Summary of the 1982 Tracer Survey
The follow-up survey of seconcdary school students in Colombia was
taken one year after they graduated. This allows for only a partial testing
of the set of hypotheses relating to the external efficiency of INEM diver-
sified schools compared with control schools. However, the available data
suggest that the provision of prevocational studies:



THE COLOMBIA CASE STUDY                                        1ll
* Has not decreased the private demand for postsecondary formal
education, nor has the introduction of INEMS
* Has not resulted in prevocational-stream graduates entering training
courses in a pattern different from academic-stream graduates
* Has slightly increased the propensity of prevocational students to
enter full-time employment immediately after graduation from sec-
ondary school'
* Has not given graduates of prevocational programs access to jobs that
are different from those taken by graduates of academic programs
(similarly, there are no significant differences based on type of school
attended)
* Has not led prevocational graduates to acquire employment earlier
than academic graduates (in fact, unemployment rates were slightly
higher among INEM graduates)
* Has not led to higher initial earnings for prevocational students
(although the average earnings for graduates of INEM and non-INEM
schools are virtually the same, the average earnings of academic
program graduates from both school types are higher than those of
graduates from all prevocational programs).
The Class of 1978 Three Years Later
Given the long time required to obtain data for longitudinal studies, it
was decided to seek some early indications of the labor market destination
and performance of secondary school graduates by a survey of pre-1981
students.
After selecting the sample of schools and students for the 1981 cohort, a
target sample of 2,000 students who had graduated from the same schools
in 1978 was randomly chosen from the school records. Attempts were
made to locate these graduates for the administration of a special ques-
tionnaire. Of the 2,000 selected, 1,826 were interviewed. The question-
naire elicited retrospective information on the graduates' further educa-
tion and occupational records between 1978 and the fall of 1981 when the
interviews took place. The questionnaire also sought the graduates' opin-
ion on the usefulness of the courses they had taken and included questions
designed to measure the degree to which they had adapted to "modern"
values and attitudes. The questionnaires were completed by personal
interviews at the home of each selected graduate.
1. This is significant only for control schools. If those studying and working are added, this
difference disappears, but then graduates from control schools as a whole appear slightly
more likely to be employed.



112  DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-41. The Determinants of 1982 Full-Time
Labor Earnings, Colombia, 1981 Cohort
Variable'            Coefficient  t-value
Out-of-school factors
Male                          - 615.281   0.96
Age                            -21.228   0.15
Family owns business           - 44.152   0.06
Books owned by household        - 1.010   0.19
Father's education              87.971   1.06
Log of family income            88.460   0.50
Father blue-collar             405.605   0.50
Father nonmanual employee     -410.076   0.69
Father is employer          - 1,346.744   1.49
Verbal aptitude                 32.610   1.04
Math aptitude                   -4.052   0.14
School factors
INEM
Academic                     466.616   0.41
Agricultural                3,558.104   0.88
Commercial                - 1,465.604   1.37
Industrial                   192.290   0.15
Social services             -730.687   0.58
Control
Academic                    - 936.234   0.96
Agricultural                 487.142   0.40
Commercial                 - 417.407   0.41
Industrial                 - 740.477   0.70
Social services             5,255.244   1.87
Test score
Academic                      49.475   1.38
Agricultural                 - 19.145   0.65
Commercial                    28.598   0.88
Pedagogicalb                 -21.985   0.79
Social services              -32.195   0.88
Labor market factors
Labor force participation rate    - 179.504   1.47
Per capita income                 4.068'  3.46
Constant                    -7,382.444   1.30
R2                                 0.13
Sample size                         274
a. The omitted school subject is pedagcgy. The omitted father occupation is farmer.
b. Pedagogical test score relates only tD control students; INEM schools do not offer
teacher training.
c. Statistically significant at 5 percent or better.



THE COLOMBIA CASE STUDY                                         113
Table 5-42. Distribution of the Sample by School Type
and Subject, Colombia, 1978 Cohort
(number of graduates)
School type
Curriculum  INEM  Control Total
Academic    181    277   458
Agricultural    76    150   226
Commercial   184    181   365
Industrial    213    156   369
Pedagogical    -    250   250
Social services  120  38   158
Total       774  1,052  1,826
- Not applicable.
This data set was used to test the same hypotheses related to diversifi-
cation as were postulated with the 1981 cohort. The exception was
achievement tests; none of the cognitive tests given the 1981 cohort had
been administered to the 1978 group while still in school. (For a more
comprehensive analysis of the 1978 cohort, see Psacharopoulos and
Zabalza 1984a and 1984b.)
The 1978 sample yielded a total of 1,826 observations distributed by
type of school and subject as shown in table 5-42. The effective sample of
INEM students corresponds to 14 percent of the total enrollment in the
eleventh-year class of INEM schools in 1978. Data for the 1978 cohort
confirm that the INEM schools draw students from the lower socioeconom-
ic groups (see table 5-43).
Destination of Graduates
Table 5-44 compares the activities of graduates immediately after leav-
ing school with what they were doing three years later in 1981. In general,
the table shows rather little change over the three-year period, but there is
a surprising increase in those who classify themselves as students. How-
ever, a cross-tabulation of initial postsecondary experiences against 1981
activities shows that the graduates actually changed their direction a good
deal during the three-year span. The variety of destinations is illustrated in
figure 5-10.
Two points are worth noting regarding the overall destination of the
1978 cohort. First, graduates are likely to remain in their initial destina-
tion (such as work rather than study) for at least three years. Hence, by
studying the initial destination of graduates, one can to a large extent



114    DIVERSIFIED  SECONDARY EDUCATION AND DEVELOPMENT
Table 5-43. Mean Socioeconomic Characteristics of Students
by Type of School, Colombia, 15978 Cohort
Characteristic        INEM   Control Overall
Father's years of schooling        9.6     9.9     9.7
Mother's years of schooling        9.4     9.7     9.6
Father's monthly income (pesos)    16,674  19,001  18,022
Family monthly income (pesos)  26,426  30,027  28,440
Number of dependents in family     1.4     0.9     1.1
Primary school repeater (percent)  25.8    20.3    22.7
Secondary school repeater (percent)   15.9    24.6    20.9
Urban born (percent)              90.3    85.5    87.5
predict their longer-term destination. Second, for those who continue to
study, a three-year follow-up cannot fully evaluate their performance in
the labor market. For these reasons, it probably would be unprofitable to
poll the 1981 cohort before 1986.
Postsecondary Education and Further Training
When the 1978 cohort was interviewed in 1981, 50.8 percent of the
sample reported they were studying full-time or part-time. There were
significant differences among students in the likelihood that they would
still be a student in 1981, which depended upon the type of school
attended or the subject studied. As shown in table 5-45, INEM students
who had been in the academic stream had the highest probability that they
would be continuing to study three years after graduation (75 percent of
them did so). Those least likely to have continued their studies were
agricultural graduates of non-INEM schools (only 20 percent did so).
Table 5-44. First Destination of 1978 Graduates
and Their Status in 1981, Colombia, 1978 Cohort
(percent)
First destination,  1981
Activity          1978       status
Study                  26.3       28.8
Work and study         26.2       22.1
Work                   36.4       36.7
Other than abovea      11.1       12.4
Total                 100.0      100.0
a. Includes those who are not participating in the labor force and those who are looking
for work.



THE COLOMBIA CASE STUDY                                          115
Table 5-45. Percentage of Students Continuing to Study in 1981,
by School Type and Curriculum, Colombia, 1978 Cohort
INEM            Control         Total sample
Full- and         Full- and         Full- and
Curriculum   part-time Full-time part-time Full-time part-time Full-time
Academic      75.1      56      53.8      38      62.2      45
Agricultural    44.7    35      20.0      15      28.3      22
Commercial    44.0      18      59.7      25      51.8      22
Industrial    51.6      29      48.1      30      50.1      29
Pedagogical    -        -       54.8      19      54.8      19
Social services   45.0  25      58.3      25      47.2      25
Overall       53.6      33      49.4      26      51.2      29
- Not applicable.
Of course, the inclination for further study is also influenced by non-
school factors, such as family background. To isolate the effect of school
type and curriculum on students' propensity to continue their education, a
logit regression was fitted, as shown in table 5-46. Graduates of academic
programs in INEM schools are the most likely to continue studying after
secondary school. And a considerably greater percentage of those who
study agriculture in INEMS went on to further study than their non-INEM
counterparts.
As the survey indicates, a proportionately higher number of INEM
graduates continue with education than do non-INEM graduates. This
finding contradicts expectations about postsecondary experiences for the
two groups. But INEM graduates receive a better secondary education, and
that allows them to succeed in the higher education entrance examina-
tions.
Of the 1,826 respondents, 800 had had a certain amount of further
training between graduation and the time of the interview. Most of this
further training took place in SENA (the national apprenticeship service), in
technical institutes, in firms, and in universities, as indicated in table 5-47.
One of the aims of diversified schools is to produce graduates who will
undertake additional training in the area of their specialization. It is
therefore important to investigate the extent to which INEMS performed
better than non-INEM schools in this respect, as well as the consistency
between field of secondary school specialization and further training
among both INEM and control students.
The mean duration of training and the proportion of graduates of each
subject who undertook training is shown in table 5-48. The results suggest
that INEMS do perform better than control schools, largely because the
postsecondary training undertaken by their graduates lasts much longer



Figure 5-10. First Destination and 1981 Status
of the 1978 Cohort, Colombia
First destination,                   1981 status                 Total for
1978                                                  each 1981 status
Study
only
308
/   Study and  |
work
only     :         ~         oonly
48                               Wor               525
/              \       1    ~~~~~~~62 
/                         1    ~~~~~~~~~~48
Study
only
81
Study and
work
/  < ;;;tj   work 23                                         Study and
work
work                                         ~~~~~~~~~~~~~~403
/      /         ~~~~~~478                     Wor03
cohort
N= 1,26                                                    Study
only
69
\ \ | ~~~~~~~~~~~~~~~Workl
\   5   only  _                                             ~~~~~~~~~~~~~~~~~only
Wk 65                         o_ 671
only~~~~~~~~~~ol
\6651    419
other
\79 
Study
only
\              /          ~~~~~~~~Study and
O1°3r                   |    Othte
78 
1'16



THE COLOMBIA CASE STUDY                                                117
Table 5-46. Factors Influencing the Probability of Studying
Immediately after Secondary School, Colombia, 1978 Cohort
Marginal
Variable                 effect   t-value
Background
Age                                 4.62a   0.00
Male                                1.62    0.58
Number of siblings                  0.02    0.96
Urban born                          1.76    0.66
Residence years in a city           0.54'   0.03
Current urban resident             -2.45     0.57
City per capita income
(thousarxds of pesos monthly)  0.0001    0.73
Primary school repeater            -3.72     0.19
Secondary school repeater           0.37    0.90
Private secondary school graduate   0.27    0.94
Father's occupationb
Employee                          0.99    0.80
Laborer                          -2.32    0.65
Self-employed                     2.25    0.58
Business owner                   -4.94    0.35
Father's years of schooling         0.53    0.18
Mother's years of schooling         1.04'    0.004
Curriculum
Academic
INEM                             32.73    0.000
Control                          25.OSa   0.000
Agricultural
INEM                             23.61a   0.000
Control                           8.29    0.27
Commercial
INEM                              7.93    0.13
Control                          11.S4a   0.02
Industrial
INEM                              9.85a   0.06
Control                          13.64a   0.01
Social services
INEM                             19.47a   0.000
Control                          21.26a   0.05
D-value                                 0.093
Sample size                             1,712
Note: Mean dependent variable = 26 percent.
a. Statistically significant at the 10 percent level or better.
b. Base category for father's occupation is farmer.



118    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-47. Location of PostsecC,ndary School Training
by School Type, 1978 Cohort
(percent)
Training location  INEM  Control
SENA             47.3   37.9
Technical institute  29.1   25.8
Firm               9.0    9.8
University        4.2    5.1
Other             10.4   21.4
Total            100.0  100.0
than that undertaken by control graduates (forty-two weeks and twenty-
nine weeks, respectively).
Graduates' further training matches their subjects of specialization in
school to a much greater extent than would result from a random alloca-
tion (table 5-49). The data also show that, particularly in those subjects of
specialty where the degree of congruence is highest (commercial and
industrial), the difference between INEMS and control schools is not signifi-
cant. However, agricultural students are more likely to enter teaching and
engineering than to continue their agricultural specialization.
Labor Force Participation
Students coming from INEM schools are less likely to participate in the
labor force or to hold a job three years after graduation than are control
students (table 5-50). The results of standardization for other factors
Table 5-48. Mean Duration of Postsecondary School Training
and Proportion of Graduates Und'ertaking Training
by School Type and Curriculum, Colombia, 1978 Cohort
Percentage of graduates
Weeks of training   undertaking training
Curriculum    INEM  Control    INEM     Control
Academic       45.0   37.0       40.7      46.7
Agricultural    25.6   14.2      34.7      49.0
Commercial    42.3   4:1.0       52.8     37.1
Industrial    43.2   3 7.4       37.8      38.6
Pedagogical    -      19.6       -         42.5
Social services   44.8   20.8    68.7      75.0
Overall       42.4   29.2        46.4     43.8
-Not applicable.



THE COLOMBIA CASE STUDY                                    119
influencing the probability of employment, by means of a logit regression
(table 5-51), reinforce the finding that INEMS do not specialize in produc-
ing graduates who enter the labor market directly (table 5-52).
Employment Characteristics
Of the sample of 1,826 graduates, more than 1,000 had jobs in 1981.
Most were employed full-time (79 percent) and in the private sector of the
economy (69 percent). Only a minority (7 percent) were self-employed.
The overall mean earnings of the 1978 cohort in 1981 (10,292 pesos per
month) is consistent with the mean earnings of the younger 1981 cohort in
1982 (7,552 pesos in 1981 prices), because the four years of extra 1bof-
market experience of the older group gives it substantially high rearnings
in real terms. It should of course be remembered -that because the two
cohorts are different, their earnings are not strictly comparable. Rapid
growth of earnings during the early career of graduates has been exten-
sively documented for both advanced and developing countries (Mincer
1974; Fields and Schultz 1980).
Again there was virtually no difference between the earnings of INEM
and control graduates (table 5-53). When the self-employed are excluded,
academic and industrial graduates of control schools enjoy the highest
earnings. The most significant difference in earnings appears between the
sexes. Women from both types of school earn less than their male counter-
parts, and INEM women earn even less than control-school women (table
5-54).
To standardize earnings for factors other than schooling, a descriptive
earnings function was fitted among those who were not self-employed
(table 5-55). The explanatory power (R2 = 16.4 percent) is consistent with
previous findings when account is taken of the cross-sectional character of
the sample and the narrow range of the age variable (the subjects were
mostly aged twenty-one to twenty-three). The important findings for the
purposes of this research are, first, that neither school type nor curriculum
program have a significant effect on earnings. The only significant vari-
ables-such as sector of employment, job experience, and hours worked-
are not related to school experience. Second, among students following the
same curriculum program, there are no significant differences in earnings
between INEM and control schools. The above conclusions were not
modified by fitting the earnings function in alternative formulations (such
as semi-log) or by including the self-employed in the study.
Table 5-56 shows mean earnings, standardized for the out-of-school
factors listed in table 5-55, and the net earnings advantage of INEM
graduates compared with the control group. Four out of five subject
specializations show negative comparative results for the INEM group.



Table 5-49. Relationship between Secondary School Subject and Postsecondary Specialization
by School Type, Colombia, 1978 Cohort
(percent)
Secondary school curriculum
Social
Academic   Commercial  Industrial  Agricultural   services
Postsecondary
specialization       I     C      I     C     I      C     I     C     I      C
Agricultural sciences     11      3     0      2    1       4   24    10    0        0
Architecture.
fine arts                4      5     0      5   12       8    0      0    0       0
Economics,
business administration   20    24   50    50    8        3    6    14   17    50
Education                 16    22    10       5   12    11   29    24   32    29
Engineering                16     8    4       3   38    44   27    17    6          7
Health sciences             8    11     1    12    3        4    6      3   17       0
Law                         8     7   10       6    3       3   10      3    4       0
Sciences                   3      4     0      2    8       9    0      3    2       0
Social services,
humanities              14    15   21       17    6       4    0    21   24    14
Other                      0      2    4       0    0       0    0      3    0       0
Total                    100   100  100   100  100   100  100   100  100   100
Note: I  INEM; C = control.



Table 5-50. Labor Force Participation and Employment Rates in 1981, Colombia, 1978 Cohort
(percent)
Total
INEM                    Control                   sample
Labor force Employment Labor force Employment Labor force Employment
Curriculum    participation   rate     participation   rate     participation    rate
Academic         49.2         37.0         61.4        46.2         56.6         42.6
Agricultural     63.2         50.0         78.7        58.0         73.5         55.3
Commercial       79.4         66.9        74.6         66.9         77.0         66.9
Industrial       73.7         62.9        72.4         66.7         73.2         64.5
Pedagogical       -            -          81.2         73.2         81.2         73.2
Social services  68.3         58.3        79.2         66.7         67.1         56.3
Overall          67.4         55.8        73.0         61.6         70.4         58.8
Note: A labor force participant was defined as one who was either working (full-time or part-time) or was looking for work. Employment rates are for
both full-time and part-time employment.



122  DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-5 1. Variables Influencing Employment in 1981,
Colombia, 1978 Cohort
Marginal
Variable                          effect  t-value
Background
Age                                                 4.5'   0.000
Male                                                7.7'   0.020
Number of siblings                                  0.1    0.656
Urban born                                          0.3    0.928
Residence years in a city                         - 0.4'   0.080
Current urban resident                            -2.2    0.641
City per capita income (thousands of pesos monthly)  0.02'  0.008
Primary school repeater                             1.6    0.588
Secondary school repeater                           0.1    0.962
Private secondary school graduate                 - 3.7    0.417
Number of postsecondary courses                     6.5'   0.000
Father's occupationb
Employee                                          3.8    0.399
Laborer                                         -2.1    0.695
Self-employed                                   -2.7    0.550
Business owner                                    5.6    0.338
Father's years of schooling                       -0.2    0.578
Mother's years of schooling                       -0.5    0.205
Curriculum
Academic
INEM                                           - 40.4'   0.000
Control                                        - 29.9'   0.000
Agricultural
INEM                                           -25.1a   0.000
Control                                        - 15.1'   0.049
Commercial
INEM                                           - 11.9a   0.030
Control                                         -1.2    0.824
Industrial
INEM                                           -18.1'   0.001
Control                                        - 11.8'   0.050
Social services
INEM                                           -22.6'   0.000
Control                                        -12.4    0.318
D-value                                               0.105
Sample size                                           1,690
Note: Mean dependent variable  58.8 percent.
a. Statistically significant at the 5 percent level or better.
b. Farmer is the reference occupational category.



THE COLOMBIA CASE STUDY                                                     123
Table 5-52. Adjusted Probabilities of Working in 1981
by School Type and Curriculum, Colombia, 1978 Cohort
(percent)
INEM
Curriculum   INEM   Control advantage
Academic      34.1a   4390a      - 9.8
Agricultural   49.20   59.0a     - 9.8
Commercial   62.4a   71.8        - 9.4
Industrial    56.40   62.2a      -5.8
Pedagogical    -      73.2        -
Social services  51.7a   61.5    -9.8
- Not applicable.
Note: Based on logit reported in table 5-51.
a. Significantly different from pedagogy (reference category).
Table 5-53. Mean 1981 Labor Earnings by School Type
and Curriculum, Colombia, 1978 Cohort
(monthly pesos)
INEM                 Control             Total sample
Excluding             Excluding             Excluding
All      self-        All      self-       All       self-
Curriculum   graduates employed   graduates employed   graduates employed
Academic         9,844     9,297     10,375    10,193      10,191      9,882
Agricultural    12,322     9,630      9,145      9,340     10,109      9,420
Commercial      9,904      9,811     10,155      9,874     10,028     9,842
Industrial     11,834    11,262      11,957    11,713      11,888    11,460
Pedagogical      -         -          9,526      9,495      9,526      9,495
Social services    8,862   8,921      9,050      9,050      8,898      8,895
Overall        10,534    10,001      10,142    10,016      10,300    10,002
Table 5-54. Mean 1981 Labor Earnings
by School Type and Sex, Colombia, 1978 Cohort
(monthly pesos)
Total
Sex    INEM    Control  sample
Male    11,473   10,822    11,131
Female   8,913    9,543      9,332
Overall  10,534   10,142    10,292



Table 5-55. Determinants of Postsecondary Students'
Labor Earnings in 1981, Colomb,a, 1978 Cohcrt
Variable              Coefficient t-value
Background
Male                               1,247a   3.45
Age                                 266'   3.10
Number of siblings                  -37     0.75
Urban born                          -44    0.09
Residence years in a city           - 17    0.61
Current urban resident              712    1.35
City per capita income
(thousands of pesos monthly)       1    1.07
Primary school repeater             243    0.76
Secondary school repeater          - 143    0.41
Father's occupationb
Employee                         - 215    0.45
Laborer                          - 521    0.89
Self-employed                      127    0.25
Business owner                   1,128    1.74
Father's years of schooling          123'   2.47
Mother's years of schooling          -3     0.05
Postsecondary characteristics
Private secondary school graduate   244    0.52
Study in 1981                        847'   2.80
Number of postsecondary courses       50    0.52
Duration of training courses           5    1.01
Months in present job                 33a   3.42
Private sector employee          - 1,323'   4.18
Full-time worker                   1,054'   2.49
Weekly hours worked                   62'   4.22
Curriculumc
Academic
INEM                             - 560    0.81
Control                             37    0.06
Agricultural
INEM                               131    0.14
Control                          -769    0.99
Commercial
INEM                             -231    0.40
Control                             64    0.11
Industrial
INEM                               153    0.24
Control                            770    1.23
Social services
INEM                             - 836    1.23
Control                          -551     0.49
Constant                         -2,635    0.96
Rz 20.164
Sample size                               884
Note: Excludes self-employed. Mean dependent variable = 10,002.
a. Statistically significant at the 5 percent level or better.
b. Farmer is the reference category.
c. Set of curriculum-specific dummies. Reference group is pedagogical graduates.
i24



THE COLOMBIA CASE STUDY                                        125
Table 5-56. Adjusted Mean Earnings in 1981
by School Type and Curriculum, Colombia, 1978 Cohort
(monthly pesos)
INEM advantage
Curriculum   INEM   Control   (percent)
Academic     9,488  10,109     -6.1
Agricultural  10,178   9,522     6.9
Commercial   9,816  10,137     -3.2
Industrial    10,199  10,834   -5.9
Pedagogical    -    10,085      -
Social services  9,211   9,304  -1.0
Note: Based on.table 5-55.
Returns to Investment in Secondary Schooling
Differences in mean earnings across school types and curriculum pro-
grams have only low levels of significance. In the following analysis,
however, these means are compared with education costs, and rough
estimates of rates of return are calculated. Because we are dealing with a
cohort three years after graduation, we can approximate a more valid "flat
equivalent" earnings differential at age twenty-two and hence use the
shortcut rate of return method (see figure 5-11, and Psacharopoulos
1981b). Monthly earnings by school type and curriculum program were
given for the 1978 cohort in table 5-54. Annual social costs for the 1981
cohort, including earnings forgone, were presented in table 5-20. From
these data, approximate rates of return were estimated for each secondary
school curriculum program and school type by the formula,
12 (Yi-Y )
6 (12 Y, + Ci)
where Yi is the mean monthly earnings of graduates of the i type of
secondary curriculum, Yp is their forgone monthly earnings while studying
(measured by the mean earnings of primary school graduates), and Ci is
the direct annual unit cost of the i type of secondary curriculum. The
results are shown in table 5-57.
Vocational education in general appears to have slightly higher returns
when provided by INEM schools, but the returns to academic education are
higher in control schools. Given the limitations of the earnings and cost
data and the simplifying assumptions of the shortcut formula, the differ-
ences in the rates of return between INEM and control schools cannot be
regarded as significant. Although some curriculum programs appear to be
more profitable than others, the differences are not great enough to



126   DIVERSIFIED SECONDAR'Y EDUCATION AND DEVELOPMENT
conclude that the introduction of prevocational tracks reduces the eco-
nomic efficiency of secondary schooling. Conversely, there is no evidence
to suggest that their introduction increases economic efficiency.
Incidence of Job Search
Eighteen percent of the 1978 cohort reported they were seeking work in
1981, with a very small overall difference between INEM and control
schools (table 5-58). Most of those looking for a job either were already
employed or were studying. When such cases are excluded, the overall
unemployment rate among the 1978 cohort in 1981 was 6 percent, with a
negligible overall difference betweenl INEM and control schools. The unem-
ployment rate was especially pronounced among control-school agricul-
tural graduates (17.3 percent).
Table 5-59 reports the results of a logit regression explaining the prob-
Figure 5-11. Flat-Earnings-Equivalent Assumption
for Approximating the Returns to Education
Earnings
True profile for secondary graduates   - -
I   Flat approximation of profile
*.               for secondary graduates
y.
True profile for primary graduates
\ Flat approximation of profile
r    Indirect                            for primary graduates
cost
13               19       22                                   Age
Direct    (1978)  (1981)
cost



THE COLOMBIA CASE STUDY                                                 127
Table 5-57. Approximate Social Rates of Return
to Investment in Secondary Education by Curriculum
and School Type, Colombia, 1978 Cohort
(percent)
Curriculum    INEM     Control
Academic      7.7      9.3
Agricultural   9.11     7.21
Commercial   8.4l       9.3l
Industrial    9.2 f88   9.9 [83
Social services  7.2J   7.2J
Note: Based on adjusted earnings, table 5-56.
a. Average of agricultural, commercial, industrial, and social services.
ability of looking for a job in 1981 among labor market participants. The
findings indicate that graduates of the academic and agricultural programs
(from either INEM or control schools) were much more likely than other
graduates to be seeking a job. According to table 5-60 (after adjustment
for out-of-school factors), control-school industrial graduates were the
least likely to be seeking a job in 1981.
Duration of Job Search
The 1978 graduates spent an average of fourteen weeks before starting
their first job (table 5-61). Those coming from INEM schools obtained jobs
Table 5-58. Job Search and Unemployment Rates in 1981
by School Type and Curriculum, Colombia, 1978 Cohort
(percent)
INEM              Control          Total sample
Search  Unemploy-  Search  Unemploy-  Search  Unemploy-
Curriculum    rate    ment rate   rate   ment rate   rate    ment rate
Academic       15.5      3.9      18.4      6.1      17.3       5.2
Agricultural    17.1     9.2      24.0     17.3      21.7      14.4
Commercial    17.9       8.2      13.8      3.9      15.9       6.0
Industrial    21.1       4.7      14.1       1.9     18.2       3.5
Pedagogical    -         -        17.2      4.0      17.2       4.0
Social services  19.2    5.0      16.7      4.0      18.4       4.8
Overall       18.4       5.8      17.4      6.2      17.8       6.0
- Not applicable.
Note: Search rate includes those who might be studying or working, or both, while
looking for a job. Unemployment rate refers to those who are strictly looking for a job and
neither working nor studying at the same time.



128  DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-59. Determinants of the Probability
of Looking for a Job in 1981, Colombia, 1978 Cohort
Variable                 Marginal effect  Significance
Background
Male                                -7.74a         0.01
Age                                  -1.03         0.16
Number of siblings                    0.22         0.58
Urban born                           - 0.07        0.98
Residence years in a city             0.12         0.60
Current urban resident                5.77         0.15
City per capita income
(thousands of pesos monthly)    - 0.02'        0.01
Primary school repeater               0.17         0.94
Secondary school repeater             1.40         0.62
Private secondary school graduate    - 1.45        0.74
Number of postsecondary courses      - 1.76a       0.05
Duration of training courses        -0.01          0.86
Father's occupationb
Employee                            1.6          0.68
Laborer                             5.73         0.20
Self-employed                       1.24         0.75
Business owner                    -5.65          0.32
Father's years of schooling           0.11         0.76
Mother's years of schooling          -0.63         0.12
Curriculum'
Academic
INEM                               17.130        0.001
Control                            13.44a        0.001
Agricultural
INEM                               13.2a         0.04
Control                            11.97a        0.05
Commercial
INEM                                7.41         0.12
Control                           - 1.81         0.73
Industrial
INEM                                9.65'        0.07
Control                             1.74         0.78
Social services
INEM                                7.9          0.16
Control                           - 1.66         0.88
D-value                                      0.048
Sample sized                                 1,168
Note: Mean of the dependent variable   17.8 weeks.
a. Statistically significant at the 5 percent level or better.
b. Farmer is the reference category.
c. Reference group is pedagogical gradua. es.
d. Refers to labor force participants.



THE COLOMBIA CASE STUDY                                                 129
Table 5-60. Standardized Probability of Looking for a Job
in 1981 by School Type and Curriculum,
Colombia, 1978 Cohort
(percent)
Curriculum   INEM   Control
Academic      25.6a   22.7a
Agricultural   19.9a   26.8a
Commercial   16.9   10.3
Industrial    14.4a    8.4
Pedagogical    -      10.6
Social services  17.1   11.1
Overall           16.4
- Not applicable.
Note: Based on table 5-59.
a. Significantly different from the pedagogical control.
one week sooner than the control group; however, school type and cur-
riculum program are statistically insignificant in explaining job search
duration (table 5-62).
Inspection of the minimum acceptable salary among those looking for a
job reveals that job-seekers had a reservation wage much higher than the
actual wages of their classmates who were already employed (table 5-63).
Clearly, most of the job-search activity among the respondents was to
improve their current economic situation.
Conclusions from Surveying the 1978 Cohort
The evidence from the earlier cohort largely corroborates the conclu-
sions about external efficiency reached on the basis of studying the 1981
cohort. In particular:
Table 5-61. Weeks Looking for First Job by School Type
and Curriculum, Colombia, 1978 Cohort
(percent)
Curriculum   INEM  Control Totalsample
Academic      11.0   12.7       12.1
Agricultural   20.1   23.8      22.3
Commercial   13.0   13.1        13.0
Industrial    12.4   12.1       12.3
Pedagogical    -      13.0      13.0
Social services  15.9   23.6    18.4
Overall       13.6   14.5       14.2



130    DIVERSIFIED  SECONDAR'C EDUCATION AND DEVELOPMENT
Table 5-62. Determinants of Duration of Search for First Job
after Secondary School Graduation, Colombia, 1978 Cohort
Variable               Coefficient I-value
Background
Age                               -0.33      0.69
Male                                1.27     0.68
Number of siblings                  0.29     1.09
Urban born                          1.75     0.72
Residence years in a city           0.48a    2.93
Current urban resident              6.31a    2.22
City per capita income
(thousands of pesos mointhly)  -0.01a    2.51
Primary school repeater           - 1.46     0.82
Secondary school repeater           1.04     0.56
Private secondary school graduate   0.65     0.24
Father's occupation b
Employee                        - 0.80     0.29
Laborer                           1.44     0.45
Self-employed                   -3.97      1.44
Business owner                  -1.88      0.51
Father's years of schooling         0.35     1.32
Mother's years of schooling       - 0,70a    2.50
Curriculumc
Academic
INEM                            -2.77      0.74
Control                         - 0.33     0.10
Agricultural
INEM                              6.07     1.42
Control                           4.63     0.95
Commercial
INEM                              0.02     0.009
Control                           0.85     0.25
Industrial
INEM                            - 1,12     0.35
Control                         - 1.63     0.46
Social services
INEM                              4.65     1.28
Control                           8.04     1.10
Constant                           30.72     -
R2                                     0.091
Sample size                              579
Note: Mean of the dependent variable   14.2 weeks.
a. Statistically significant at the 5 percent level.
b. Reference group is farmer.
c. Reference group is pedagogical gradus.tes.



THE COLOMBIA CASE STUDY                                           131
Table 5-63. Reservation Wage of Those Looking for a Job
in 1981, Colombia, 1978 Cohort
(monthly pesos)
Subject    INEM   Control Totalsample
Academic    14,988  10,618    12,167
Agricultural  13,500  12,971    13,115
Commercial   13,678  12,739    13,285
Industrial    14,423  15,191    14,683
Pedagogical    -    13,125    13,125
Social services 11,000  10,000    10,793
Overall     13,713  12,480    13,019
-Not applicable.
* The simple examination of unstandardized data suggests that earn-
ings differences between graduates from INEMS and from traditional
(control) schools are not very large. Any differences observed seem to
be attributable largely to the higher earnings of self-employed people.
* The raw data on employment status suggest that INEMs have not
noticeably changed the employment pattern produced by the tradi-
tional school system. If anything, the data suggest that INEMS (particu-
larly in the case of males) have accentuated the propensity of high
school graduates for higher education. (The detailed results by sex are
given in Psacharopoulos and Zabalza 1984a.)
* INEMS do not differ from traditional schools in the proportion of their
graduates who enroll in Colombia's best universities, nor in the pro-
portion who are unemployed after three years.
* When earnings are adjusted for influences other than differences in
secondary schooling, the conclusions of the descriptive analysis are
largely maintained. Overall, the data do not support the hypothesis
that INEMS generate an earnings structure different from that gener-
ated by traditional schools.
* Although male earnings grow significantly with labor market experi-
ence, female earnings show hardly any growth at all. Tenure on the
job seems to be a significant determinant of earnings for women but
not for men. Male graduates who specialized in agricultural studies
earn significantly less than other male graduates, while female grad-
uates who specialized in commercial studies earn significantly more
than women graduates working in other fields.
* When data on employment status were standardized to take into
account the influence of variables that determine the relative attrac-
tiveness of work versus study, the main differences found for women



132   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
were not between INEMS and traditional schools, but between differ-
ent subjects of study. For men, there were statistically significant
differences between INEMs and traditional schools.
Male INEM graduates, in particular those who followed an academic
curriculum, are much more likely to end up studying full-time than are
their counterparts from tradirional schools. Graduates following a
commercial curriculum in traditional schools are, however, much
more likely to be working full-time than their peers from other cur-
riculum programs. For graduates of other programs, the probability
of "working only" is higher, and that of "studying only" lower, for
INEM graduates than for control graduates.
Our overall conclusion is that INEMS have not increased the propensity
of high school graduates to enter the labor force. In fact, where a signifi-
cant difference can be identified, INEMS have tended to produce graduates
who opted for further studies, rather than for working immediately after
graduation. INEMS have not influenced earnings, but a final conclusion on
this must await the availability of data from university graduates, since
INEMS have sent a greater proportion of their graduates to universities than
have traditional schools.



THE COLOMBIA CASE STUDY                                                133
Table 5-64. Mean and Standard Deviation
of Selected Variables, Colombia, 1981 Cohort
Standard
Variable                     Mean   deviation
Base-year sample (N = 8,050)
Background
Male                                     0.509     0.501
Current urban residence                  0.873     0.333
Age                                       18.5    1.750
Number of siblings                        5.17      2.84
City labor force participation rate (percent)    15.10  18.10
City per capita income (pesos per month)  1,041      862
Family income (pesos per month)         27,007   47,036
Father's occupation
Blue-collar                            0.132    0.339
Nonmanual employee                     0.365    0.482
Self-employed                          0.252     0.434
Farmer                                 0.131     0.337
Employer                               0.110     0.314
Father's years of schooling               6.46      3.60
Ability
Verbal aptitude                          51.12      9.09
Math aptitude                            51.36      9.65
Attitudes
Individual "modernity" score             19.40      6.32
Plans further full-time education        0.254     0.473
Expects private sector job               0.096     0.240
Expected earnings upon graduation (pesos)   10,754  7,017
Expected waiting time to first job (weeks)  16.10   14.3
Achievement score (nonstandardized)
Academic                                 11.78      4.67
Agricultural                              4.27      2.40
Commercial                                4.47      2.41
Industrial                                7.24      3.85
Pedagogical                               3.68      2.20
Social services                           5.90      2.75
School type and curriculum
INEM                                     0.404     0.491
Academic                               0.108     0.311
Agricultural                           0.022     0.145
Commercial                             0.119     0.324
Industrial                             0.108     0.310
Social services                        0.047    0.212
(Table continues on the following page.)



134    DIVERSIFIED SECONDAFY EDUCATION AND DEVELOPMENT
Table 5-64 (continued)
Standard
Variable                    Mean  deviation
School type and curriculum (continued)
Control
Academic                              0.183    0.386
Agricultural                          0.046    0.210
Commercial                            0.121    0.326
Industrial                            0.117    0.322
Pedagogical                           0.095    0.214
Social services                       0.034    0.181
Other school-related variables
Repeated primary school               0.237    0.426
Per pupil annual school expenditure  628.565  1,125.60
Private student expenditure           1,874      626
Student-teacher ratio                 18.85     4.94
Private school graduate               0.194    0.395
1982 Follow-up (N  4,925)
Destination
Studying                                0.374    0.484
Working                                 0.297    0.457
Working and studying                    0.104    0.305
Looking for work                        0.312    0.463
Further schooling
Full-time university study              0.194    0.396
Other coursework                        0.174    0.373
Postsecondary training                  0.256    0.437
Labor market outcomes
Salaried employment                     0.265    0.417
Self-employment                         0.067    0.250
Earnings per month (pesos)              9,650    4,086
Hours worked                             41.9    13.15
Weeks waited for first job              24.47    33.09
Weeks worked                            64.14    56.09
Sector of economic activity
Agriculture                             0.018    0.105
Construction                            0.053    0.067
Commercial/financial                    0.426    0.465
Mining/industry                         0.117    0.276
Public services                         0.037    0.062
Social services                         0.269    0.317
Transport                               0.031    0.052



Table 5-65. Zero-Order Correlation Matrix among Selected Variables, Colombia, 1981 Cohort
Variable               (1)      (2)      (3)     (4)     (5)      (6)     (7)     (8)     (9)    (10)    (11)    (12)    (13)    (14)
1. Verbal ability              1.00
2. Mother's education          0.15     1.00
3. Father's education          0.17    0.59    1.00
4. Father's income             0.02    0.11    0.14    1.00
5. Father is employer          0.04    0.05    0.00    0.06    1.00
6. Academic achievement        0.42    0.12    0.12    0.03    0.03      1.00
7. Agricultural achievement    0.19    0.03    0.03    0.01    0.00    0.42    1.00
8. Commercial achievement      0.09    0.02    0.03    0.03    0.02    0.24    0.23    1.00
9. Industrial achievement      0.12    0.03    0.03    0.00    0.02    0.40    0.45    0.22    1.00
10. Earnings in 1982         -0.01   -0.02      0.03    0.01   -0.04     0.04    0.00    0.01    0.03      1.00
11. Hours worked               0.01    0.01   -0.02     0.03    0.01    0.01    0.04    0.01    0.01    0.17    1.00
12. Weeks worked               0.05    0.01    0.01   -0.03   -0.01   -0.01      0.01   -0.04   -0.05   -0.05      0.02     1.00
13. Looking for work           0.01    0.01    0.02   -0.01      0.01   -0.02   -0.03     0.01   -0.01     0.15    0.02   -0.02     1.00
14. INEM graduate            -0.06   -0.16   -0.12   -0.09   -0.04       0.04    0.11    0.11    0.17    0.02    0.01    0.01   -0.04        1.00



136  DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 5-66. Means and Standard Deviations
of Selected Variables, Colombia, 1978 Cohort
(Number surveyed = 1,826)
Standard
Variable             Mean  deviation
Background
Male                      0.521    0.499
Age                      21.859    1.864
Number of siblings        5.608    2.838
Father's years of schooling    9.775    3.302
Father's occupation
Farmer                  0.117    0.322
Laborer                 0.111    0.315
Employee                0.354    0.478
Professional            0.297    0.457
Business owner          0.086    0.281
Family income (pesos)    28,440   32,516
Urban born                0.875    0.330
School type and curriculum
INEM                      0.427    0.494
Academic                  0.250    0.433
Agricultural              0.123    0.329
Commercial                0.199    0.400
Industrial                0.202    0.401
Pedagogical               0.136   0.343
Social sciences           0.086    0.281
1978 destination
Studying                  0.262    0.440
Working and studying      0.261    0.439
Working                   0.364    0.481
Self-employed             0.025    0.158
Earnings                  5,405    3,748
Hours worked             39.054   14.235
Weeks to find job        14.157   17.784
1981 Status
Studying                  0.287    0.452
Working and studying      0.220    0.414
Working                   0.367    0.482
Self-employed             0.039    0.195
Earnings                 10,292    5,474
Hours worked             40.205   12.403
Looking for work          0.177    0.382
Weeks looking            27.254   30.596



Table 5-67. Zero-Order Correlation Matrix among Selected Variables, Colombia, 1978 Cohort
Variable                 (1)      (2)     (3)       (4)     (5)      (6)     (7)      (8)     (9)     (10)    (11)    (12)    (13)    (14)
1. Male                         1.00
2. Age                          0.09     1.00
3. Urban born                 -0.03   -0.16       1.00
4. Years city resident          0.04     0.08     0.23    1.00
5. Resident of big city         0.03   -0.16      0.09    0.15     1.00
6. City per capita income     -0.07   -0.27       0.18    0.14    0.78      1.00
7. INEM graduate                0.16   -0.06      0.07    0.09     0.13     0.22    1.00
8. Repeated primary school    -0.00      0.17   -0.01     0.09   -0.01   -0.02      0.07    1.00
9. Father is farmer             0.04     0.23   -0.31   -0.22   -0.23   -0.30   -0.13   -0.00         1.00
10. Father is employee        -0.01   -0.11      0.07    0.02    0.05      0.08    0.03   -0.03   -0.03       1.00
11. Student in 1981           -0.01   -0.28      0.07   -0.04      0.09    0.18    0.04   -0.06   -0.11       0.03     1.00
12. Worker in 1981              0.03     0.13   -0.01     0.00    0.06    0.06   -0.06       0.04    0.00     0.02   -0.32      1.00
13. 1981 earnings               0.16     0.06    0.02     0.02    0.11     0.11     0.04    0.03   -0.04      0.00     0.02     0.00     1.00
14. Private sector employee     0.08   -0.11     0.02   -0.01      0.19    0.23    0.16   -0.03   -0.13   -0.03        0.01     0.00   -0.02     1.00



6. The Tanzania Case Study
WITH A PER CAPITA INCOME of $3 DO a year, Tanzania faces far greater
constraints on its economic development than does Colombia, which is
semi-industrialized. Tanzania has an agrarian economy, and most people
live and work in the countryside. During the 1970s, gross domestic
product (GDP) increased at an annual average rate of 5 percent. With 51
percent of the population of working age, Tanzania's labor force grew
at an average yearly rate of 2.7 percent. In 1978 agriculture accounted for
53 percent of GDP (down 6 percent from 1960); manufacturing stood at 10
percent (up 4 percent from 1960); and the service sector stood at 36
percent (also up 4 percent from 1960). During the same period, agricul-
ture's percentage of the total labor force decreased by 6 percent, from 89
to 83 percent; industry's percentage rose from 4 to 6 percent; and services'
share increased from 7 to 11 percent.
Nonetheless, Tanzania remains heavily dependent on agriculture; more
than half of the farm product is still derived from subsistence farming. The
country has an abundance of labor and considerable uncultivated land,
but it is critically short of capital, both physical and human. As in most
other African nations, well-trained manpower in the modern sector has
been in short supply since independence.
Background of the Study
A major feature of Tanzania is its planned economy and socialist form
of production, which stress the public over the private sector, the opera-
tion of state enterprises, and government control of wages. Ever since the
Arusha Declaration and President Julius Nyerere's statement spelling out
in detail the concept of "Education for Self-Reliance" (see Nyerere 1967,
1977), political ideology has dictated the role of education in national
development. In order for decentralized socialism to succeed in abolishing
poverty, integrating a nation with a diverse tribal heritage, and achieving
national self-reliance, Tanzania's policymakers decided that education
had to accomplish two goals. First, it must impart useful knowledge and
skills; second, it must give a practica[ orientation to learning. The fact that
1.38



THE TANZANIA CASE STUDY                                    139
96 percent of all primary school students received no further schooling
suggested that primary education should be terminal-geared to the over-
all needs of a rural life and the manual skills suitable for rural employment,
and fostering attitudes favorable to manual work. Primary schools were
also expected to help support themselves by planting cash crops, raising
chickens, and vegetable gardening. (Mbilinyi 1974a; Dodd 1969; Boesen,
Madsen, and Moody 1977; Mmari 1976a, 1976b).
Diversification and Self-Reliance
The goals of secondary schools, however, were conceived very differ-
ently. At independence, Tanzania's acute shortage of skilled manpower
forced the educational system to train a local cadre as quickly as possible
to manage the public sector and replace the skilled expatriate work force.
Because of that need, policies were focused on secondary and higher
education, which historically had offered traditional academic curricula.
Tanzania's public secondary education system at first glance appears to
be elitist and highly circumscribed. Those who complete it are supposed to
be guaranteed appropriate jobs by the government, and the curriculum is
geared to preparing graduates for work in specific economic sectors. It
includes agricultural, commercial, technical crafts, and home economics
subjects.
Government policy attempted to fine-tune the supply of and demand for
educated workers so that the country could attain self-sufficiency in
qualified manpower. Because the state controls a large part of the econ-
omy and, hence, the demand for educated labor, the public school system
was permitted to expand only as much as the projected growth in demand
for educated labor.
The content of Tanzanian education, like its goals, is also influenced by
socialist philosophy. As expressed by President Nyerere, self-reliance re-
fers not only to self-sufficiency in the skills and knowledge needed to
achieve economic independence, but also to the attitudes and social values
needed to promote national development. For example, invidious distinc-
tions between town and country, or mental and manual work, were
considered counterproductive. Likewise, it was thought that schools
should not stress individual welfare, but, rather, cooperative effort and the
welfare of the group.
Educators were directed to encourage behavior and attitudes conducive
to the concept of self-reliance. Education was expected to stress the virtues
of rural life, and thus help to slow migration to towns. At the same time,
attempts were made to distribute government services more evenly
throughout the country. Schools were directed to play down any distinc-



140   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
tion between work and education: work is not for the uneducated, and
theory without practice is useless. Both academic knowledge and practical
experience in the local environment were to be considered of equal value.
Education's role in fostering social change thus came to be viewed as
critical. The curriculum became the centerpiece of an education policy
designed to link education to work, skills to jobs, and graduates to job
openings. Primary schools were to teach crafts and skills useful in the rural
community, where more than 80 percent of the children would eventually
live after leaving school. At the secondary level, diversified education was
to prepare students for work and for advanced training in broadly defined
economic sectors such as commerce, agriculture, and industry. As a con-
sequence of this policy, it would seem that no public secondary school
place would be created if manpowver planning did not warrant it.
Hence, a major test of whether a diversified curriculum prepares stu-
dents for their post-school experi znces is whether students trained in a
specific vocational curriculum are eventually placed in training and em-
ployment situations that utilize the skills they have acquired in school.
Educational System
A general view of Tanzanian education by scholars studying the system
has been that, at the time of inclependence in 1961, the country had
inherited "an underresourced school system, racially differentiated, urban
in bias, elitist in form and culturally distant from Tanzanians themselves,
which had to face up to two sets of potentially inconsistent demands. The
needs of the new state for manpower related to development goals had to
be balanced against the demands of parents for some tangible benefit of
independence through increased educational opportunities" (Williamson
1979, p. 115).'
Secondary education in Tanzania consists of a four-year cycle (Forms I
to IV), followed by a two-year cycle (Forms V and VI), as shown in figure
6-1. Tanzania has a very steep educational pyramid; for example, in 1980
only 4 percent of the 14-17 year-o:!d cohort was enrolled in Forms I to IV
in public schools (see figure 6-2 and table 6-1). The 4 percent secondary
school participation rate is in sharp contrast to the 25 percent participa-
tion rate in Colombia (see figure 5-2) and represents one of the major
differences between the two countries' systems of education. Stiff selec-
tion, based on merit tests, is in part responsible for the particular shape of
Tanzania's pyramid. Because of a strict one-to-one correspondence of
enrollment to job placement, secondary education is prevented from ex-
1. See also Austen (1968), Cameron and Dodd (1970), Court (1976), Furley and Watson
(1977), Mbilyini (1974b), Sanyal and Kinurtda (1977), Nyerere (1967), and Omari (1984).



THE TANZANIA CASE STUDY                                             141
Figure 6-1. Flow Chart of the Tanzanian Educational System
Lower secondary
Diversified
(Agricultural,
commercial,      Upper
Primary           technical, domestic   secondary    University
I  _1 I III I _vi  /       science)
Nondiversified
*|I |II III IVR
(Academic)
panding more rapidly than the number of suitable job openings. Unless
enrollment is controlled, policymakers fear that a glut of unemployed
educated manpower will result.
Students are selected to Form V on the basis of equity and merit. To
qualify for university entry, students must (1) complete Form VI with at
least minimum university requirements, (2) participate in a one-year
national service (military) training program, (3) work for at least two
years before joining the university (except for female and engineering
students, who have been exempted2) and (4) demonstrate a commitment to
serve the people in the course of their work.
During the four years spent in lower secondary school, students are
assigned to, or allowed to choose from, one of four vocational curricula:
commerce, agriculture, technical trades, and home economics. Vocational
courses supplement the academic course load, and the amount of time
devoted to all subjects-vocational and academic-is prescribed by the
Ministry of Education. Considerable time is spent on vocational subjects:
between 25 and 40 percent of the course-work periods a week are spent on
the vocational field of concentration during the fourth year alone.
Tanzania has proceeded far in institutionalizing vocational subjects
2. This is considered to be a temporary measure to compensate for the social and
education disadvantages suffered by Tanzanian women in the past, and to ensure that
expensive technical equipment and teaching capacities at the university are fully utilized
during the transition to the new system.



Figure 6-2. Educational Pryamid, Tanzania, 1961, 1971, and 1980
MALE                                     FEMALE
University (first year)
VI Secondary Form
V
IV
III
11
VII   rllllaly    Ld11a.....                                                  A
VI                                            X 7 ;.;
400    350    300    250    200    150    100        50       0      50    100    150    200    250    300    350
Enrollment
1980  1971  1961                                             (thousands)
Source: Tanzania, Ministry of National Education, Sectoral Planning Department.



Table 6-1. Enrollment Data for Public Schools by Grade and Sex, Tanzania, 1961, 1971, and 1980
Female
Male                1961             1971             1980
Total
Grade       1961    1971    1980   Number Percent Number Percent Number Percent  1980
University
(first year)     70      624      693       6     8        75    11        216   24         909
Secondary school form
VI              251    1,240    1,230      25     9       196    14        407    25       1,645
V               211    1,354    1,340      25    11       254    16        413    25       1,761
IV             1,121    5,274    6,550    482    30      1,170    25      2,770    25      9,320
III            1,540    5,375    5,908    548    26      1,947    27      2,737    32      8,645
II             2,514    5,426    5,827    1,019   29     2,197    29      2,785    32      8,612
1              2,967    5,554    5,887    1,229   29     2,016    27      2,960    33      8,847
Primary school standard
VIl           11,322   46,563  123,749    -      -      23,939    34     86,954    41    210,703
VI            13,061   54,011  210,534    -      -      31,256    37    167,445    43    377,973
V             14,737   63,300  246,579    -      -      39,308    38    209,329    46    455,908
IV            65,152   82,396  269,474  30,239    32    55,850    40   237,363    47    506,837
III           65,553   89,616  413,125  34,787    35    59,689    40    383,375    48    706,500
11            67,647   98,954  268,827  41,345    38    67,636    41    257,609    49    526,436
1             72,773  111,018  246,827  48,613    40    79,873    42    240,038    49    486,865



144   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
along with the academic curricuilum. Its diversified education system
produces, at the end of four years, students with a general knowledge not
only of academic subjects but also of a particular vocation. From there,
students can pursue more highly specialized training, either on the job or
in other formal training programs run by the government or private
institutions. The primary purpose of Tanzania's diversified vocational
system is thus not to prepare students for the university. Nor does the
system try to provide vocational training without a meaningful academic
education.
Sampling Procedures
With fewer than 200 secondary schools and 18,000 Form IV (twelfth-
grade, lower secondary) students in Tanzania, a random sample was made
of one-third of all schools and approximately 25 percent of all Form IV
students in the 1981 graduating class. The relatively large sample ensures
sufficient numbers for statistical comparisons among subgroups (such as
region, sex, employment in publicly or privately owned enterprises), and it
compensates for unavoidable attrition in follow-up surveys. Furthermore,
in order to compare schooling outcomes across the major vocational
curricula, it was necessary to ensure representative subsamples of the
school population: those coming from agricultural, commercial, and tech-
nical streams and a control group of nonvocational (academic) schools.
The home economics stream was not included in the study because it was
offered by only a few schools, and was not fully implemented even where it
was available. So that the schools chosen would be fairly representative of
schools throughout the country, a large ratio (1:3) of sampled to total
schools was adopted.
The sample was stratified by first selecting schools and then selecting
students within schools. Enrollment data for Form IV students attending
secondary schools around the country were obtained from Ministry of
Education lists. It was then possible to determine the overall size of the
population of the Form IV class of 1981 and the percentage studying in
each curriculum. Schools were selected according to their field of spe-
cialization and whether they were publicly or privately run. All technical
schools, which are the fewest in the nation, were selected outright. With
the help of Ministry of Education officials, random samples were made of
commercial, agricultural, and academic schools. Once each subcategory
of schools was established with a representative mix of schools by curricu-
lum, researchers proceeded to tesm: all Form IV students in each selected
school, unless the total number exceeded 125, in which case a random
selection of students was made.
The result was a sample of stucients and schools that closely approxi-



THE TANZANIA CASE STUDY                                        145
mated the distribution of the total student population among the commer-
cial, agricultural, technical, and academic curricula. The eventual sample
of more than 4,000 students represents about one-quarter of the eligible
population of Form IV students (table 6-2). Secondary schools in Tanzania
are widely scattered throughout the nation (figure 6-3); many are boarding
schools located far from towns. The twelve regions of the country sur-
veyed in this study account for three-quarters of all schools in the nation.
The uneven distribution of schools follows the location of early missionary
centers, where formal education was introduced (Ishumi 1974). (Descrip-
tive statistics of the Tanzania sample are provided in tables 6-51 and
6-52.)
Socioeconomic Profile of Form IV Students
Data gathered on selected mean characteristics of the sample of 4,181
students in the twelve regions confirm that the regions of Arusha, Kiliman-
jaro, Iringa, and Morogoro and the coastal regions of Tanga and Dar es
Salaam are advantaged areas, in the sense that their incomes per capita are
not in the lowest quartile of regional per capita incomes in Tanzania, and
that they have fewer mothers and fathers who are illiterate. Dodoma,
Mtwara, and Pwani are disadvantaged by the same measures of family
background (see table 6-3).
In large-scale studies, it is typical that measures of father's occupation,
income, and education significantly correlate. For secondary school stu-
dents in Tanzania, these correlations are also significant and indicate the
tendency for education to be linked with occupational placement in the
modern sector and wage employment.
The measurement of socioeconomic background or status is important
for two reasons. First, equity concerns demand that we test whether
differences in socioeconomic status arise across curricula. Second, to the
extent that students in Tanzania are differentially recruited to a given
Table 6-2. Sample Distribution by Curriculum,
Tanzania, 1981 Cohort
Curriculum  Students Schools
Academic   1,025    13
Agricultural  1,380    18
Commercial  1,284    19
Technical    492     7
Total      4,181    57



Table 6-3. Selected Characteristics of Students in the Sample by Region, Tanzania, 1981 Cohort
(percent)
Sample  Born in  Mother's  Father's  Father's  Father
size    same   education  education  income      a     Private  Regional
Region     (number)  region'    lowb      lowc      lowd   farmer' school    GDPg
Arusha            317      51       54         41        19       47      38       763
Dar es Salaam     428      45       51         35       19        24      24      3,838
Dodoma            267      18       74         58       33        51       8       415
Iringa            344      58       53         52       27        40      33       519
Kagera            213      33       76         55       31        49       7       504
Kilimanjaro     1,096      56       57         39       26        44      39       674
Mara               81      21       62         46        15       17      40       400
Morogoro          328      37       60         33       21        40      39       566
Mtwara            117      54       67         44       24        62       0       460
Pwani             245      27       75         49       28        60       0       617
Singida            99      19       56         36       25        51       0       439
Tanga             577      55       55         39       30        39      33       730
a. Percentage of students attending schools in the same region in which they were born.
b. Percentage of mothers with less than three years of education.
c. Percentage of fathers with less than three years of education.
d. Percentage of fathers in lowest quartile of income.
e. Percentage of fathers who farm for living.
f. Percentage of students attending private schools.
g. Gross domestic product (GDP) per capita in 1973 Tanzanian shillings.



THE TANZANIA CASE STUDY                                             147
educational stream on the basis of their socioeconomic backgrounds, and
to the degree that the various social groups differentially prepare their
children for successful competition in school, the results of schooling must
be corrected to allow for these out-of-school influences. In this way, we
can avoid attributing to schools alone, learning outcomes that are in fact
due to family circumstances.
In table 6-4, a set of common measures of family background, generally
accepted as indicators of social well-being, reveals some differences among
groups with respect to certain variables. For example, commercial stu-
dents have the most educated mothers in the sample (4.9 years), while
technical students have the least educated mothers (3.4 years). Students
Figure 6-3. Location of the Schools Surveyed, Tanzania
.   Schools in sample
Cities
t  - fLa.*c  |     @   National capital
f   UGANDA~~
- -  N- Region boundaries
B./.<  Etuk.ba.  t   -   ~-\   ---  International boundaries
\ zzES 1       M        . Muom
RWANDA  WT              C    MARA    N                  MILES
J or LAKE \   - qt  \--I           <sS      KILOMETERS
-  t~yn    \MWANZA g          w     
BURUNDI) es,A            J        ]       *          oh
A- 0Mo'shi
SHINYANGA       J       Arusha  \'
/~ ~ ~~_                       ARUSHAX\KIlMNlR
_J ~ ~ I          -
/.y                     x( )~~J  J  /N
KIGOMA   Q               Sing&da        )
TABORA    \  .           /        T
il   "y                  ?                        N\ *  GA
i- t\     <-')              %  $  l DODOAIA\                    IediaD
RUKWA         A                    7 * %50r050r0 /)
MBEYA     / .    ,, o    F
k 5 ~~~~~~~~LINDI f
ZAMBIA                     fjj+
'   MOZg RUVUMA          111WAU A*
(          ~~~MOZAMBIQUE



Table 6-4. Socioeconomic Characteristics of Secondary School Students by Curriculum, Tanzania, 1981 Cohort
Percentage of
fathers                             ~~~~~~~~~Percentage
Father's         fathers         Father's   Mother's   Number    of families
ncome       In ucage            educat:on  education       ? --L  - S' g Engli
Curriculum   (shillings)A  employment Farmers  (years)a   (years)'   at home'     at home
Academic       7,181         54         45        5.9       4.4         37           15
Agricultural    6,656        50         48        5.7       4.0         33           15
Commercial    7,834          50         31        6.6       4.9         44           23
Technical      7,088         60         47       5.2        3.4         33           14
a. Mean values.



THE TANZANIA CASE STUDY                                             149
from high-income families tend to go to commercial schools, while those
from low-income families are drawn to agricultural schools.
The questionnaire recorded how frequently English is spoken in the
home, because it was assumed that families communicating in English,
rather than in Kiswahili or some other East African language, probably
represented "elite" households. Twenty-three percent of the commercial
students come from families using English frequently, compared with 15
percent of students pursuing other courses.
Primary School Attrition and Secondary Selection
How do the socioeconomic backgrounds of incoming secondary school
students compare with those of primary school graduates who did not
obtain sufficiently high marks in their final examination to be accepted
into secondary school?
In Tanzania it was found that children of educated parents stand a much
better chance of getting into secondary school, even under a system of
meritocratic selection. On the basis of 1969 census data on population
distributions for paternal education, indices of the representativeness of
the sample of both public and private school students were calculated
(table 6-5). The indices show that the children of parents with higher levels
of education are overrepresented in public secondary schools by a factor of
six in relation to their percentage in the total population. The index is still
higher for private school children. If the distribution of parental education
in Tanzania has changed since the census, it has undoubtedly shifted
Table 6-5. Selectivity Indices for Students in Public
and Private Secondary Schools, Tanzania, 1981 Cobort
Selectivity
Father             Student         index'
In total    Public   Private
Education  population   school   school  Public  Private
(years)   (percent)  (percent) (percent) school school
(1)       (2)       (3)      (4)     (5)    (6)
0          70        11       26     0.16    0.37
1-4        20        37       22     1.85    1.10
5-8         7        40       32     5.70    4.60
8+          2        12       20     6.00   10.00
Source: Cols. (1) and (2): World Bank data based on 1969 census data. Cols. (3) and (4):
our study sample.
a. Selectivity index = percentage in school - percentage in total population.



150   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
upward, and this would imply slightly lower representation indices for
children of educated parents.
A second method of assessing the degree of selectivity of secondary
students is to compare the number of students from farm backgrounds in
the entire school-age population with those in the sample. Approximately
90 percent of all Tanzanians live and work in rural areas, and of that
percentage about 90 percent are engaged in strictly farm-related work. On
this basis, about 80 percent of all Tanzanian children in any age cohort
should have fathers engaged in farming. Normally, however, by the end of
primary school, disproportionate numbers of rural children will have
dropped out of school because the,  are needed for agricultural work. If 20
percent of rural students drop out before completing primary school, one
would expect about 75 percent of secondary school students to come from
homes in which fathers are farmers, if little selectivity by social class
occurred at entrance to secondary school.
In our secondary school sample, however, only 47 percent of the stu-
dents had fathers engaged in farming. By extrapolating backwards to the
number of farm children in the population who entered lower secondary
school (Form 1), we can be reasonably sure that students with a farming
background are underrepresented in secondary schools in proportion to
their number at the end of primary school, and that this reflects inequality
of selection by socioeconomic background.
Distribution of Students by Sex
Another way to measure equity is to examine sex distributions in the
student sample. Relatively few females are enrolled in all levels of the
education system. In the entire Form IV secondary school population, and
in our sample, only about 35 percent of the students are females. This is
mainly because historically fewer secondary schools were built for females
than for males. Within the sample, one can note whether the percentages
of females studying various curricula are similar or dissimilar, especially in
the public schools, where close to 35 percent of the students should be
female if the system is selecting equally across curricula.
Among all public school students, 32 percent are females, but across
curriculum tracks, the figures are quite skewed. Only 7 percent of female
students are enrolled in technical ccurses, compared with 24 percent in the
agricultural curriculum, 52 percen: in the commercial, and 38 percent in
the academic. There are also some wide disparities in the representation of
girls in the curricula of private schools (table 6-6).
Overall, female-to-male enrollment rates are far from the 50 percent
parity found in the population at large, but the ratios are similar to the
male-female ratio found in Standard VII classes (last year of primary



THE TANZANIA CASE STUDY                                             151
Table 6-6. Enrollment by Secondary Curriculum, Sex, Parental
Education, and School Ownership, Tanzania, 1981 Cohort
Enrollment        Father's
as percentage of    education
curriculum group     (years)
Curriculum and
school ownership    Male Female    Male Female
Academic
Public           62    38         5.2   6.4
Private          66    34         6.5   6.8
Agricultural
Public           76    24         5.4   7.1
Private          65    35         7.2   7.5
Commercial
Public           48    52         5.9   6.7
Private          56    44         6.5   7.3
Technical
Public           93     7         4.8   6.4
Private          99      1        6.0   7.2
school). A few disparities in these proportions can be found between
public and private schools, although by curriculum, much greater differ-
ences arise in commercial and technical programs (and undoubtedly
domestic science).
Public versus Private Schools
Still another way to look at Form IV enrollment variations is to compare
students attending public and private schools. Private schools in Tanzania
are generally not as popular as the state-run public schools. Public schools
have no fees, most offer boarding facilities, and, best of all, their students
have a very good chance of going on to Form V, a teacher training course,
or a government service job. Because entry into the public schools is
principally determined by an examination taken in the last year of primary
school, students accepted into state schools are already considered among
the country's elite.
Private schools, in contrast, are viewed as second-choice institutions for
those who cannot get into public schools. Children of affluent parents who
are not admitted into public schools attend these schools in disproportion-
ate numbers. Not only are the fees considerable, but students cannot
automatically expect government employment after graduation. Transfers
to public schools are impossible.
A comparison of public and private school enrollments indicates that, as
expected, private schools are attended by children of higher-income fami-



152   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
lies (table 6-7). As shown below, these students perform less well in
achievement tests than their public school counterparts, and this differ-
ence may partly explain why they are in private schools to begin with.
However, many children who do well in the primary school final examina-
tion are not admitted to public schools because Tanzania has a regional
quota system. Table 6-7 also indicates that private school students tend to
have more educated parents than public school students. Indeed, it is
commonly believed that educated parents who work in the modern urban
wage sector of the economy, but whose children do not gain entrance to
state schools, are more willing than other groups to pay for private
schooling in order to secure secondary education for their children.
Although there are no major social class differences across curricula,
certain types of schools may appeal to students of different backgrounds.
A cursory inspection of seventeen public schools judged to be "elite" in the
eyes of Tanzanian parents, on the basis of the quality of teaching staff or
the high proportion of graduates going on to Form V, showed that such
schools do not draw students disproportionately from socioeconomic
backgrounds of higher status. Me:'itocratic selection procedures appear to
have successfully negated social influences in these schools.
In summary, although distinctions of sex, region of birth, socioeco-
nomic background, and innate ability undoubtedly contribute to some
degree of educational inequality in Tanzania, it appears that the state-run
secondary school system is relatively free of social class bias.
Measures of Students' Ability
Two measures-one of verbal ability, the other of quantitative ability-
were used to test whether students of high innate intelligence fall dispro-
portionately into any one curriculum. Because Form IV students are by
definition a very select group to begin with, any differences in ability that
arise across curricula represent second-order variations. The ability tests
were designed to measure thinking with numbers and associated word
Table 6-7. Parental Income and Level of Education by Type
of School Students Attend, Tanzania, 1981 Cohort
Father's
Father's  Mother's  monthly
Type of educationi education income
school  (years)   (years)  (shillings)
Public    5.6     4.0     6,713
Private    6.8    5.3     7,594



THE TANZANIA CASE STUDY                                           153
problems and were intended for use as control measures for innate ability
(non-curriculum-based knowledge) and more general learning skills. (For
examples of test items, see Psacharopoulos and Loxley 1984.) The test
results reveal significant differences across the four curricula in favor of
technical students (table 6-8).
When further divided by sex, state or private school ownership, and
parental education, the results suggest that recruitment into the technical
track is the most selective. Possibly, the science curriculum demands more
proficiency with numbers, as well as with words, than do other curricula.
Males significantly outperform females on both measures, which is some-
what surprising because one would expect females-being underrepre-
sented in the school population-to have been more strictly selected for
ability.
State schools apparently do far better than private schools in recruiting
talent, according to our measures of aptitude. A primary reason for this is
that state schools are generally looked upon as high-caliber institutions in
Tanzania. The state-private school distinction also helps explain why
commercial schools appear to do most poorly: twice as many commercial
students as other students are enrolled in private schools, and their lower
mean ability as a group dampens the average ability scope of students in
the commerical curriculum as a whole. The differences in mean ability
between males and females and between public and private schools are of
the order of five points, or half the standard deviation (highly significant).
To complete the profile of Form IV students by curriculum, we asked
students a series of questions about their attitudes and opinions. Eighty-
three percent of all respondents, regardless of curriculum, had started
Table 6-8. Mean Ability Score by Selected Student Characteristics,
Tanzania, 1981 Cohort
Verbal Quantitative
Characteristic  aptitude  aptitude
Sex
Male            52       52
Female          46       46
School ownership
Public          52       52
Private         46       45
Curriculum
Academic       51        51
Agricultural    51       51
Commercial     43        48
Technical       52       54
Note: Test scores are standardized to a mean of 50 and standard deviation of 10.



154   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
primary school in the same year, thus, age in the sample is fairly homo-
geneous. With a similar degree of uniformity, 85 percent of all students
claimed to be studying the subject of their choice. Seventy percent of
agricultural and technical students and 80 percent of commercial and
academic students planned to apply for Form V. No significant differences
arose across curricula in students' own assessment of their academic
performance in secondary school. Forty-five percent of all students consid-
ered their schoolwork "above average," and 55 percent of all students,
regardless of curriculum assessed the quality of English spoken in their
classrooms as "above average."
Internal Efficiency of Schools
One of the more important Eunctions of diversified curricula is to
facilitate further training for occu Dational specialization and the transmis-
sion of cognitive skills that will later enhance the graduate's labor market
performance. Better cognitive performance in school should make a per-
son more productive on the job, even if productivity is not reflected in
higher wages and salaries because of a government wage cap. Thus, it is
important to examine the extent to which schooling in the different
curricula improves cognitive performance.
Cognitive Achievement
As discussed in chapter 4, special vocational achievement tests in com-
merce, agriculture, and technical trades were prepared by the National
Institute of Education at the University of Dar es Salaam. Test items were
based on questions closely tied to secondary school curricula. Reliability
for these tests was high enough (in the mid-0.60s) to warrant confidence
that the tests were indeed measuring knowledge in the specific field of
concentration. No academic achievement test was administered because
achievement scores in languages vEnglish and Kiswahili), math, and sci-
ence were already available from the Form IV individual examination
booklets.
Performance scores on four achievement tests, distributed by curricu-
lum, sex, and public or private school ownership are provided in table 6-9.
The scores are normalized to a mean of 50 and a standard deviation of 10.
Figure 6-4 displays in the form of a bar graph standardized performance in
these four subjects. Students do best in their own specialization; males do
better than females in agricultLtre, technical, and academic subjects,
although not in commerce; and public school mean scores exceed those of
private schools in all subjects.



THE TANZANIA CASE STUDY                                               155
Table 6-9. Mean Achievement Score by Selected Student
Characteristics, Tanzania, 1981 Cohort
Test score
Characteristic   Academic Agricultural Commercial Technical
Sex
Male             52         51          50         51
Female           47         49          51         47
School ownership
Public           52         52          51         52
Private          46         46          49         46
Curriculum
Academic         52         50          49         50
Agricultural     50         53          48         50
Commerdal        49         47          54         47
Technical        50         50          48         57
Note: Academic test refers to mathematics, Kiswahili, and English. All test score compari-
sons between groups are statistically significant at the 0.0001 level.
Boldface scores are in the students' own area of specialization.
Having observed mean test score differences across curricula, we em-
ployed regression analysis to see which groups performed better relative to
the academic control group, once a set of background characteristics were
controlled for. The findings for each of five achievement tests-three
vocational and two academic-are presented in table 6-10. The control
variables include sex, age, parental characteristics, and natural ability.
Each vocational curriculum dummy is based on the omitted comparator
group of academic-track students. The coefficients of the dummy variables
represent points gained or lost on a test relative to the reference group of
academic-track students.
In practically all cases, private school students scored significantly be-
low public school students. For each vocational test, the group trained in
the subject of that test did better than the others, as expected. The
regression findings simply confirm previous observations of the mean
scores, when out-of-school characteristics as independent variables were
not controlled for.
Effect of Course Time on Achievement
Having observed differences in test performance across the various
subjects, we examined the effect on achievement of variations in time
devoted to different courses. The measurement in this case is based on the
number of class periods per week attended by individual students in
Form IV.



156   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Figure 6-4. Academic and Vocational Test Scores by Curriculum,
Tanzania, 1981 Cohort
60                                                         57
52
50M       50450   50 505050
40-~~'
-5 30-
"a 20-
r.~ ~~~~~~~~~~~$
10-
Academic     Agricultural   Commercial      Technical
C U RRICULU M
Mean = 50
Standard deviation = 10
Achievement test
B Academic         Agricultural    Commercial      Technical



Table 6-10. Vocational and Academic Achievement as a Function of Curriculum
and Selected Socioeconomic Characteristics, Tanzania, 1981 Cohort
Vocational test score                         Academic test score
Independent         Agricultural       Commercial         Technical          English             Math
variable       Coefficient t-value  Coefficient t-value  Coefficient t-value Coefficient t-value  Coefficient t-value
Background
Male                    0.134     0.37   - 0.379    0.10     1.692a    4.93     2.413'    7.24    2.986'    9.38
Age                    - 0.140    1.09   - 0.317a    2.38   - 0.277'   2.32   - 0.636'    5.47   - 0.277'    2.49
Books at home         -0.002      0.37   -0.002     0.46   -0.008      1.87     0.20'    4.64     0.002     0.51
Urban born              0.480     1.20     1.793'   4.34     0.912'   2.44   -0.468       1.29   -0.911'    2.63
Father's education    -0.005      0.09     0.108a   2.01     0.002     0.04   -0.050      1.06   -0.065     1.45
Log of father's earnings    0.129  1.23    0.063    0.58     0.008     0.08   -0.112      1.17    0.032     0.36
Father self-employed    0.946a    2.84     0.821a   2.38     0.345     1.11     0.143    0.48     0.573a    1.99
Father farmer         -0.392      0.96   -0.089     0.21   -0.142      0.37     0.276     0.74   -0.730'    2.06
Father blue-collar      0.634     1.16     0.202     0.36    0.829     1.63   - 0.263     0.53   - 1.077a    2.28
Father white-collar     0.108     0.21     0.194     0.35   -0.260     0.53   -0.203      0.43   - 1.071'   2.35
Father professional     1.502'   2.69      2.137a    3.70    0.927     1.78   -0.024     0.05   -0.185      0.38
Verbal aptitude         0.227a   12.91     0.147'   8.08     0.196'   11.96     0.265'   16.63    0.198'   12.97
Math aptitude           0.108'   6.17      0.094a    5.19    0.233'   14.28     0.267a   16.88    0.432a   28.56
Private school         -2.961     8.09   -2.474'   6.54   - 2.402'   7.03   - 3.292'    9.92   -0.004       0.01
Curriculum
Agricultural            1.608'   4.11   -1.131a    2.79   -0.967'   2.65   - 2.240'    6.31   - 0.468       1.38
Commercial             -2.604     6.51     5.431a   13.11   -1.897'   5.07   -2.274a    6.26   - 1.559'    4.50
Technical              - 1.606'   3.05   - 1.138'   2.09     5.264a   10.69   - 4.049a    8.47   - 1.924a    4.21
Constant               35.880a  12.30    42.176'  13.96    33.872'  12.42    37.620'   14.21    23.377'    9.24
R2                          0.18               0.13              0.28               0.33               0.39
Sample size 3,903
Note: Curriculum dummy variables are evaluated relative to students in the academic track (omitted reference category). The occupational
groups listed are evaluated relative to those not listed.
a. Statistically significant at the 5 percent level or better.



158    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-11. Percentage of Time Spent on Core Courses,
by Curriculum, Tanzania, 1981 Cohort
Official  Student
Curriculum syllabus assessment
Academic    100       100
Agricultural    28     20
Commercial    32       27
Technical     42       26
Sources: Tanzanian Ministry of Education (1980), and Diversified Secondary Curriculum
Study.
The core courses in academic sub jects include civics taught in Kiswahili,
languages (Kiswahili and either English or French), sciences (chemistry,
biology, physics, and mathematics), and social sciences (geography and
history). Academic streams emphasize strictly academic material, and
students often get a heavy dose of English, math, and science.3 The core
courses taken by the vocational students depend upon the curriculum they
are following. Essentially, those in commerce take business education,
which includes commerce, economics, bookkeeping, typewriting, and
office routine. Commerce students often take additional social science
courses, such as geography and history. Technical core subjects include
mechanical, civil, and electrical engineering, designed to prepare students
for jobs as trainee craftsmen in these areas and, if they are able to pass the
examination at the Form IV level, for a technical college or the Form V
science stream. Agriculture students take core courses in the introduction
to agriculture, soil science, anima. science, crop production, and agro-
mechanics.
The official course syllabus for Form IV students dictates that technical
students spend nineteen periods a week on technical subjects, out of a total
of forty-five periods. The figures for agriculture are ten and thirty-six
periods, respectively; for commei-ce, about fourteen and thirty-nine.
Academic students take academic subjects exclusively. A comparison of
the survey figures and the students' own assessments of the time devoted to
different courses in Form IV is shown in table 6-11. Apparently, in the
Tanzanian secondary education system, wide deviations occur between
scheduled and actual course loads. These are probably a result of insuf-
ficient teachers or facilities, or only partial implementation of diversifica-
tion.
3. Although the official number of periods for the core curriculum is the same for all
curricula, including the academic track, students can take optional subjects such as addi-
tional mathematics, arts, foreign languages, and music.



THE TANZANIA CASE STUDY                                       159
Regressions were fitted separately for each curriculum using three
academic scores (math, English, and Kiswahili), with the score in the
curriculum for which the student was registered as the dependent variable.
Again, a host of out-of-school characteristics were included in the set of
independent variables to serve as controls so that any effect of sex, age, or
ability was removed before the impact of the specific course-time-exposure
variable was assessed. The findings for agriculture students are presented
in table 6-12; results for technical, commercial, and academic students,
respectively, are contained in tables 6-13, 6-14, and 6-15.
The findings reported in these tables highlight several important effects
of course-load distribution on learning. (Note that it is the amount of class
time specifically devoted to courses rather than the undifferentiated total
number of course hours that affects test scores.) First, in practically all
cases, related courses have a powerful and significant impact on the test in
the subject they complement. For example, in table 6-12, the figures show
that exposure to commerce and science has a positive effect on math
achievement. Second, it appears that in different curricula combinations of
courses affect learning in different ways, depending on the subject tested.
To use commercial students as an example, table 6-14 reveals that social
studies courses aid commercial students in learning math and English;
commerce courses, as expected, aid these students in learning commerce
skills, but apparently at the expense of English learning. Language and
social studies courses have some surprising effects on test performance.
For academic students (table 6-15) language courses do not aid learning in
English, Kiswahili, or math, whereas social studies courses affect perfor-
mance positively in these academic areas. For commerce students (table
6-14) a different pattern is revealed: language courses negatively affect
math performance and have no effect on commerce. For agriculture
students, language courses have a negative effect on agriculture and on
math test scores (see Karweit 1983).
Effect of Date of Implementation on Achievement
Do students attending schools that have a relatively long history of
offering a given curriculum specialization perform better than students in
schools that introduced it later? The survey data show that student scores
in schools that established a specialization before 1977 are indeed higher
than those of students in schools that began to diversify after 1977.
When exam date (pre-1977 = 0 and post-1977 = 1) and course periods
in individual vocational subjects are included in regressions on vocational
test scores by curriculum (table 6-16), we are able to observe more fully
how "time on task" affects learning. In table 6-16, each regression is run
only for students majoring in the given subject area. For example, in the



Table 6-12. Effects of Class Time on Achievement, Agricultural Students, Tanzania, 1981 Cohort
Test score
Independent           English         Mathematics         Kiswahili         Agriculture
variable       Coefficient t-value  Coefficient t-value  Coefficient t-value  Coefficient t-value
Background
Male                    2.654a    4.40     3.927'   6.44     4.998'   8.17   -1.539'   2.08
Age                    - 0.474'   2.39   -0.384     1.92   -0.486a   2.42       0.080    0.33
Books at home           0.019'   2.57      0.001    0.14     0.009    1.19   -0.003      0.32
Urban born             - 1.232    1.93     0.084     1.13   - 1.853'   2.86     0.220    0.28
Father's education      0.002     0.03   -0.144      1.82   -0.118    1.49      0.180    1.87
Log of father's earnings  -0.068  0.47     0.187    1.29   -0.039    0.27       0.472a   2.67
Father self-employed    0.327     0.64     0.730     1.41    0.044    0.08      0.160    0.25
Father farmer          -0.711     1.17   -0.608     0.99   -0.766    1.25       0.589    0.79
Father blue-collar      0.373     0.43   - 0302    0 34   -0.161    0.52       0.031    0.02
Father white-collar   -0.420      0.48   -1.610      1.83   -2.108a   2.39      0.604    0.56
Father professional    - 2.025'   2.29   -0.753     0.84   - 1.486    1.66      1.770    1.63
Verbal aptitude         0.212a    8.29     0.181a   6.99     0.173a   6.66      0.234a   7.43
Math aptitude           0.266'  10.24      0.433'  16.47     0.248'   9.43      0.087'   2.75
Private school         -5.458a    7.30   -1.802'   2.38   -6.992a   9.22    -3.382    3.69
Class time periods
Agriculture            -0.416a    4.11   -0.498'   4.86      0.086    0.83      0.749a   6.02
Commerce                0.174     1.46   -0.307a   2.54      0.278a   2.30      0.011    0.07
Language              -0.258      1.74   -0.160     1.07   -0.120    0.80   -0.448a   2.46
Science                 0.572'   8.12      0.405a   5.68     0.143'   2.00      0.174'   2.01
Social studies          0.069     1.01   - 0.094    1.37     0.449'   6.49    - 0.064    0.76
Constant               30.362     6.38    23.571a   4.89    32.289a   6.68    27.437'   4.70
R 2                          0.37              0.43               0.37              0.18
Sample size 1,252
a. Statistically significant at the 5 percent level or better.



Table 6-13. Effects of Class Time on Achievement, Technical Students, Tanzania, 1981 Cohort
Test score
Independent           English          Mathematics         Kiswahili          Technical
variable        Coefficient t-value  Coefficient t-value  Coefficient t-value  Coefficient t-value
Background
Male                     2.405    1.44      0.549    0.30      1.889    1.13    -0.180    0.10
Age                     -0.152    0.43    -0.730    1.93    -0.248    0.71    -0.469    1.23
Books at home            0.028a   2.05      0.020    1.30      0.023    1.64    -0.002    0.14
Urban born              -3.278'   2.04    -4.366'   2.50    -3.464'   2.15    -3.151    1.79
Father's education      -0.135     0,98    -0.109    0.73    -0.114    0.82    -0.075    0.50
Log of father's earnings  0.060    0.20     0.058    0.18      0.026    0.08    -0.005    0.01
Father self-employed     1.413    1.52    -0.257    0.25       1.253    1.35      0.514    0.50
Father farmer           - 1.639    1.56   -1.408    1.23    - 1.739    1.65       0.061    0.05
Father blue-collar      -0.148     0.11     0.498    0.33    -1.059    0.78       0.938    0.63
Father white-collar     -2.315     1.46   -3.846    2.23    -1.716      1.08    -2.893    1.67
Father professional     -0.216    0.11    - 1.738    0.82    -0.864    0.44       1.051    0.49
Verbal aptitude          0.213'   3.85      0.227'  3.78       0.173'   3.13      0.099    1.63
Math aptitude            0.220'   4.41      0.324'  5.96       0.289'   5.78      0.430a   7.88
Private school         -15.395'  11.04    -8.184'  5.39   -19.956'  14.29    -5.573'   3.65
Class time periods
Agriculture             -1.621     1.50    - 0.280    0.23     1.456    1.34    -0.399    0.33
Commerce                 0.399    0.29      1.100    0.75    -2.331     1.73    -0.687    0.47
Language                 0.052    0.17    -0.791'  2.39        0.680'   2.24      0.350    1.05
Science                  0.090    0.54      0.901'  5.04    -0.414'   2.51    -0.165    0.92
Social studies           0.032    0.08      0.572    1.30    -0.840'   2.08       0.225    0.51
Technical               -0.255    1.24    -0.070    0.81    -0.239    0.53        0.143    1.09
Constant                28.468'   3.40    31.938a  3.51       35.519'   4.24    38.634a   4.22
R2                            0.49               0.45               0.57               0.32
Sample size 457
a. Statistically significant at the 5 percent level or better.



Table 6-14. Effects of Class Time on Achievement, Commercial Students, Tanzania, 1981 Cohort
Test score
Independent           English          Mathematics        Kiswahili         Commercial
variable       Coefficient t-value  Coefficient t-value  Coefficient t-value  Coefficient t-value
Background
Male                    4.069a   6.906     1.765'   3.737    2.913'  4.938   -0.004    0.007
Age                    -0.832'  3.682   -0.175    0.970    0.917    0.873   -0.251    1.052
Books at home           0.026a  3.353      0.001    0.246    0.007    0.957   -0.008    1.011
Urban born             -0.857    1.360   -1.653'   3.272   -0.385    0.611   -0.205    0.309
Father's education     -0.134    1.441   -0.002    0.036    0.005    0.063      0.087    0.890
Log of father's earnings  0.066    0.336   0.145    0.918     0.166    0.839    0.363    1.735
Father self-employed   -0.545    0.974     0.318    0.709   -0.232    0.415   -0.383    0.648
Father farmer            1.591'  2.212   -0.514    0.893    0.486   O(675   -0 124    0.164
Father blue-collar     -0.483    0.532   - 1.282    1.759    0.046    0.051   -0.742    0.772
Father white-collar     0.703    0.870   - 0.008    0.013    0.534    0.661     0.524    0.614
Father professional     0.390    0.443     0.482    0.682   -0.285    0.324   -0.114    0.123
Verbal aptitude          0.218'  6.714     0.160'   6.141    0.173    5.322     0.127a   3.696
Math aptitude           0.191a  6.084      0.350'  13.872    0.170a  5.406      0.209'  6.285
Private school         -5.808'  8.975      1.716a   3.310   -5.677'  8.763   -2.158a   3.155
Class time periods
Agriculture              1.234    1.834    1.305'   2.419    0.770    1.143     0.310    0.436
Commerce               -0.207'  3.119      0.079    1.500   -0.065    0.982     0.246'  3.510
Language               -0.010    0.099   -0.248'   2.820    0.310   2.821      0.008    0.077
Science                 0.004    0.067     0.195'   3.624   - 0.221'  3.286   - 0.065    0.922
Social studies          0.440a  7.754      0.154'   3.401    0.102    1.811     0.061    1.020
Constant               40.779a  7.648    21.750a   5.089   27.588a   5.168    39.008a   6.919
R2                            0.36              0.39               0.24              0.13
Sample size 1,024
a. Statistically significant at the 5 percent level or better.



Table 6-15. Effects of Class Time on Achievement, Academic Students, Tanzania, 1981 Cohort
Test score
Independent            English         Mathematics          Kiswahili
variable       Coefficient t-value  Coefficient t-value  Coefficient t-value
Background
Male                    -0.100     0.16     5.025'    7.70   -3.449a   5.33
Age                     -0.198     0.93     0.041     0.18   -0.195      0.86
Books at home            0.016'    2.10     0.013     1.62     0.013     1.55
Urban born              - 0.991    1.44   - 1.035     1.40   - 2.344'   3.20
Father's education       0.091     1.04   -0.004      0.05     0.188a   2.01
Log of father's earnings   - 0.207  1.12   - 0.172    0.86   - 0.194    0.98
Father self-employed    -0.083     0.14     0.181     0.30     0.605    1.02
Father farmer            0.077     0.11   - 1.192     1.60     0.606    0.82
Father blue-collar      -0.660     0.71   -0.988      0.99   -0.136      0.13
Father white-collar     -1.601     1.71   -1.907      1.90   -0.818      0.82
Father professional      0.114     0.11   -0.987      0.95     0.965    0.94
Verbal aptitude          0.318'   11.03     0.166'    5.36     0.211'   6.90
Math aptitude            0.234'    7.77     0.432'   13.55     0.221'   6.90
Private school          -0.180     0.30     1.483'    2.30   -2.471'   3.87
Class time periods
Agriculture             - 0.179    1.26   - 0.281     1.84     0.210     1.39
Commerce               -0.134      0.69   -0.487'    2.32      0.614'   2.96
Language                -0.035     0.45   -0.142      1.68   -0.016      0.20
Science                  2.398a    9.75     1.286a    4.87     0.888'   3.39
Social studies           0.414'    6.12     0.390a    5.37     0.420'   5.85
Constant                24.138'    4.90    16.245'    3.07    27.598'   5.27
R2                             0.45               0.48               0.32
Sample size 922
a. Statistically significant at the 5 percent level or better.



Table 6-16. Influence of Exam Date and the Number of Course Periods on Learning
for Vocational Students, Tanzania, 1981 Cohort
Achievement score
Independent         Agricultural       Commercial           English           Technical
variable       Coefficient t-value  Coefficient t-value  Coefficient t-value  Coefficient t-value
Background
Male                     0.47'    2.5     -0.13      0.7       2.74'    6.0      0.21      0.4
Age                      0.01     0.6       0.03     0.9       0.99a    6.2      0.10      0.6
Household possessions    0.16a    2.3    -0.10       0.5       0.07     0.8    -0.10       0.5
Log of father's income   0.25     2.2       0.27'    2.5       0.01     0.7    -0.13       0.9
English spoken in home   -0.11    0.6    -0.13       0.8      -0.19     0.9    -0.02       0.8
Books at home          -0.00      0.5    -0.01       0.7       0.01a    2.5    -0.00       0.7
Father professional      0.17     0.8       0.13     0.9     - 1.60a    2.5      0.17      0.7
Father self-employed     0.09     0.7       0.09     0.8     -0 15      0.6      0.15      .9
Father white-collar    - 0.25     0.4       0.17     0.8     - 0.92     0.7    - 0.67      1.0
Father blue-collar     -0.19      0.5    -0.09       0.9     -1.70a    2.4    -0.07        0.8
Father's education       0.04     1.0       0.04     0.7     -0.07      0.5   -0.01        0.4
Mother's education     -0.05      0.2    -0.01       0.9     -0.19a    2.5    -0.02        0.7
Verbal aptitude          0.19'    9.0       0.09a    4.5       0.93a   16.0      0.11a    2.2
Math aptitude            0.10a    4.2       0.19a    7.7       1.13a   16.0      0.41'    8.5
Exam date after 1977   -0.02'    2.0    -0.06'    2.0        -0.03      0.5    -0.05 a    4.6
Class time periods
Agriculture              0.21'    8.0       -        -         -         -       -         -
Commerce                 -         -       0.05'     2.5       -         -       -
English                  -         -        -        -       -0.25      0.8      -         -
Technical                -         -        -                  -         -      - 0.05     0.6
Constant                 0.22               6.56            -32.9                2.37
R2                           0.175              0.132             0.306              0.280
Sample size                  1,380              1,284             4,181                492
Note: The reference occupational category is farmer.
a. Statistically significant at the 5 percent level or better.



THE TANZANIA CASE STUDY                                           165
agriculture equation, only the 1,380 students majoring in agriculture are
included. The same is true for the 492 technical students and the 1,280
commerical students. All commercial, technical, and agriculture students
are included with the academic students in the regression for English,
because everyone takes that subject. The exam date variable is significantly
negative for each of the three specialized biases; in other words, schools
that offered a specialization and were prepared to offer exams in that
subject prior to 1977 produce students with significantly higher scores in
their bias subject. Students in those schools do not have higher English
scores, however, and their academic performance is not significantly bet-
ter. The variable representing the course periods in the specialization is
positive and significant for agriculture and commerce, but not for techni-
cal and English achievement scores.
In summary, students attending schools which specialized before 1977
do significantly better in agriculture, commerce, and technical subjects.
Effect of School Quality on Achievement
To what extent do school expenditures and related quality indicators
improve test scores? A breakdown of selected private and public school
expenditures by curriculum is presented in table 6-17.
Achievement scores in vocational and academic subjects were regressed
on a block of school quality inputs, along with the previous list of out-of-
school independent variables. Table 6-18 presents the results, with scho-
lastic achievement in English, commerce, agriculture, and industrial arts
used as the dependent variables. As can be seen, school inputs differen-
tially affect learning outcomes. As the number of students per teacher
increases, test scores in English decrease. Teachers' educational levels are
significantly related to test scores in commerce, but not in other subjects.
Table 6-17. Annual School Expenditure per Student
by Curriculum, Selected Items, Tanzania, 1981 Cohort
(shillings)
Expenditure    Academic Agricultural Commercial Technical
Government-supplied
Teacher salaries  1,050      1,441      1,185     1,378
Materials           441       352        230       763
Family-supplied
Textbooks           172       156        210        150
Writing supplies    120        96        121       108
Fees                748       340        811       399
Note: Expenditures are for both public and private schools.



Table 6-18. Achievement as a Function of Background and School Factors, Tanzania, 1981 Cohort
Test score
ON                         Independent           Technical        Agricultural        Commercial           English
variable       Coefficient t-value  Coefficient t-value  Coefficient  t-value  Coefficient t-value
Background
Male                     2.031a    4.80     0.477    1.02       0.321    0.68      2.385a    5.78
Age                    -0.255      1.82   -0.271     1.76      -0.345a   2.23    -0.447a    3.27
Books at home          -0.003     0.67      0.006    1.17       0.005    0.89      0.018a    3.63
Urban born             -0.010     0.02      0.231    0.45       1,419a   2.79   -0.613      1.36
Father's education     -0.055     0.98   -0.118      1.90       0.024a   0.38      0.007    0.13
Log of father's earnings   -0.107  0.97   -0.059     0.49      -0.157    1.30   -0.020      0.18
Father self-employed     0.278    0.77      0.952a   2.41       0.853'   2.15     0.619     1.77
Father farmer          -0.215     0.49    -0.208     0.43       0.027    0.05   -0.185      0.43
Father blue-collar       0.783     1.34     0.442    0.69       0.338    0.52   -0.666      1.17
Father white-collar      0.038    0.06      1.079    1.69       0.069    0.10   -0.295      0.52
Father professional      0.708    1.10      1.781a   2.53       2,158a   3.06     0.009    0.01
Verbal aptitude          0.190a   9.75      0.220a  10.29       0.126a   5.89      0.250a  13.17
Math aptitude            0.233a  12.37      0.137a   6.65       0.091'   4.40      0.262'  14.27



School
Curriculum
Agriculture          - 2.275a   5.15    - 0.385    0.79     - 4.616a   9.49    -2.802a   6.51
Commercial           - 1.923a   3.52    -3.261a   5.45       2.861a   4.76    -3.936a   7.40
Technical              4,937a   8.02    - 2.877a   4.25     - 3.000a   4.42    - 5.4S3a   9.08
Private school         - 1.753a   3.59    - 2.54a   4.77       0.092    0.17    - 1.575a   3.31
Teacher education       0.122    1.36   -0.135       1.37      0.357a   3.62     0.095     1.09
Per student expenditure  0.74E-03a   4.56  0.55E-03a   3.07    0.29E-04    0.16  0.16E-03    1.05
Student/teacher ratio  -0.007     0.28   -0.046     1.68      -0.027    0.97   -0.119a    4.87
Teacher salaries       _0.OO1a   3.41    -0.001a   2.47  -0.51E-04    0.11  0.27E-03    0.69
Academic periods        o.O15a   1.99      0.008    0.97      -0.001    0.21     0.003    0.46
Constant               31.997a   8.68    42.304'  10.46       41.767a  10.29    34.243'   9.53
R 2                           0.32              0.19                0.15               0.37
Sample size 2,803
Note: The occupational groups listed are evaluated relative to those not listed.
a. Statistically significant at the 5 percent level or better.



168   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Teacher salaries and per studeni: expenditure are inversely related to
technical and agricultural scores. The time devoted to academic subjects
has a significant effect on technical scores only. In sum, there is no clear
overall pattern for the effect on achievement of common indicators of
school quality.
A convenient method of evaluating the effectiveness of the vocational
programs is presented in table 6- 19. This table shows that higher mean
scores are obtained by vocational students on measures of vocational
learning than on academic subjects. These scores are derived from the
regression equation after adjustment for background characteristics and
measures of ability (see table 6-10). In all cases, the gains in vocational
learning are obtained at the expense of some English language achieve-
ment. Because the government's policy objective is to develop vocational
skills, however, it may be concluded that the curriculum design does meet
its intended goal.
Modernity Attitudes
As discussed in previous chapters, schools are believed to influence the
values and attitudes of students and make them more "modern." Such
changes toward modernity are judged to contribute to national and eco-
nomic development. To measure psychological modernity, students were
asked to agree or disagree (on the Lickert scale4) with hypothetical be-
havioral statements relating to a variety of day-to-day situations. These
items were intended to elicit choices that would identify persons as either
"'traditional" or "modern" in their value orientation and outlook. For
example, the following statement asked students to agree or disagree:
"Success depends more on luck than on hard work." Presumably, less
traditional individuals will believe less in fate and more in the results of
their own labor and thus view change as something they can control.
Another statement reads: "If I were given a choice of 20 shillings today or
40 shillings next month, I would talce my 20 shillings today." The modern
individual might be more inclined to disagree with the statement, deferring
present consumption for future greater investment. A final example states:
"The only people one really can trust are one's family and relatives.>'
Modern individuals probably are more inclined to view outsiders as
trustworthy. In all, twenty-one questions were asked. The coefficient of
reliability for the sample was equal to 0.75, which indicates that the items
seemed to measure a single concept of what we call individual modernity.
4. A Lickert scale measures the respondent's reaction to a question along a continuum
from less to more: (1) disagree strongly, (2) disagree, (3) no opinion, (4) agree, (5) agree
strongly. The scale is extensively used in opinion and attitude surveys.



THE TANZANIA CASE STUDY                                                169
Table 6-19. Achievement Score Gains: Vocational Curriculum
Compared with Academic Control, Tanzania, 1981 Cohort
Achievement
Curriculum     test      Gain
Technical    Technical     5.26
English      - 4.05
Commercial  Commerce       5.43
English      -2.27
Agricultural  Agriculture  1.61
English      - 2.24
Academic    Commerce   -2.74
Technical    - 4.09
Agriculture   -2.24
Note: All achievement score differences are statistically significantattheS percentlevel or
better. Background and ability are held constant.
The mean score of the modernity scale by selected characteristics of the
education system is presented in table 6-20. Technical students enrolled in
public schools outscore all others. Although not shown in the table,
differences in parental education among students in the various curricula
were not significantly associated with differences in modernity.
Of course, the question remains whether the modernity attitudes preex-
isted or were formed by the type of school attended. In order to control for
ability and other out-of-school characteristics, a regression was fitted with
modernity score as the dependent variable and the various curricula as a
Table 6-20. Mean Modernity Score by Curriculum
and Type of School, Tanzania, 1981 Cohort
Curriculum and      Mean
school ownership modernity score
Agricultural        5.9
Public            6.1
Private           5.0
Technical           6.3
Public            6.5
Private           5.6
Commercial          5.6
Public            6.0
Private           5.0
Academic            5.8
Public            6.0
Private           5.4



170  DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
series of independent dummies. The results show that, after controlling for
background characteristics, very few modernity differences survive among
curriculum programs, although private school students score significantly
lower than public school students on modernity (table 6-21).
When students are compared on the basis of whether they wish to
pursue formal education, those desiring to enter Form V are much more
modern than those who do not expect to continue with their studies. This
is equally true for all curricula.
School Experience and Expectations
Do curriculum programs differentially alter a student's view about
occupations or his expectations about the job he will eventually take? As
vocational students invest more and more time in specialized subjects, they
Table 6-21. Individual Modernity as a Function of School
and Out-of-School Characteristics, Tanzania, 1981 Cohort
Independent
variable        Coefficiency  t-value
Background
Male                     0.323      0.84
Age                    -0.008    -0.06
Books at home            0.015 a    3.00
Urban born             -0.728    -1.74
Father's education     - 0.046    -0.85
Log of father's earnings    0.029   0.27
Father self-employed   - 0.791a   -2.28
Father farmer            0.186      0.43
Father blue-collar     -0.641   -1.13
Father white-collar    - 0.500    -0.91
Father professional    -0.996    -1.71
Verbal aptitude          0.199%   10.89
Math aptitude            0.120a    6.62
Curriculum
Agriculture            0.529      1.29
Commercial             0.669      1.60
Technical              0.973      1.77
Private school         -1.453a   -3.80
Constant                34.0ola   11.20
R 2                           0.11
Sample size                  3,903
Note: Academic bias is the omitted reference category.
a. Statistically significant at the 5 percent level or better.



THE TANZANIA CASE STUDY                                      171
may come to realize that those special subjects are of limited value outside
certain occupations. One might expect these students to be prepared to
continue with their specialized course of training so as not to lose the
advantage they have built up in the course work already completed. In
contrast, students taking only academic courses have invested most of
their time in general subjects that facilitate entry to many kinds of employ-
ment. The expectation is that these students would therefore be less likely
to tie themselves to any specialized occupational field.
In the survey, students were asked to examine a list of sixteen occupa-
tions and to rate each in terms of its importance in Tanzania. Ratings
varied from very important to unimportant. For some students, impor-
tance may have implied personal social prestige, rather than value to the
nation, but in every case students rated occupations on the basis of their
own perceptions. The occupations were chosen from existing lists used to
tap occupational expectations elsewhere in Africa (see Foster 1966) and
were hierarchically ranked from professional-managerial, clerical, and
skilled labor down to unskilled manual work. A few new items were
introduced so that the scale fit more neatly with the occupational catego-
ries found in Tanzania (such as, party worker and youth worker).
Differences in how Tanzanian students within the various tracks viewed
the importance of specific occupations is reported in table 6-22. Because
most of the findings are of only borderline statistical significance, it is
likely that students in the sample, regardless of their curriculum, viewed
the Tanzanian occupational structure in much the same way. It appears
that clergy, sales persons, and youth leaders were not highly valued, while
most students, regardless of track placement, listed agricultural scientist,
nurse, and school teacher as highly valuable and useful occupations.
Depending on the school enrolled in or the course of instruction, stu-
dents may view their own prospective careers differently. For example,
more than 70 percent of all students (72 percent of agricultural, 71 percent
of technical, 81 percent of commercial, and 82 percent of academic)
planned to apply for admission to Form V (upper secondary school), even
though fewer than 20 percent of all Form IV graduates would eventually
be accepted (table 6-23). Sixty-six percent of agricultural students planned
to apply for employment training upon completion of Form IV. Compara-
ble figures for technical, commercial, and academic students were 69
percent, 55 percent, and 53 percent, respectively. Although the differences
are relatively small, agricultural and technical students appear to be
slightly more interested in pursuing careers in their training fields than in
pursuing traditional academic course work in Form V. Academic students
planned to apply for Form V in large numbers (which is to be expected,
given their stream's close link to the university curriculum), but many
planned to apply for jobs and nonuniversity training programs as well,
perhaps as a form of insurance in case they were not accepted into Form V.



172    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-22. Percentage of Studenzts Asserting that a Given
Occupation Is of Importance for Tanzania, 1981 Cohort
Curriculum
Occupation       Agricultu}ral Technical Commercial Academic
Agricultural scientist    87         82         82         84
Auto mechanic             57         65         53         59
Building trades           34         55         34         40
Clergy                    19         19         22         31
Customs inspector         40         44         47         44
Farm manager              62         53         52         54
Government clerk          44         35         45         44
Lawyer                    55         55         62         61
Nurse                     79         65         77         80
Party worker              46         53         47         S2
Police Officer            66         59         70         67
Primary school teacher    76         74         72         76
Sales worker              24         22         31         27
Secondary school teacher  76         78         74         75
University professor      69         74         69         68
Youth leader              38         39         38         35
Note: Figures refer to percentages responding that the occupation is of "above average
importance" and "very important."
With regard to the choice of sectlor of employment (public sector exclu-
sively, versus parastatal or private), little difference was evident across
curricula, with more than 50 percent of all respondents planning to seek
public sector employment.
Students also were asked to select from a list of thirteen job training
areas their first, second, and third choices for a career after graduation.
The list included two agricultural choices (farming and veterinary train-
ing), three commercial (secretarial training, hotel management, and book-
keeping), three technical (electrical, mechanical, and automotive training),
and two academic (teacher and cultural officer training), along with some
miscellaneous categories (tailoring, food technology, and day care), which
were used as distractors. These choices were then grouped to reflect the
four broad fields of agricultural, commercial, technical, and academic
pursuits, so that it was possible to compare student occupational choices
with their curriculum program.
As first choice, agriculture students selected their own field 53 percent of
the time, technical students pickedl their specialization 75 percent of the
time, and commerce students chose theirs 59 percent of the time.
Academic students seemed to scatter their choices more widely; only 22
percent wished to pursue such traditional academic courses as teacher
training or cultural affairs training (table 6-24).



THE TANZANIA CASE STUDY                                             173
Table 6-23. Post-Graduation Plans and Expectations,
Tanzania, 1981 Cohort
(percent)
Curriculum
Characteristic        Academic Agricultural Commercial Technical
Plans to apply to
Form V                           82        72          81        71
Employment training              53        66         55         69
Plans to enter
Self-employment                  22        19          16        14
Salaried employment               5         6           6         4
Salaried employment in
Government service               60        69         59         61
Parastatal                       34        25         32         35
Private sector                    6         6           9         4
Expectations
Starting salary (shillings a month)    852  824      833        856
Weeks to get a job               23        23         21         21
Like the Colombian secondary school cohort examined in chapter 5,
Tanzanian students were asked to state their earnings prospects with and
without a Form IV education. The perceived value of a Form IV education
indicates how realistically students view education's worth in the labor
market.
Expected mean earnings by curriculum bias are listed in table 6-25. The
first column refers to expected earnings if students dropped out of Form IV
and immediately went to work. The second column refers to expected
earnings from employment on the assumption that students would com-
plete Form IV a few months after the survey was made. The difference
between columns 1 and 2 represents the value attributed to a Form IV
certificate. The third column refers to what students hoped to be earning in
five years' time. (In every case students were asked to discount for effects of
inflation.) There are few differences in expected initial earnings across
curriculum programs, a finding undoubtedly due to the government's
egalitarian pay policy.
Comparing Costs with Outcomes
The unit costs for Tanzania's secondary schools are more easily calcu-
lated than was the case for Colombia. In Tanzania, each school has only
one area of specialization, and hence expense accounts can be differenti-
ated. The assumptions underlying the cost comparisons are briefly de-



Table 6-24. First, Second, and Third Career Choices by Curriculum, Tanzania, 1981 Cohort
(percent)
Career choice
Academic            Agricultural         Commercial            Technical
Curriculum First Second Third First Second Third First Second Third First Second Third
Academic    22      24      34    33     26      21    25      28     24    20      22      21
Agricultural  12    17      41    53    45       17    18    20       21    17      18      21
Commercial  12      19      26    14     22      16    59    45       31    15      14      17
Technical    11     16      35      6      8      6      8     13     17    75    63        42
Note: This table summarized thirteen specific job choices in four general areas; students from each curriculum ranked three of the four. Figures
in boldface indicate choices in the students own field of specialization.



THE TANZANIA CASE STUDY                                         175
Table 6-25. Expected Monthly Earnings by Curriculum,
Tanzania, 1981 Cohort
(shillings)
Without    With Form IV certificate
Form IV
certificate Upon graduation In five years
Curriculum    (1)       (2)          (3)
Academic     798         852         1,763
Agricultural   735       824         1,675
Commercial   746        833          1,831
Technical    776         856         1,763
scribed below, and the detailed calculations are to be found in Hinchliffe
(1983).
Unit Costs by Curriculum
Estimates of forgone earnings, relevant to both private and social costs,
have been based on students' expected earnings without a Form IV di-
ploma. Other data in the study suggest that students generally overesti-
mate earnings by about 20 percent. As a result, we have reduced the stated
levels of earnings forgone by that amount. The final figure of 624 shillings
a month is consistent with findings of a 1980 survey of wage employment
in Tanzania, where twenty-one- and twenty-five-year-old primary and
secondary school leavers were earning a monthly average wage of 596 and
745 shillings, respectively. (Special tabulations from the survey are de-
scribed in Knight and Sabot 1981.) Annual per student capital costs, have
been calculated directly from World Bank appraisal reports for education
diversification projects in Tanzania. Recurrent school costs were reported
in some detail on the questionnaires. Data on private expenditures by
school curriculum and ownership (state or private) and by item (text-
books, fees, and so forth) were available, as were data on school expendi-
tures for salaries, materials, and room and board for both state and private
schools.
Annual private and social costs by school curriculum are presented in
table 6-26, while table 6-27 converts a number of the cost sets into indices,
with the unit cost of the academic track set at 100. Total social costs of
each type of school are dominated by forgone earnings; and differences in
these according to curriculum are not statistically significant. The varia-
tions in total social costs by curriculum therefore appear small-the
average for vocational schools is only 3 percent above that for the
academic schools. In contrast, total recurrent costs in state schools are 14



176  DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-26. Annual Average Cost of Secondary Schooling
per Student, Tanzania, 1981
(shillings)
Curriculum
Item         Agricultural Technical Commercial Academic
Recurrent cost
Borne by individual
State schools          749        854       1,114       1,112
Private schools      1,328      1,129       1,556       1,065
Private school fees  1,987      1,981       1,841      2,226
Borne by rest of society
State schools        2,700      2,409       2,046       1,776
Private schools             -                -          -
Total recurrent cost
State schools        3,449      3,263       3,160       2,888
Private schools      3,315      3,110       3,397       3,291
State schools
Total direct costa     5,798      5,612       5,509       5,237
Total social costb    13,290     13,104      13,001      12,729
-Not applicable.
a. Including capital costs.
b. Including forgone earnings.
percent higher in the diversified schools than in those offering only
academic subjects. In the case of government-financed recurrent costs, the
diversified schools are on average 34 percent higher.'
Whether differences in costs between diversified and nondiversified
schools are regarded as significant depends on the type of cost considered.
Total social costs show little variation. The implications for public ex-
penditure, however, are substantial.
Cost-Effectiveness of Cognitive Achievement
To compare the costs and achievements of different curricula, the same
procedure was followed as in Colombia. The comparison is limited to
5. Secondary school costs have also been estimated by the Tanzania Ministry of National
Education (1979a). These estimates differ in major ways from those based on our question-
naire, for two main reasons. First, the ministry figures cover a much wider set of components,
some of which, it can be argued, are incidental to teaching. Second, ministry figures are all
accounting costs, whereas those used in the present study reflect social opportunity costs.
When these two factors are taken into account, the two estimates of the costs of commercial
schoots are virtually the same; the differe ace is 8 percent. For technical schools, however,



THE TANZANIA CASE STUDY                                                177
Table 6-27. Indices of Annual Private and Social Costs per Student
for State and Private Schools, by Curriculum, Tanzania, 1981
(academic cost = 100)
Curriculum
Item             Agricultural Technical Commercial Academic
Private direct expenditure
State schools                 67         77         100        100
Private schools              102         95         106        100
Private direct cost
State schools                 67         76         100        100
Ptivate schools              125        106         146        100
Government financed
Recurrent cost               152        136         115        100
Recurrent and capita] cost    122       115         106        100
Total recurrent social cost
State schools                119        112         109        100
Private schools              101         95         103        100
State scbools
Total recurrent and capital
social cost                  111        107         105        100
Total social cost
(including forgone earnings)  104       103         102        100
public schools, which have the most relevance for government policymak-
ers. Because most secondary schools would continue to be academically
oriented if diversification had not been introduced, the academic stream
has been treated as a control group to contrast with vocational students.
Table 6-28 presents average annual recurrent costs by curriculum pro-
gram, along with achievement scores in both academic and vocational
subjects. Test score means have been adjusted to remove the influences of
innate ability, sex, age, and other nonschool factors potentially influencing
educational achievement.
Annual costs per student are higher in all vocational schools than they
are for the academic control group: 13 percent higher for technical stu-
dents, 19 percent for agricultural students, and 9 percent for commercial
students. How do such increased costs translate into achievement im-
provements?
The adjusted mean achievement scores in table 6-28 have a mean of 50
substantial differences remain; ministry estimates are 40 percent greater than those obtained
from our questionnaire.



178    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-28. Recurrent Costs and Achievement Scores
in Public Schools, Tanzania, 1981 Cohort
Cost      Adjusted mean achievement score
diffe;'ence,                 Vocational
Annual  diversifiedl   Academic
unit cost   corstrol              Agri-  Com- Tech-
Curriculum    (shillings)  (per5ent)  Math  English  culture merce nical
Academic (control)   2,888    --       51     52     52     50    52
Agricultural       3,449      + 19     52     52     54     -      -
Commercial         3,160     + 9       49    S1      -      55    -
Technical          3,263      +13      55    51      -      -      58
Note: Means are for public school students only, and adjustments have been made to
remove the effects of ability and out-of-5chool characteristics on achievement.
Scores in boldface are those for students' area of specialization.
and a standard deviation of 10. An advantage of ten points means that 34
percent of the group in question outscore the control group (see figure
5-5). For example, agriculture students scored 54 on the agriculture test,
while academic students averaged only 52. The two-point difference in
favor of agricultural students corresponds to a 0.2 difference in standard
deviation and indicates that about 7 percent more agricultural students
than academic students score above the sample mean. For 19 percent more
resources per student, agricultural schools are able to boost agricultural
achievement by about 7 percent; at the same time, math achievement is
raised by 3 percent, and English achievement is not raised at all. For the
extra resources committed, gains i n agricultural knowledge seem modest,
while the academic skills acquired are neither better nor worse than those
obtained by academic students in the control group.
Comparisons can also be made between technical and control students.
For 13 percent more resources, technical schools can boost technical
achievement by 0.6 standard deviation, or 23 percent-a rather large
impact. In return for the additioncal resources, math achievement is raised
by 16 percent, though English achievement drops by about 4 percent. The
technical stream appears to be cost-effective in raising technical and math
achievement levels, at virtually no loss to English skills.
Students in commercial schools cost 9 percent more on average to
educate than those in the control group. In return, commercial achieve-
ment is raised by 0.5 standard deviation, or about 19 percent. Although
commercial schools appear to sacrifice academic skills for commercial
skills, the loss is small.
Thus, it appears that technical schools, though costing more than
control schools, yield a substantia.l increase in both academic and voca-



THE TANZANIA CASE STUDY                                        179
tional knowledge over that of the academic control group. Likewise, both
agricultural and commercial students gain in vocational knowledge as the
result of higher per student costs, but they show little or no increase in
academic knowledge relative to the control group.
The Class of 1981 One Year Later
Of the 1981 base-year sample (4,181 students), a total of 2,426 re-
turned the follow-up questionnaire distributed one year after gradua-
tion-a response rate of nearly 60 percent. This response rate was ex-
pected when the original 1981 sample was designed. Although it appears
low by U.S. or European standards of survey research, the figure is highly
satisfactory by developing nations' standards, particularly given the con-
ditions in Tanzanian rural areas (long distances, poor roads and trans-
portation, limited postal service, and the movement of students back and
forth between rural and urban areas).6
Response Rate
Because fewer than two-thirds responded, however, it is essential to
determine whether there were any differences in the characteristics be-
tween those who did and did not respond. Response bias arises when the
missing cases are not a random subsample of the original sample. If
missing cases are randomly distributed by key sample characteristics,
however, the omission is of little consequence because it simply leaves a
smaller but still random sample of the original population.
To detect any systematic response differences, selected characteristics of
the base-year sample, follow-up respondents, and follow-up nonrespon-
dents were compared. According to table 6-29 there are no significant
differences between respondents and nonrespondents based on sex, apti-
tude, curriculum bias, and father's education or employment. Slight differ-
ences arise regarding public school graduates and the urban born. The
latter, who constituted only 18 percent of the 1981 base-year cohort,
responded to the 1982 follow-up in greater numbers than expected. But
when response rates by region are compared, only three regions (the Dar es
Salaam area, Dodoma, and Arusha) significantly deviate from their ex-
pected response rates.
The discrepancy in the Dar response rate is mitigated by the fact the
capital contained less than 10 percent of the 1981 base-year sample to
begin with. The nine remaining regions, including the two most populous
6. A Land Rover was secured to trace sample participants in the countryside.



180    DIVERSIFIED  SECONDARY EDUCATION AND DEVELOPMENT
Table 6-29. Base-Year Cohort, Follow-up Respondents,
and Nonrespondents by Selected Background Characteristics,
Tanzania, 1981 Cohort
(percent)
1981        1982     1982 non-
Characteristic        base sample  respondents respondents
Curriculum
Agricultural                     32.8        32.2        33.9
Technical                        11.6        11.5        12.1
Commercial                       30.2        30.6        30.9
Academic                         25.3        25.6        23.1
Male                               65.7        66.0        67.8
Public school graduate             72.3        74.3        69.8
Urban born                         17.9        19.8        15.6
Applied to Form V in 1981          76.9        76.2        77.5
Top quartile scorer (math aptitude)  25.0      25.9        23.7
Top quartile scorer (english)      25.0        26.5        23.5
Top quartile scorer (math)         25.0        25.6        24.3
Father secondary school graduate   14.9        16.2        13.9
Father professional/manager        10.1        10.7         9.3
Father farmer                      41.9        41.6        42.3
in the sample (Kilimanjaro and Tanga), had response rates in the 1982
follow-up identical to those observed in the 1981 base year.
Post-Form IV Opportunities
The major post-Form IV avenues available to graduates are:
* Entry into Form V-a preferred option because it is likely to lead to
university and a professional career
* Placement in teacher training or other public sector training, which
may lead to a career in a government ministry
* Salaried employment or self-employment in the private sector
* Looking for a job (particularlyr a teaching, clerical, or technical job in
the public sector-very rarely in farming) or further training
* Being voluntarily inactive.
The last two groups might include individuals who scored low on their
Form IV examinations, private school graduates who were never led to
expect public sector employmen: in the first place, or those who for
personal reasons never desired further training or immediate employment.
As mentioned previously, Tanzania's secondary school system is tied
closely to estimated needs for manpower in the public sector. Originally,



THE TANZANIA CASE STUDY                                    181
students admitted into public secondary schools were expected to enter
public sector or parastatal employment upon graduation. With the expan-
sion of secondary education, however, that objective has been reviewed,
and it is now understood that public school graduates will also work in the
private sector or be self-employed.
Even on its present scale, the secondary education system has not been
able to place all its graduates. (Placement is defined here as government
assignment of students to Form V, further training, or full-time public
sector employment.) Ntuah (1981) reports that only 10 percent of Form
IV graduates in 1977-78 obtained suitable further education or employ-
ment in fields related to their secondary education. In addition, he found
that 31 percent of the 1977 cohort were unemployed two years after
graduation, and that more than 43 percent of the 1978 cohort had not
found jobs or further training one year after graduation. Ministry of
Manpower statistics reveal that from 1978 to 1981, more than one-third
of all Form IV graduates were never placed (posted to Form V, given
teacher or other public sector training, or employment). A fair number of
those not placed undoubtedly came from private schools, whose students
are often not considered in government manpower forecasts and whose
standards are generally lower than those of public schools.
Graduate Destinations
By comparing the destinations of graduates from diversified and control
schools, it is possible to see whether the diversification of secondary
education has had any impact on graduates' career opportunities. As in the
case of Colombia, we asked whether diversification:
* Reduces private demand for upper secondary education
* Raises the probability of postgraduate labor force participation
* Reduces unemployment and the time spent seeking work
* Raises the initial earnings of those working
* Raises the social rates of return to investment in schooling.
Answers to these questions help assess whether diversification of the
secondary curriculum has its intended effects, at least during the early
career of graduates.
The first question in the 1982 follow-up survey asked students to
indicate the activity which most nearly described how they spent the
preceding month-whether they were in school, further training, salaried
employment, or self-employment or were looking for work or a place in
further education. Only one individual marked the category of "other"
than the above activities. Clearly, virtually all 1981 graduates saw their
future in terms of competitive employment, absorption into government



182    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-30. The 1982 Activity of Form IV Graduates,
Tanzania, 1981 Cohort
(percent)
198:! Form IV curriculum
1982
activity    Academic Agricultural Commercial Technical Overall
Form V             27.5       25.3       34.2       33.3     29.5
Training           45.0       45.1       35.9       39.4     41.6
Working            14.1      -3.7        14.2       17.4     14.4
Looking for work    13.4     ;.5.8       15.7        9.8     14.5
Total             100.0      100.0      100.0      100.0    100.0
service, or further education. Form  IV graduates represent a select com-
petitive group who are not likely to choose to remain without a job or a
place in further education or trainling.
Table 6-30 shows the distribution of Form IV graduates among activi-
ties, and figure 6-5 illustrates their various destinations by curriculum.
Curriculum appears to make no major difference to graduates' initial
destination. Technical students, however, are more likely than others to be
employed and less likely than others to seek schooling or training or be
unemployed (not shown in figure). Consequently, one can infer that with
the possible exception of technical students, all Form IV graduates stand
nearly the same chances of landirng a job.
Figure 6-5. Distribution of Respondents by 1982 Activity,
Tanzania, 1981 Cohort
Diversified
(all vocational curricula)             Control (academnic)
Taning                                TraininA
40%           Working                               Worin
g            /    ~~~~15%                          /      4 
Looking                                Looking
15%1%
\   Form V  \      /        \/          Formo V   \      /
30%             \/<                         28%  



THE TANZANIA CASE STUDY                                                 183
The 1982 activities of Form IV public and private school graduates by
curriculum are listed in table 6-31. Private school graduates appear to
have been placed in almost equal proportions to the public school gradu-
ates, which implies that government recruitment procedures have not
discriminated against them.
Table 6-32 relates graduates' current activities to their achievement
scores in school. In general, graduates who are self-employed or are
looking for work, schooling, or training-all of whom constitute the
residual of unplaced-do not appear to have been low achievers.
Table 6-31. Graduate Destination by Form IV Bias
and School Ownership, Tanzania, 1981 Cohort
(percent)
Form IV school
Form IV curriculum
and 1982 activity    Public  Private
Academic
Form V               28      27
Training             44      46
Salaried employment    8     10
Self-employed         8       2
Looking for training  8       6
Looking for work      5       9
Agricultural
Form V               28      20
Training             46      48
Salaried employment    9      7
Self-employed         5       7
Looking for training  7      13
Looking for work      7       5
Commercial
Form V               36      37
Training             45      29
Salaried employment    7     13
Self-employed         7       8
Looking for training-   8    12
Looking for work      8       5
Technical
Form V               33      35
Training             38      45
Salaried employment   15      6
Self-employed         4       4
Looking for training  5       6
Looking for work      4       4



184    DIVERSIFIED  SECONDA:1Y EDUCATION AND DEVELOPMENT
Table 6-32. Achievement Score by Curriculum and 1982 Activity,
Tanzania, 1981 Cohort
Vocational
Academic                             Non-
Form IV curriculum                   Agri-   Com-  Tech-  verbal
and 1982 activity   Math  Eiglish  cultural mercial nical Aptitude
Academic                51     53       52      51     51      51
Form V                52     53       54      51     53      52
Training              51     53       51      50     50      50
Salaried work         52     50       53      52     52      50
Self-employed         51     55       48      48     50      49
Looking for training    52   53       51      51     51      51
Looking for work      52     50       54      52     52      51
Agricultural            51     51       53      48     50      51
Form V                52     51       53      48     51      52
Training              51     51       53      48     50      51
Salaried work         49     51       51      46     49      50
Self-employed         49     51       52      47     49      48
Looking for training    51   50       54      49     49      51
Looking for work      52     53       54      46     49      54
Commercial              47     47       47      54     47      47
Form V                46     47       47      54     47      47
Training              48     49       47      53      48     48
Salaried work         47     45       48      54     47      47
Self-employed         46     47       46      56     44      48
Looking for training    47   47       45      56     46      46
Looking for work      47     46       48      54     47      48
Technical               53      50      50      48      57     55
Form V                52      51      50      48      56     55
Training              51      50      51      49     56      53
Salaried work         57      54      51      47     59      58
Self-employed         53      54      52      50     58      57
Looking for training    52    49      47      49     54      52
Looking for work      56      48      48      48     52      56
Overall                 50    .50       50      50     50      50
ENROLLMENT IN FORM V
At completion of Form IV in 1981, well over 80 percent of students
reported they had applied to Form V. As noted, only a small fraction of
Form IV graduates have sufficiently good examination results to be
selected for Form V. To what extent does the probability of going on to
further education relate to the typo of school attended and its specializa-
tion?
The follow-up survey of the 1981 cohort provides information on the
way students were actually selected into Form V-whether on the basis of
high academic achievement, as measured by Form IV examination results



THE TANZANIA CASE STUDY                                               185
Table 6-33. Form V Major Subject by Form IV Curriculum,
Tanzania, 1981 Cohort
(percent)
Form IV curriculum
Form V
subject    Academic Agricultural Commercial Technical Overall
Academic       71.6       79.3        79.6       81.8     77.9
Agricultural'   1.2        5.3         4.6        0.0      3.4
Commercial      9.9       10.6         9.6       10.2     10.0
Technical      17.3        4.8         6.3        8.0      8.7
Total         100.0      100.0       100.0      100.0    100.0
Note: Figures in boldface indicate students who continued to study in their Form IV field of
specialization.
a. Refers to those enrolled in the MATi and LITI agricultural and livestock training
institutes.
and the course content, or of nonschool factors such as sex, age, and
socioeconomic background.
The follow-up survey asked those who were currently attending Form V
to list the courses they were taking. Such courses were clustered into three
major fields: technical (civil or electrical engineering), commercial (com-
merce or principles of accounting), and agricultural. An agricultural
stream in Form V was not officially adopted until the 1983-84 school year,
so the students polled by the study were those taking courses in the
training institutes run by the Ministry of Agriculture and the Ministry of
Livestock.7 All students not in these three fields were considered to be in
the "academic" stream.
The survey included a cross-tabulation of pre- and post-Form IV sub-
jects (table 6-33). Seventy-eight percent of all those in Form V took
academic subjects (math, history, languages, and social and physical sci-
ences) exclusively and were evenly distributed by Form IV curriculum. Of
the remaining 22 percent of respondents, students pursuing technical
courses in Form V were drawn mostly from the Form IV academic track,
and many of those majoring in agricultural sciences had majored in
commerce in Form IV. Even though diversification was not fully im-
plemented at the time of the survey, the continuity between Form IV and
Form V subject specialization seems rather loose.
We examined test scores for graduates of each curriculum in two
categories: those who stayed in their same Form IV field of specialization
7. Although these courses are not strictly equivalent to Form V level, graduates of them
receive a diploma in the appropriate field, just as the Form VI leavers do.



186   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
and those who moved into acadernic streams. Among agricultural students
math, English, Kiswahili, and agriculture scores did not vary between the
two groups. Presumably, student:s who pursue some agricultural courses
in Form V are selected for reasons other than examination results in that
subject, since only 5 percent of those who finished Form IV and had above
average achievement scores in agriculture stayed in that field in Form V.
For the two groups of commercial students, there were no differences in
math, English and Kiswahili scores. Students who remained in commerce
after Form IV, however, obtained significantly higher scores in commer-
cial achievement tests. Thus, staying in commerce is related to better
scores in that subject-which means that recruitment into Form V com-
mercial subjects is based in part on achievement in Form IV commercial
studies. Among technical and academic Form IV students, no differences
emerge on test scores between those who pursued academic and non-
academic careers in Form V.
Among Form V students, 58 p,-rcent said they were studying the curri-
culum of their first choice. This statistic varies by field: agriculture, 56
percent; commerce, 79 percent; technical, 62 percent; and academic track,
51 percent.
Form V students were also asked about their occupational expectations.
Seventy-one percent wanted to pursue professional or managerial careers
rather than jobs in agriculture, technical fields, and commerce. By field, the
figures are 68 percent for agriculture, 40 percent for commerce, 100
percent for technical, and 77 percent for the academic bias.
An examination of the number of academic and vocational courses
taken in lower secondary school. by students who went on to Form V
reveals no significant differences. This is also the case for students' nonver-
bal ability.
The selection of students for Fcrm V appears to be essentially unrelated
to the subjects they studied in lower secondary school, with the possible
exception of commercial students. More than three-quarters of Form V
students pursued academic courses-whatever their Form IV course may
have been. The remainder, who specialized in nonacademic courses, did
not differ significantly from academic students in terms of innate ability,
socioeconomic background, scholastic attainment in Form IV, occupa-
tional expectations, or forgone and expected earnings. In short, although
ability, achievement, social class, and other characteristics may be an
important determinant of obtaining a place in Form V, once accepted, a
student's choice of a particular curriculum program does not appear to be
contingent on past performance or previous field of study.
Table 6-34 presents schooling expenses and expected labor market
outcomes among those in Form 'V. Again, no major differences arise on
any of the characteristics. The rather lengthy (thirty-six weeks on average)



THE TANZANIA CASE STUDY                                               187
Table 6-34. Monthly School Expenses, Expected Earnings,
and Weeks to Find Employment by Students Enrolled in Form V,
Tanzania, 1981 Cohort
Current
Expected Expected schooling Earnings
Form IV   weeks to  earnings  expenses  forgone
curriculum  find work (shillings) (shillings) (shillings)
Academic      31      2,116      245      603
Agricultural    37    2,534      224      715
Commercial    38      2,491      251      611
Technical     36      2,263      205      568
Overall       36      2,381      236      632
waiting period to obtain employment is described as "normal" by Tanza-
nia government officials because of the time necessary to process the
examination results and place the graduates.
FURTHER TRAINING
The group of nearly a thousand Form IV graduates placed in training
programs (including teacher training) constituted 42 percent of the 1982
follow-up respondents. When this group is subdivided according to Form
IV curriculum and the graduate training program, a pattern emerges as
shown in table 6-35.
Nearly half of the group was engaged in teacher training, most of them
coming from Form IV agricultural and commercial curricula. Those who
entered teacher training did not come disproportionately from the
academic track. The most interesting result in this table is the small
proportion of commerical graduates and the high proportion of technical
Table 6-35. Further Training in 1982 by Form IV Curriculum,
Tanzania, 1981 Cohort
(percent)
1981 Form IV curriculum
1982 training    Academic Agricultural Commercial Technical Overall
Agriculture              7.5        5.4        11.1        6.7      7.6
Clerical occupations    10.2        8.4         5.6       21.2      9.5
Primary school teaching    47.9    55.8        59.1       45.2    53.3
Other                   34.3       30.4        24.2       26.9    29.5
Total                  100.0      100.0       100.0      100.0   100.0



188   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
graduates who took clerical training courses. This would seem to imply
that post-Form IV clerical courses are not seen as adding to a commercial
student's job opportunities, but that a substantial proportion of technical
students who opt for further train, ng expect to find openings in the clerical
field. Finally, the proportion of agricultural graduates taking agricultural
training courses is slightly lower than the proportion of graduates of other
curricula studying agriculture.
There were no sex differences by type of further training. But, when
public and private school graduates were compared, it was found that
public school graduates were more likely than their private school coun-
terparts to continue in the field they studied in Form IV.
Occupational expectations of tIrainees revealed that the largest propor-
tion (44.4 percent) expected to become teachers. Eleven percent expected
to enter clerical occupations, with technical and academic students prefer-
ring such roles.
Two cther questions were designed to elicit how contented respondents
were with their post-Form IV trair[ing program. On the question whether,
given a choice, graduates would prefer to work, be in school, or continue
in training, 70 percent expressed satisfaction with training, 25 percent had
a desire to go on to Form V, and only 5 percent would have preferred to be
employed. No significant differences were found in such opinions by Form
IV track or between public and private schools. However, slightly more
males than females preferred workc (7 percent compared with 4 percent).
On the question whether, given the choice, respondents would prefer a
form of training different from their current one, 35 percent expressed a
desire to change; agriculture students had the highest percentage (41
percent) wishing a change, and technical students the lowest (25 percent).
Respondents did not vary in their opinions according to sex and public-
private school background.
When questioned about the number of weeks they expected to wait
before obtaining employment and their expected earnings from initial
employment, students evidenced no significant differences by Form IV
curriculum (table 6-36). To an outside observer, the expected waiting
period appears unusually long, but this is accepted as normal because
Form IV graduates must await placement by the government.
EMPLOYMENT
Nine percent of the respondents had found salaried employment one
year after graduation, and another 5 percent were self-employed. A com-
parison of their jobs and earnings allows some conclusions about whether
one curriculum offered any advantage over another to graduates in the
labor market. Because self-employment and salaried employment differ



THE TANZANIA CASE STUDY                                        189
Table 6-36. Labor Market Expectations of Those
in Training in 1982, Tanzania, 1981 Cohort
Expected                  Forgone
number  Expected Monthly  earnings
of weeks  monthly  training  while in
Form IV   to find   earnings  expenses  training
curriculum employment (shillings) (shillings) (shillings)
Academic     31      1,264    236      690
Agricultural    35   1,279    231      673
Commercial   32      1,175    222      680
Technical    36      1,172    247      769
Overall      33      1,254    229      683
substantially, the two groups are treated separately. And because the
government sets wages in Tanzania, those in public employment are
distinguished from those in the private sector.
Of the graduates in salaried employment, 48 percent were in govern-
ment, 12 percent were in private, and the rest were in parastatal employ-
ment (table 6-37). By economic sector, 29 percent were in commerce, 9
percent were in agricultural jobs, and 17 percent were in manufacturing,
with the remaining 45 percent in all other sectors. Most respondents had
found their jobs after waiting six months. They worked, on average,
forty-four hours a week and earned 726 shillings a month before taxes.
When these workers were studied on the basis of their Form IV curricu-
lum, we found that 20 percent of agricultural students went into private
sector employment and that 57 percent of technical graduates went into
government work, as did 55 percent of commercial graduates. Thus, it
appears that agricultural students are not as readily hired into the public
sector as students from other specialties. Moreover, agriculture graduates'
private sector activities do not necessarily include farming because one-
third of them work in commercial or secretarial occupations.
Among all respondents in salaried employment, half are in public sector
clerical jobs, with no differences by Form IV curriculum or public-private
school background. Females, however, are twice as likely as males to end
up in private sector employment. No significant differences emerge in
earnings and hours worked by curriculum bias.
The self-employed were working on average 42 hours a week and
earning twice as much as those engaged in salaried employment. Twenty-
three percent worked on their own farms, and another 48 percent were
engaged in family agricultural help. Those who owned their own farms
worked significantly fewer hours a week than those who worked on farms
as laborers-38 and 44 hours, respectively. The remaining 30 percent



190    DIVERSIFIED  SECONDARY EDUCATION AND DEVELOPMENT
earned their living from personal services, odd jobs, or nonfarm
businesses. This group waited longer than the others before deciding to
enter self-employment (35 as opposed to 28 weeks). Self-employed com-
mercial graduates earned substantially less than those coming from other
specialties. Females were significantly less likely to be earning their living
from "kibarua," or odd jobs, i:han were males. On measures of the
socioeconomic  background, innrate ability, and  modernity  of self-
employed graduates no differences were found by Form IV track.
In the light of government equalizing wage controls, it is of interest to
examine how respondents perceived their earnings prospects as they pro-
gressed through the various educational, training, and employment pro-
grams, and what they in fact earned as secondary school graduates in the
public and private sectors of the economy.
Table 6-38 presents expected and actual earnings for respondents in the
1982 follow-up on the basis of their Form IV curriculum. Reports of
students' expected earnings before and at the time of Form IV graduation
were obtained from the base-year survey, while the expected earnings cited
in the follow-up survey are on the assumption that any schooling or
Table 6-37. Selected Characteristics of Those Employed
in 1982, Tanzania, 1981 Cohort
Form IV curriculum
Characteristic      Academic Agricultural Commercial Technical Overall
Economic sector and employer
(percent)
Government                    49        34          55        57       48
Parastatal                    40        46          35        37       40
Private sector                11        20          10         6       12
Agriculture                    S        12           7        10        9
Manufacturing                 10        18          25        14       17
Commerce                      30        40          20        21       29
Administration and services   55        30          48        55       45
Salaried employment           57        60          60        69       62
Self-employment               43        40          40        31       38
Those in salaried employment
Weeks waiting for job         26        29          27        33       28
Hours worked                  42        44          45        42       44
Monthly earnings (shillings)    762    736         710       699      726
Those in self-employment
Weeks waiting for job         39        38          32        20       35
Hours worked                  44        38          39        42       42
Monthly earnings (shillings)   1,532  1,676      1,252      1,560    1,504
Sample size                     83       101         100        45      329



THE TANZANIA CASE STUDY                                           191
Table 6-38. Expected Monthly Earnings at Three Levels
of Education and 1982 Actual Earnings by Sector of Employment,
Tanzania, 1981 Cohort
(shillings a month)
Expected earnings           Actual earnings
With
post-                 Self-
Form IV   Without   With  Form IV  Public Private  employ-   All
curriculum  Form IV  Form IV  studies  sector  sector   ment  sectors
Academic    792      883    1,252   746   787    1,532   1,065
Agricultural   755   830    1,325   722    776    1,676   1,100
Commercial   754     852    1,356   697   820    1,252    932
Technical    773     844    1,247   702    580    1,560    892
Average      766     852    1,307   716    777    1,504   1,013
training activity in the interim will have been successfully completed.
Actual earnings were computed from follow-up data. Expected and actual
earnings differed somewhat by Form IV curriculum, but with only a low
level of statistical significance.
LOOKING FOR SCHOOLING OR WORK
About 7 percent of the respondents were looking for full-time schooling
or further training and another 6 percent were looking for work. These
respondents constitute a residual group who had neither been placed in
any public sector activity nor found a job in the private sector at the time of
the follow-up. Only one person wished to remain inactive. The ratios of
males to females and public to private school graduates in this residual
category are similar to those who were already placed in further training
or jobs (tables 6-39 and 6-40). Therefore, no particular group is dispro-
portionately looking for work or training.
Of those looking for work only, 65 percent were looking exclusively for
government employment, and 29 percent were seeking employment in
parastatals; only 6 percent were interested in private sector work. The
attraction of government employment varies little across Form IV curric-
ula, sex, or public-private school affiliation, which further suggests that
those still looking are actually hoping for placement in government jobs.
The notable exception is technical school graduates, who seek private
employment in greater proportion to other students. About half of those
looking for work were interested in secretarial or accounting positions,
regardless of their curriculum in 1981. All but 15 percent started looking



192   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-39. Characteristics of Those Looking for Schooling
or Training in 1982, Tanzania, 1981 Cohort
Form IV curriculum
Aca-   Agri-  Com-  Tech- Over-
Characteristic       demic cultural mercial nical  all
Sex (percent)
Male                           61    67      40   100   59
Female                         39    33      60      0   41
Form IV school (percent)
Public                         80    75      48    79   67
Private                        20    25      52    21   33
Family income (shillings a month)  5,934  7,841   7,280  6,040 7,048
Sample size                     48    62       58    16  184
for jobs sometime in 1982, although all the private school graduates in the
group (24 percent) started looking as early as 1981.
Of those looking for work 70 percent were supported by families and 30
percent by odd jobs. Among those supported by families, there was a slight
preponderance of Form IV agricultural graduates, males, and private
school graduates. On average, they expected to wait eleven more weeks
before finding a job, which 50 peront described as secretarial or clerical in
nature. The minimum acceptable nonthly salary averaged 957 shillings.
There were no significant differences among job-seekers with respect to
ability and family background.
Likewise, among the 7 percent looking for training a year after graduat-
ing from Form IV, little difference could be discerned between males and
females, public and private school graduates, or curricula. Twenty-three
percent were looking for training in clerical-commercial subjects, 17 per-
cent in agriculture, 14 percent in trades, 10 percent in teaching, 8 percent
in a profession, and 27 percent training in other occupations.
It is interesting that those still loDking for work put "minimum accept-
able earnings" at a much higher level than the salaries of their counterparts
who were already working. The data are summarized in table 6-41,
together with the amounts that those who were in training and in school
expected to earn once they entered the labor market.
As a group, those looking for work or training did not vary by aptitude
or social class from the entire 1981 sample, and within the group the
variations in achievement as mea,ured by national examination scores
were also similar to those found for the entire 1981 sample: technical
students did better in math, academic and agricultural students did better
in English, and students in each curriculum did best in their respective



THE TANZANIA CASE STUDY                                                 193
Table 6-40. Characteristics of Those Looking for Employment
in 1982, Tanzania, 1981 Cohort
Form IV curriculum
Aca-   Agri-   Com-  Tech- Over-
Characteristic         demic cultural mercial nical  all
Sex (percent)
Male                              77     79        47    92    69
Female                            23     21        53     8    31
Form IV school (percent)
Public                            66     89        68    83    76
Private                           34     11        32    17    24
Expected sector of employment (percent)
Public                            97     96        91    82    93
Private                            3      4         9    18     6
Weeks looking for job
In current search                 32     32       28    16    31
Additional expected               11     11         9     8    11
Reservation wage (shillings a month)   859    944    1,051   958   957
Source of financial support (percent)
Family                            70     73        70    64    70
Other sources                     30     27       30    36    30
Sample size                         35     53       47    12   147
specialty. In conclusion, those graduates who did not immediately go on to
Form V, training, or employment do not noticeably differ from other
graduates on the basis of their Form IV curriculum.
UNEMPLOYMENT
One of the main goals of diversification was to provide sufficient train-
ing so that graduates of agricultural, technical, and commercial curricula
would experience a shorter period of unemployment than they might
otherwise have encountered after graduating from an academic school.
However, the data in table 6-42 do not show that this has happened. One
year after graduation, 13 percent of academic students were still looking
for either work or training while the percentages for technical, commer-
cial, and agricultural students were 10, 16, and 16, respectively. Academic
graduates did not believe that they had to wait any longer for some activity
than other students. Form IV graduates in Form V or in training courses
also showed no differences by curriculum in the period they expected to
have to wait before obtaining work once they entered the labor force.



194   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-41. Actual and Expected Monthly Earnings
of Form IV Leavers, Tanzania, 1981 Cohort
(shillings)
Form IV curriculum
Monthly earnings   Academaic Agricultural Commercial Technical
Actual in 1982a           762       736         710       699
Expected
Those looking for work  859        944       1,051       958
Those in training      1,264     1,279       1,175     1,172
Those in school       2,116      2,534      2,491      2,263
a. Those in salaried employment.
Predicting Post-Form V Elestinations
Having examined several descriptive dimensions of the sample, we
turned to the question of how school and background characteristics
affect the decision to study, work, obtain further training, or look for
work. What characteristic most encourages a move into upper secondary
education? Is it high grades, the type of lower secondary school attended,
or the previous curriculum track? As noted above, 27 percent of graduates
in the follow-up survey entered Fotrm V. We wanted to know whether test
scores in the vocational secondary school curriculum helped students to
get accepted into Form V. It is especially interesting to see how, in a
centrally monitored system such as Tanzania's, merit, school type, curric-
ulum, and socioeconomic background affect the eventual destination of
graduates.
To do this, we applied logit analysis, using control variable for family
background, sex, age, scholastic achievement, vocational scores, curricu-
Table 6-42. Unemployment Rates in 1982 and Expected Periods
of Job Search, Tanzania, 1981 Cohort
Form IV curriculum
Item           Academic Agricultural Commercial Technical
Unemployment rate (percent)    13     16          16         8
Expected weeks to find work
Those looking             10         11          9         9
Those in Form V           38        34          36        38
Those in training         31        35          32        36



THE TANZANIA CASE STUDY                                    195
lum, and measures of school quality, all of which might potentially affect
the probability of entry into Form V. The findings reported in table 6-43
reveal that six variables affect entry into full-time study (statistical signifi-
cance at the 0.10 percent level or better): being male, coming from a
commercial-track high school (either private or public), coming from a
school with a lower level of English usage, having scored higher on the
technical test, and having scored lower on the modernity test. Test scores
in academic subjects do not influence entry, and the technical achievement
score barely affects selection into Form V (t = 1.6).
Access to Training
More than 40 percent of the 1981 Form IV graduating class went on to
further training, primarily teacher training. Using a logit model, we again
found only six variables that significantly predicted access to further
training (table 6-44): attending a public agricultural school, attending a
school with high-quality resources, coming from a curriculum program
with few vocational courses, scoring high in English achievement or
individual modernity, and having a less-educated mother. Such factors
increased the chances for selection into government-sponsored training.
At the margin, higher English and modernity scores, along with an agri-
cultural curriculum, helped students get into training program. Surpris-
ingly, more vocational course work hurts rather than helps students in
gaining entrance to further training.
Entry into Salaried Employment
About 10 percent of the 1982 respondents were in salaried employment
(excluding those who were self-employed), most of them in public or
parastatal jobs. Some countervailing tendencies can be noted in table 6-45.
Graduation from a public technical school aids in finding employment,
whereas graduation from a public commerical school makes it more
difficult to find employment. Salaried workers tended to have scored lower
in English, taken more vocational courses, graduated from poorer quality
schools, and privately spent more on schooling than the rest of the 1982
respondents.
When self-employed workers are added to salaried employees, the
combined group represents 15 percent of the 1982 sample. The only
difference regarding the prediction of employment status is that in the
combined group males are less likely to land jobs in 1982. The explanation
for this is that the self-employed are predominantly female graduates who
reside on farms.



196  DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-43. Maximum  Likelihood Estimates of the Determinants
of Form V Selection in 1982, Tanzania, 1981 Cohort
Marginal
Variable               effect  t-value
Background
Male                            0.086a  3.50
Age                             0.011   1.30
Urban residence                 0.019   0.70
Books in home                   0.001   0.20
English usage in home           0.023   1.20
English usage in school       -0.042a  1.60
Mother's education              0.014   0.20
Father's education              0.002   0.50
Father professional             0.011   0.30
Father white-collar             0.023   0.70
Father blue-collar              0.040   1.20
Father teacher                  0.052   1.20
Father's income                 0.001   0.30
Regional GNP                    0.001   0.30
School quality index            0.034   1.00
Modernity score               -0.011V   2.50
Achievement score
Agricultural                    0.002   0.50
Commercial                      0.002   0.40
Technical                       0.006a  1.60
Math                            0.001   0.20
English                         0.001   0.40
Form IV school
Academic private                0.034   0.60
Agricultural public           - 0.029   0.70
Agricultural private          -0.086   1.40
Commercial public               0.064a  1.70
Commercial private              0.087'  1.50
Technical public                0.018   0.40
Technical private               0.072   0.80
Private school expenditure      0.001   0.20
Teachers' salaries              0.001   0.20
School maintenance expenditure    0.002   0.30
Science periods                 0.017   0.70
Vocational periods              0.001   0.80
Sample size                         1,722
D-value                             0.026
Mean dependent variable                 0.208
a. Significant at the 10 percent level or better.



THE TANZANIA CASE STUDY                                                197
Table 6-44. Maximum Likelihood Estimates of the Determinants
of Recruitment into Training Programs in 1982,
Tanzania, 1981 Cohort
Marginal
Variable            effect  t-value
Background
Male                          - 0.028   1.00
Age                            0.001   0.50
Urban residence                0.001   0.20
Books in home                  0.001   1.00
English usage in home          0.001   0.60
English usage in school       - 0.001   0.40
Mother's education            - 0.009'  2.00
Father's education             0.001   0.80
Father professional            0.024   0.70
Father white-collar            0.001   0.20
Father blue-collar            - 0.003   1.00
Father teacher                 0.001   0.20
Father's income                0.001   0.20
Regional GNP                  -0.001   1.10
School quality index           0.008a  1.70
Modernity score                0.006a  1.60
Achievement score
Agricultural                  -0.001   0.90
Commercial                    - 0.001   0.50
Technical                     -0.001   0.90
Math                          - 0.001   0.90
English                        0.OOi'  1.80
Form IV school
Academic private              -0.001   0.50
Agricultural public            0.058a  1.60
Agricultural private           0.073   1.30
Commercial public              0.001   0.40
Commercial private            -0.001   0.50
Technical public               0.006   0.20
Technical private              0.078   0.90
Private school expenditure     0.001   0.40
Teachers' salaries             0.001   0.40
School maintenance expenditure    0.001   0.30
Science periods                0.001   0.10
Vocational periods            -0.010a  3.50
Sample size                        1,722
D-value                            0.028
Mean dependent variable                 0.416
a. Significant at the 10 percent level or betrer.



198  DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-45. Maximum Likelihood Estimates of the Determinants
of Salaried Employment One Year after Graduation,
Tanzania, 1981 Cohort
Marginal
Variable            effect  t-value
Background
Male                          -0.008   1.00
Age                             0.001   0.60
Urban residence               -0.004   0.50
Books in home                   0.001   0.50
English usage in home         - 0.008   0.90
English usage in school         0.014   1.00
Mother's education            -0.001    1.00
Father's education            -0.001   0.90
Father professional             0.003   0.40
Father white-collar             0.026   1.30
Father blue-collar              0.005   1.00
Father teacher                -0.036    1.00
Father's income                 0.001   1.20
Regional GNP                    0.001   0.80
School quality index          -0.006a  1.90
Modernity score               -0.001   0.60
Achievement score
Agricultural                    0.001   0.50
Commercial                    -0.001    0.50
Technical                       0.002   1.00
Math                            0.001   1.00
English                       - 0.00la  1.50
Form IV school
Academic private                0.011   0.80
Agricultural public           - 0.002   0.50
Agricultural private          - 0.049   1.30
Commercial public             -0.044a  1.70
Commercial private            -0.036   1.00
Technical public                0.041'  1.60
Technical private             -0.057   1.00
Private school expenditure      0.001a  1.70
Teachers' salaries              0.001   0.50
School maintenance expenditure   - 0.001   1.20
Science periods                 0.001   1.00
Vocational periods              0.002'  2.40
Sample size                         1,722
D-value                             0.021
Mean dependent variable                 0.82
a. Significant at the 10 percent level or better.



THE TANZANIA CASE STUDY                                    199
Recruitment by the Government
As has been mentioned, recruitment into Tanzania's large public sector
is based heavily on manpower planning considerations. Nearly 80 percent
of the base cohort in the sample was absorbed by the government in 1982
in one way or another. As table 6-46 shows, most of the factors influencing
that absorption are similar to the ones discussed in the above sections.
First, families with less-educated mothers and families with fathers who
are white-collar workers were most likely to have children recruited into
the public domain. Second, males and older graduates were more likely to
be absorbed by the government than others in the sample. Respondents
coming from poorer economic regions (areas outside Dar es Salaam and
Dodoma) were more likely to have entered government-sponsored activi-
ties, as were respondents with more traditional attitudes. Finally, those
who graduated from public technical and commercial schools and those
who took fewer vocational courses were more likely to find themselves in
government activities. With regard to achievement scores, it appears that
students who scored high on technical tests or low in commercial tests
stood a better chance of being selected. In short, coming from a poor
region, having a father who is a clerical worker, having oneself traditional
values, and being male increase one's chance of getting selected into
government sector activities.
Graduates Still Seeking Placement
Twenty percent of the sample had not been placed in school, training, or
work by 1982, but were seeking to participate in one of these activities
(table 6-47). Respondents whose fathers were not clerical workers,
mothers who were more highly educated, graduates living in high-income
areas, those with more modern outlooks, high achievers on commercial
and math tests, those who took more vocational courses, and those who
were younger were more likely to be searching for work, training, or
employment one year after graduation.
Many of these characteristics suggest that these individuals could afford
to wait for a job or further training. Having higher modernity scores,
which implies a willingness to take risks, and coming from wealthier
regions of the nation and being younger may predispose a graduate to
search longer. Those looking for training or employment opportunities
may have been offered positions by the ministries of Education or Man-
power, but turned them down for personal reasons.



200  DIVERSIFIED SECONDA.RY EDUCATION AND DEVELOPMENT
Table 6-46. Maximum Likelihood Estimates of the Determinants
of Being Recruited into the Pubtic Sector in 1982,
Tanzania, 1981 Cohort
Marginal
Variable             effect  t-value
Background
Male                            0.050a  2.30
Age                             0.016'  2.10
Urban residence                 0.017   1.00
Books in home                   0.001   1.00
English usage in home           0.025   1.50
English usage in school       -0.030   1.40
Mother's education            -0.010a  2.90
Father's education              0.006   1.40
Father professional             0.040   1.20
Father white-collar             0.066a  2.10
Father blue-collar              0.012   0.90
Father teacher                  0.026   0.80
Father's income                 0.001   0.20
Regional GNP                  -0.001a  3.00
School quality index          -0.001    0.30
Modernity score               - 0.001   1.50
Achievement score
Agricultural                  -0.001    0.50
Commercial                    -0.001a  1.50
Technical                       0.001a  1.50
Math                          - 0.001   1.10
English                         0.001   0.90
Form IV school
Academic private                0.021   0.70
Agricultural public             0.024   0.80
Agricultural private          - 0.046   1.00
Commercial public               0.045a  1.50
Commercial private              0.033   1.00
Technical public                0.078'  1.90
Technical private               0.099   1.30
Private school expenditure      0.001   0.50
Teachers' salaries              0.001   1.00
School maintenance expenditure   - 0.001'  1.70
Science periods                 0.002   0.90
Vocational periods            - 0.004'  1.80
Sample size                         1,722
D-value                             0.023
Mean dependent variable                 0.758
a. Significant at the 10 percent level or better.



THE TANZANIA CASE STUDY                                                  201
Table 6-47. Maximum Likelihood Estimates of the Determinants
of Being Unplaced in 1982, Tanzania, 1981 Cohort
Marginal
Variable             effect  t-value
Background
Male                           - 0.029   1.50
Age                            -0.015a  2.10
Urban residence                - 0.028   1.40
Books in home                  -0.001    1.00
English usage in home          - 0.014   0.90
English usage in school         0.001   0.70
Mother's education              0.010a  2.90
Father's education             -0.001    1.70
Father professional            -0.084   1.40
Father white-collar            - 0.069a  2.40
Father blue-collar             - 0.015   1.00
Father teacher                 -0.009   0.50
Father's income                - 0.001   0.90
Regional GNP                    0.oola  2.30
School quality index            0.001    1.00
Modernity score                 0.003'  2.00
Achievement score
Agricultural                    0.001   1.00
Commercial                      0.003a  1.80
Technical                      - 0.001   1.00
Math                            0.OOla  1.80
English                        - 0.001'  1.90
Form IV school
Academic private                0.019   0.70
Agricultural public            - 0.009   0.90
Agricultural private            0.036   0.80
Commercial public              -0.026   0.90
Commercial private             - 0.025   0.90
Technical public               - 0.063   1.50
Technical private              -0.077    1.20
Private school expenditure      0.001   0.80
Teachers' salaries              0.001   0.70
School maintenance expenditure    0.001   1.00
Science periods                - 0.009   0.90
Vocational periods              0.001'  2.30
Sample size                         1,722
D-value                             0.026
Mean dependent variable                  0.123
a. Significant at the 10 percent level or better.



202   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Summary of Logit Analysis
The best school-based strategy for getting into Form V would be to
attend a commercial-track school and score well in technical subjects.
Form IV graduates were more likely to be recruited into further training if
they came from agricultural public schools and scored well in English
tests. To obtain salaried employment, chances are better for private com-
mercial or technical school graduates who have taken several vocational
courses and (surprisingly) scored poorly on the English test. A lack of
English may not be a handicap for this group of public employees because
those who work in party-affiliated firms, utilities, and corporations are less
likely to find foreign language ability a necessity.
Those currently unemployed may be holding out for something un-
usually attractive. As a rule, memnbers of this unplaced group are high
achievers, modern in attitudinal outlook, urban, and younger and wealth-
ier than the rest of the sample. Tk ey graduated from all combinations of
public and private curriculum biases except technical schools.
The Determinants of Labor Earnings
Because several hundred of the 1981 graduates sampled were actually
earning wages in 1982, it is possible to test whether monthly earnings
differed by curriculum. The respondents who graduated from Form IV in
academic programs were used as a control group against which to com-
pare other wage earners, but there were no differences in earnings attribut-
able to the diversified curriculum programs (table 6-48). (The earnings by
curriculum subject were adjusted for the other independent variables in
the earnings function in table 6-49.)
Path Analysis
If it can be theorized that a given variable affects another either because
of an expected causal link or because of some temporal ordering, path
analysis can be used to test the actual relationship between given socioeco-
nomic factors, schooling, and earnings. Following the model presented in
chapter 4 and tested for Colombia in chapter 5, we examined the rela-
tionship among father's education, student ability (a composite of verbal
and nonverbal), school type (diversified or control), achievement scores,
and earnings one year after graduation from Form IV in Tanzania. The
model hypothesizes that family background, ability, school attended, and
achievement scores all influence earnings.



THE TANZANIA CASE STUDY                                               203
Table 6-48. The Determinants of 1982 Labor Earnings,
Tanzania, 1981 Cohort
Variable           Coefficient t-value
Background
Male                          2.01   0.09
Age                         - 30.07   0.42
Urban residence              -1.69   0.42
Father farmer               -22.50   0.20
Father's education           19.38   1.21
Regional GNP                  0.04   0.92
Nonverbal ability             0.05   0.49
Attended private school       5.30   0.04
Achievement score
Agricultural                  10.09'  1.75
Commercial                   - 5.70   1.00
Technical                    -3.97   0.50
Math                         12.77'  1.90
English                      -1.69    0.35
Work characteristic
Self-employed               543.77a  4.50
Started work in 1981       - 116.37   0.52
Work in private wage sector   7.15   0.06
Form IV curriculumb
Agricultural                102.35   0.52
Commercial                  -27.92   0.39
Technical                  -110.27   0.42
Constant                   2,051.48
R2                               0.224
S"ample size                       302
a. Statistically significant at the 10 percent level.
b. Omitted school program variable is the academic control.
Figure 6-6 presents the results of the path model, with the agricultural
score as the intermediate achievement variable. The model is based on the
complete data which were available for 254 respondents. Neither their
achievement scores nor the type of school attended had a significant effect
on earnings. When the agricultural score is replaced with scores in other
subjects, the results show that technical, commercial, math, and English
achievement scores also do not have a significant influence on earnings
(not shown in figure). Thus, the effect of the academic curriculum on
earnings is no different from that of the agricultural, technical, or commer-
cial curriculum.



204   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Figure 6-6. Path Model: Effect of Diversification on Achievement
and Earnings, Tanzania, 1981 Cohort
S iiversisaey g
_    schooli
/<\0.0~~0.21
TheReturnstoEdatin0.043r
\9 / /~~~~~~~~~~0.083
Until the same cohort of graduates has been surveyed again some years
hence, the data set for Tanzania will not permit a calculation of the social
rate of return to investment in secondary education. First, only a very
small fraction (14 percent) of the 1981 cohort was employed in 1982.
Second, most of those employed worked in the noncompetitive public
sector of the economy, and hence their earnings did not necessarily reflect
the marginal product of labor. Third, the graduates were surveyed too
early in their careers for their earnings to approximate the overall lifetime
equivalent earnings differential assumed by the short-cut method (see the
corresponding section in chapter 5, above).
Subject to these qualifications, the set of adjusted earnings reported
above has been used to calculate indicative social rates of return to
investment in the four curricula, but merely to summarize the different
costs and benefits associated with them.



THE TANZANIA CASE STUDY                                               205
Table 6-49. Adjusted 1982 Earnings by Curriculum,
Tanzania, 1981 Cohort
Mean earnings Difference from
(shillings    academic control
Curriculum    per month)     (percent)
Academic       1,014
Agricultural   1,116            10.1
Commercial       986           - 2.8
Technical        902          -10.8
Note: Monthly earnings for 302 public, private, and self-employed workers. Adjustment
factors are listed in table 6-48.
Table 6-50 presents the estimated "rates of return," which highlight the
cost and benefit differences among the four curricula. The academic
(control) program exhibits the highest rate of return (6.3 percent) and the
technical bias the lowest (1.7 percent). It should again be emphasized that
the usefulness of this calculation lies more in the comparison of curricula
than in the absolute level of the returns.
Conclusion
Although the follow-up survey of school leavers, made one year after
graduation, allows sufficient time to test some of the hypotheses about
external efficiency, others require a longer period. However, from the data
available it appears that in Tanzania diversification has NOT:
* Reduced private demand for schooling beyond Form IV of secondary
school
* Achieved a close correspondence between vocational curriculum and
subsequent training
Table 6-50. Indicative Social Rates of Return to Investment
in Secondary Education, Tanzania, 1981 Cohort
(percent)
Form IV curriculum Rate of return
Academic (control)     6.3
Agricultural           5.4
Commercial             3.2
Technical              1.7
Overall                3.7
Note: Returns are based on regression-adjusted earnings (table 6-49).



206    DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Table 6-51. Means and Standard Deviations of Selected Variables,
Tanzania, 1981 Cohort
Standard
Variable                   Mean  deviation
Base-year sample (sample size = 4,181)
Background
Male                                          0.668    0.047
Age                                             19.1    1.30
Urban residence                               0.179   0.384
Number of siblings                             6.36   2.720
Mother's education                             4.33   3.045
Father's education                             5.97   3.296
Father's annual income (shillings)            5,318   11,010
Regional per capita income (shillings)          944     980
Father farmer                                 0.419   0.493
Father professional                           0.101    0.301
Father white-collar                           0.119    0.324
Father teacher                                0.060    0.237
Father blue-collar                            0.101    0.302
Father otherwise employed                     0.200   0.400
Father self-employed                          0.526   0.499
Expectations
Will seek private employment                  0.051   0.221
Expected post-Form IV earnings (shillings a month)  836  649
Achievement score (nonstandardized)
Kiswahili                                    46.500    9.499
Math                                         36.400   16.300
English                                      41.500   12.300
Commercial                                    6.060   2.720
Agricultural                                  6.140   2.590
Technical                                     6.940   2.999
Ability
Verbal aptitude                              10.353    3.822
Math aptitude                                 7.330    3.200
Modernity score                               5.830    2.570
Form IV school
Private agricultural                          0.053    0.255
Public agricultural                           0.277   0.447
Private technical                             0.021   0.143
Public technical                              0.097   0.296
Private commercial                            0.129   0.335
Public commercial                             0.179    0.383
Private academic                              0.074   0.262
Public academic                               0.170   0.402
Private expenditure per student (shillings/year)  1,759    1,292
Public teacher expenditure per student        1,045     870



THE TANZANIA CASE STUDY                                                207
Table 6-51 (continued)
Standard
Variable                    Mean  deviation
1982 Follow-up (sample size = 2,297)
Destination
Form V                                         0.295    0.456
Training                                       0.416    0.493
Salaried employment                            0.091    0.288
Self-employed                                  0.053    0.224
Looking for placement                          0.144    0.351
Expectations
Expected earnings of Form V students           3,450    7,653
Expected earnings of students in training      1,282     944
Labor market outcomes
Earnings of salaried workers                     699     153
Earnings of self-employed                        319     304
Satisfied with training program                0.653    0.476
Hours worked per week (salaried worker)         42.8    16.8
Hours worked per week (self-employed)           41.6    23.3
Sector of economic activity
Public sector                                  0.077    0.266
Private sector                                 0.010    0.102
Self-employed                                  0.053    0.224
* Generally raised the proportion of secondary school graduates di-
rectly entering employment (although a higher proportion of techni-
cal graduates than academic students were employed in 1982-17
percent compared with 14 percent)
* Increased the likelihood that the few who are employed will obtain
work in their field of specialization (however, after all students have
completed their post-Form IV education and training, this situation
may change)
- Reduced the amount of time needed to find work (though technical
students do find employment earlier than all others)
* Raised the initial earnings of graduates of vocational programs above
those of academic graduates (again, however, it is possible this finding
may change when all students have entered the labor force).
An evaluation of the diversification of secondary school curricula in
Tanzania must also take account of the political rationale for the



Table 6-52. Zero-Order Correlation Matrix among Selected Variables, Tanzania, 1981 Cohort
(1)    (2)    (3)    (4)    (5)    (6)    (7)    (8)    (9)    (10)   (11)   (12)  (13)  (14)
1. Mother's education     1.00
2. Father's education     0.66    1.00
3. Father's earnings      0.03    0.02    1.00
4. Father in private sector  -0.21  -0.24    0.09    1.00
S, C. FatLIcr fama  .    .l.-v.27  -u.31    0.04    0.41    1.00
6. School fees            0.17   0.14    0.03  -0.05  -0.18    1.00
7. Expected earnings      0.01    0.00    0.03  -0.02  -0.01    0.01    1.00
8. Verbal aptitude      - 0.04  - 0.03  -0.01    0.06    0.06  - 0.26  - 0.02    1.00
9. Modernity score      -0.04  -0.03    0.02    0.00    0.06  -0.14    0.01    0.30    1.00
10. Labor earnings in 1982   0.11    0.05    0.07   0.01  - 0.07    0.11  -0.14  -0.05    0.08    1.00
11. In Form V in 1982      0.02    0.02  - 0.02    0.01  -0.03  - 0.02    0.02    0.02  -0.04  -0.09    1.00
12. Working in 1982        0.00   0.00    0.01   0.01   0.01   0.00   0.04    0.00  -0.02    0.03    0.03    1.00
13. Looking for work in 1982 0.03    0.01    0.01  -0.02    0.00   0.02   0.00    0.00    0.02   0.00  -0.12  -0.09  1.00
14. Diversified school     0.01  -0.02  -0.01  -0.01  -0.03  -0.11  -0.01  -0.04    0.01  -0.01    0.03  -0.01  0.02  1.00



THE TANZANIA CASE STUDY                                           209
innovation.8 The previous education system was acknowledged to be
elitist, inadequate, and irrelevant to Tanzania's needs. The new system, of
which diversified secondary schooling is a part, constitutes one of the
means of implementing the political ideology of socialism and self-
reliance. By making secondary school curricula more practical, policy-
makers hoped to instill positive attitudes toward work-a sense of dignity
and respect for workers at all levels of skill-and to provide students with
training that would enhance individual development and that of the
community and the nation at large.
Difficulties in implementation may mean that diversification is not
initially cost-effective, but it may still be considered worthwhile if the
political objectives for which it was adopted are being realized. Thus, its
value should be judged by the extent to which it advances realization of the
national objectives of socialism and self-reliance and the extent to which
individuals themselves profit from their educational experiences.
8. Based on Tanzania Ministry of National Education, letter to the Diversified Secondary
Curriculum Study research team, January 25, 1984, pp. 19-20.



7. Summary and Conclusions
THIS STUDY has been concerned with whether the diversification of curric-
ula in secondary schools has led to differences in postschool experiences
between students who enrolled in prevocational courses and those who
concentrated solely on academic programs. In chapter 4, a number of
positive results of diversification were noted, including higher cognitive
achievement and better labor market performance of students following
diversified curricula. But the question that remains is how these outcomes
relate to the higher costs involved. If, for instance, a diversified curriculum
imparts cognitive skills and other attributes that are superior to those
taught in a traditional program and is also judged to be cost-effective,
those findings would provide some support for diversification. Similarly, if
graduates of diversified prograrms find jobs more easily and at higher
salaries than do graduates of more traditional academic programs, a case
could be made in favor of any adclitional cost. Conversely, if the results of
diversification are not significantly better than those of traditional educa-
tion, yet extra costs are involved, the case for this type of curriculum
reform is weakened.
The individual country findings described in detail in the previous two
chapters are summarized and consolidated in the following pages in an
attempt to make some initial judgments about the success of such curricu-
lum innovation. We begin with iss bies of internal efficiency and look briefly
at learning achievements of students in different curriculum programs.
There follows a discussion of students' postschool experiences in further
educational programs and in the labor market.
Internal Efficiency: Colonibia
Achievement Scores and Course Hours
In Colombia, students attending the diversified Institutos Nacionales de
Educacion Media (INEMS) were fotnd to come from less affluent socioeco-
nomic backgrounds and to have slightly lower aptitude scores than non-
INEM students, regardless of curriculum. Yet their achievement scores in
210



SUMMARY AND CONCLUSIONS                                      211
all vocational subjects and in academic subjects surpassed those of the
non-INEM control group after socioeconomic background, natural ability,
and other nonschool factors were controlled for. The results point to
substantial differences in achievement by curriculum stream and school
type, which cannot be attributed to out-of-school factors.
When students' learning achievements by school type and by academic
and vocational specialization are considered, the following patterns
emerge:
* INEM students always score much higher than non-INEM students on
vocational tests in general.
* Regardless of the type of school attended, students concentrating in a
particular curriculum always have higher scores on tests of their own
specialization than do others not taking that subject.
* INEM students' scores are better than those of control students on
academic achievement tests as well.
Apparently, not only have INEM schools been successful in raising the level
of learning in vocational subjects, but they have achieved this success
without sacrificing academic knowledge.
One way to test whether the learning differences between INEM and
non-INEM students are due to the curriculum followed is to see whether
greater exposure to vocational subjects enhances vocational test achieve-
ments. Variations between INEM and control schools in course hours were
substantial and contributed heavily to explaining the higher INEM achieve-
ment scores. Because INEM schools were all developed at the same time,
they offer more similar numbers of course hours in vocational subjects
than do the non-INEM schools. In addition, all INEM students devote more
time to their subject of specialization during their final year than do
students in other schools.
Influence of Home and School
Another measure of internal efficiency concerns the relative influence of
home and school on achievement scores. Previous studies have generally
shown that secondary school inputs affect academic achievement less than
do personal and family characteristics. Does this pattern also hold for
students of vocational subjects? Vocational achievement, by its nature,
may be more dependent on learning in school than is achievement in
languages, the humanities, science, and mathematics.
The impact of home and school variables on test scores was measured
for the entire sample. Only 20 percent of the total explicable variance in
the academic achievement test (humanities and science) was attributed to



212   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
school inputs. But in the vocational achievement tests school factors
explained up to four times as much: 79 percent in industrial, 53 percent in
commercial, 60 percent in agriculture, and 53 percent in pedagogy.
One possible reason for the large difference is that levels of innate
ability, as measured by tests of verbal and numerical reasoning, do not
correlate equally with achievement levels for each school subject. Even
though natural ability is the most powerful out-of-school influence on
performance in vocational tests, its overall impact in that area is smaller
than it is on academic scores. In o Lher words, test performance in academic
subjects is more predictable from measures of aptitude than is test per-
formance in vocational subjects. Thus, there is more room for school
factors, shaped in part by government policy, to influence achievement in
vocational subjects.
Unit Costs by Curriculum
The Colombia study focused on two issues: whether learning and
postschool experiences differ between students pursuing an academic
course and those in prevocational studies, and whether outcomes vary
according to school type (INEM4, non-INEM). Before trying to assess
whether and how differences in oitcomes are related to the resources used
to produce them, it is necessary to attempt two types of cost analysis: first,
a comparison of costs among the different curricula (academic, industrial,
commercial, social services, and agriculture) offered within INEMS; and,
second, a comparison of each of these with the non-INEM schools teaching
those subjects.
The cost of schooling can be looked at from three viewpoints-society,
government, and the individual (or household). The cost components that
may be relevant to one or more of these are forgone output or earnings;
capital costs of buildings, furniture, and equipment; and direct recurrent
costs. For calculations of social costs it is immaterial how capital and
recurrent costs are financed. From the viewpoint of the government or
individual, however, the source o' expenditure is obviously important. An
analysis of recurrent costs revealed that:
* In INEMS, recurrent costs differ little among curricula.
* INEM academic and commercial tracks are more expensive than their
counterparts in control schools.
* INEM agricultural and industrial curricula are significantly less expen-
sive than their control counterparts, and the same is true for the social
services but to a lesser extent.
The social costs of all curricula in both types of school are dominated by



SUMMARY AND CONCLUSIONS                                      213
earnings forgone. Differences in social costs are thus small; they lie within
a range of 12 percentage points. While control schools specializing in
academic and commercial studies have costs slightly less than those of
INEMS, costs of schools for agriculture, social services, and industry are
higher. For capital and recurrent costs alone, again variations among
INEMS are very slight. There are, however, wide differences between the
average costs in INEMS and those in single- and multi-track control
schools, with agriculture and industry studies being one-third and one-
quarter more expensive respectively in control schools.
For total social costs, differences between school types (INEM and
non-INEM) and between academic and prevocational tracks are small. For
capital and recurrent costs alone, however, there are substantial differ-
ences among non-INEM schools and between them and the INEM schools.
INEM academic and commercial schools are 16 and 9 percent, respectively,
more expensive than their control counterparts, while INEM industrial,
agricultural, and social services curricula are between 10 and 22 percent
less expensive.
Cost-Effectiveness of Cognitive Acbievement
To relate costs to achievement, the differences in the capital and recur-
rent costs of producing various types of secondary graduates are compared
with the differences in the graduates' achievement. There are two relevant
sets of comparisons: the first compares students following the same track
in different types of school, and the second compares the control group of
academic students with prevocational students.
Students in the academic track cost 16 percent more resources a year in
INEM schools than in control schools. Twelve percent of INEM students in
the academic track outscore control students on academic tests of achieve-
ment. In the commercial track, a 9 percent extra cost per student generates
no additional academic achievement. The INEM industrial, agricultural,
and social services curricula, however, are cost-effective since with less
resources they produce equivalent or higher academic achievement scores.
Cognitive differences between students in INEM and control secondary
schools favor INEM schools. Although INEM students come from slightly
less advantaged socioeconomic backgrounds and score lower on aptitude
test, their achievement scores are higher than average on the relevant
vocational tests and higher than average on the academic tests in three out
of five of the curricula.
A comparison of the control group of students taking academic subjects
and students who have followed prevocational courses shows that INEM
students do better than academic control students in the prevocational



214   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
tests. Any apparent concurrent reductions in academic test performance
are relatively slight in the raw data and vanish when statistical controls are
introduced.
Internal Efficiency: Tanzania
Achievement Scores and Cou?se Hours
In Tanzania, students in vocational courses always scored higher on
achievement tests of their own specialization than did the academic con-
trol group. Conversely, students in vocational curricula scored below the
academic control students on all academic tests.
We also tested whether students' achievement was affected by the
proportion of their total course load spent on vocational subjects. The
higher proportion of time spent on agriculture and commerce did appear
to raise achievement in those areas, but additional course work in techni-
cal subjects did not. We also found that the earlier a school had begun
teaching vocational subjects, the higher were the achievement scores of its
students.
Influence of Home and School
Another issue in assessing internal efficiency is the extent to which
school resources affect learning, home influences and ability. As in Colom-
bia, the hypothesis tested was thai: school factors would be more powerful
in predicting scores for vocational subjects than for academic subjects.
Confirmation of this hypothesis would imply that vocational skills could
be taught in schools more successfully than, say, on the job. In contrast to
the findings in Colombia, however, in Tanzania school factors had no
greater impact on vocational achievement than on achievement in
academic subjects.
Unit Costs by Curriculum
In Tanzania the unit costs of secondary schooling are more easily
determined than in the case for Colombia, because vocational and
academic subjects are taught in different types of school and therefore
separate accounts are kept.
Total social costs of each type of school are dominated by forgone
earnings, and differences accordir.g to school curriculum are statistically
insignificant. Because of this, the variation in total social costs by curricu-
lum appears small-the average for vocational schools is only 3 percent



SUMMARY AND CONCLUSIONS                                      215
above that for control schools. From the point of view of public expendi-
ture, however, the differences in cost are substantial. Total recurrent costs
are 14 percent higher in the vocational schools than in those emphasizing
purely academic subjects. Government-financed recurrent costs, however,
are on average 34 percent higher for the vocational schools.
Cost-Effectiveness of Cognitive Achievement
Unit costs were highest in agricultural and technical schools, and lowest
in academic control schools. Differences in achievement were compared
with differences in cost per student for the various curricula in order to
assess the internal efficiency of schooling in Tanzania.
The unit costs for agricultural students were 19 percent higher than for
academic students; math and English scores were virtually the same for the
two groups; and agricultural students performed 8 percent better on the
agriculture test. In this case, the increase in learning appears small relative
to the higher unit cost. For technical students costs were 13 percent higher,
math scores 16 percent higher, and technical test scores 23 percent higher
than those of the control groups. In the commerce stream, unit costs were
9 percent above those for the academic stream and commercial students
scored 19 percent higher on tests in commercial subjects.
In summary, in Tanzania all prevocational curriculum tracks are more
costly than academic ones. For technical and commercial students the
gains in vocational learning appear substantial; for agricultural students,
they are small. For all students, there is a tradeoff between gains in the
vocational subject and losses in the academic ones.
External Efficiency
The focus of this study is on the results of education in different types of
secondary schools, and the relationship between these results and the costs
of providing schools. One result is learning, as measured by achievement
tests in both prevocational and conventional academic subjects. This was
extensively discussed in chapters 5 and 6 and summarized in the previous
section. Another result is the school leaver's ability to undertake further
education or training and ultimately to find work. Variations in this ability
associated with different forms of schooling are the subject of this section.
An important qualification at the outset is that only a minority of
secondary school leavers in Colombia and Tanzania had entered the labor
market within a year of graduation. As a result, the follow-up surveys
conducted in each country in late 1982 can only suggest the effects of
different forms of secondary schooling on the labor market performance



216   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
of school leavers. Further tracing of the same cohorts is planned in order to
complete the picture.
The indicators revealed by the study are therefore only tentative, but
they are nonetheless important, since the main rationale for introducing
vocational or prevocational studies has been to increase the economic
relevance of schooling. Higher achievement scores and positive changes in
attitudes toward work are not ends in themselves, but are presumed to
raise the economic contribution of schooling.
There are two ways to measure the relevance or economic impact of
diversified schooling. The first is to examine how closely students' post-
school activities match their subject of specialization; that is, to assess the
correspondence between secondary school curriculum and the content of
subsequent training or education, the occupation entered, or the period of
unemployment after graduation. In this approach, the success of diver-
sified schooling would be judged by how many more jobs were filled by
appropriately prepared people than would otherwise have been possible.
The usefulness of the information collected, however, depends in part on
the degree to which the governrment controls labor market placements.
The more both supply and demand are controlled, the less informative is
the degree of fit.
An alternative approach is to go straight to the earnings data. If, other
things being equal, a graduate who has followed a vocational track earns
more than one who has followed a purely academic track, and the differ-
ence in earnings at least matches any additional costs per student of
providing vocational education, diversification can be regarded as contrib-
uting more to economic developnr ent than does traditional education. The
validity of this approach depends on the degree to which earnings reflect
actual productivity rather than, say, government policy to keep wages
egalitarian.
In the rest of this chapter, the results of the tracer surveys are used to test
hypotheses that graduates' careels differ according to the secondary cur-
ricula followed and, in the case of Colombia, the type of school attended.
The hypotheses are:
a. Diversification reduces individual demand for postsecondary
schooling.
b. Diversification leads to a closer correspondence between school
curriculum and the field of postschool training.
c. Diversification leads to a higher propensity to seek work after sec-
ondary school.
d. Diversification leads to employment in the field of prevocational
specialization.



SUMMARY AND CONCLUSIONS                                   217
e. Diversification leads to a shorter period of job search after gradua-
tion.
f. Diversification enables graduates to earn more initially than gradu-
ates of other programs.
Colombia
For several years Colombia has had secondary schools that concentrate
on one or more prevocational subjects as well as schools that teach
conventional academic subjects. What is new is the introduction in the
1970s of the INEM schools, each offering a set of prevocational subjects in
addition to an academic curriculum.
If the purpose in the Colombia study were simply to compare the
outcomes of studying academic and prevocational subjects, then data on
all students taking each subject could be aggregated across school types. It
was also important, however, to measure the effects of the new INEM
schools and to identify the most appropriate ways of introducing prevoca-
tional subjects (if they should be introduced at all). Thus, data on students
from INEM and non-INEM schools need to be differentiated. Where possi-
ble, therefore, outcomes were analyzed by both curriculum and school
type.
The labor market in Colombia is more competitive than in Tanzania.
The availability of secondary school places is not so rigidly determined by
forecasts of manpower requirements, the private sector is large, and there
is no government policy to narrow wage differentials. As a result, differ-
ences among graduates in finding employment, in their earnings, and in
the degree of correspondence between their secondary school curriculum
and their occupation will be reasonable indicative of the interaction
between the effective demand for and supply of labor-and hence the
demand for and supply of secondary school graduates of different curric-
ula.
HYPOTHESES RELATING TO OUTCOMES
Respondents to the 1982 tracer follow-up were divided into those
studying full time, those working full time, those combining study and
work, and a residual group of the unemployed and voluntarily inactive.
Exactly the same proportion (37 percent) of graduates from INEM and
non-INEM schools took up further study, and a slightly higher proportion
of control graduates were working (31 percent as opposed to 29 percent of
INEM graduates). Thus in the choice between further study and work, there



218   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
is no systematic difference betwe en graduates of vocational and academic
tracks.
The results for the hypotheses presented above are as follows:
a. Diversification reduces individual demand for postsecondary
schooling.
Since there is virtually no difference in the proportions of academic and
prevocational graduates from either type of school who continue studying,
it appears that neither curriculurn diversification nor the introduction of
INEMs has reduced the desire of secondary students to continue with some
type of formal education. Prevocational studies, like academic studies,
seem to be used to qualify for Further education. There are, however,
differences between tracks and schools in the type of further education
followed. A higher percentage of INEM academic students (67 percent)
enter university than do the control group of academic students (61
percent).
b. Diversification leads to a closer correspondence between school
curriculum and the field of postschool training.
Postsecondary education in Colombia can be at a university, at a
technical institute, or in a nonaccaedited program. One year after gradua-
tion, 25 percent of non-INEM graduates and 24 percent of INEM graduates
were currently or had previously been enrolled in a postsecondary training
program. These programs have been grouped into humanities, commer-
cial, technical, and a residual category.
INEM graduates are more likely to follow their secondary school spe-
cialization (academic or vocational) in their postsecondary studies. But
from every track, more graduates take commercial courses than any other
type. Forty percent of the courses taken are commercial, and the demand
for them is the same among graduates of all curricula. Among INEM
graduates, those having taken industry and agriculture are more likely to
take industrial courses than are the others; INEM academic graduates take
humanities-related courses in a greater proportion than do other gradu-
ates. Among non-INEM graduates., the proportion of industrial graduates
taking technical courses is surprisingly low.
c. Diversification leads to a higher propensity to seek work after sec-
ondary school.
As with the first hypothesis, it is assumed that with some prevocational
training fewer students will deman d further scooling, and that they will be
less likely to require additional training. In short, they will possess
appropriate attitudes and skills and be better prepared to enter the labor
force.



SUMMARY AND CONCLUSIbNS                                    219
One year after finishing secondary school, 30 percent of the graduates
were working full time. So as to focus on the data more sharply, the
observations for agricultural, commercial, and industrial studies are
aggregated under the heading of vocational subjects.
There was virtually no difference in employment rates between
academic and vocational graduates of the INEMS, or between these groups
and non-INEM academic graduates. Moreover, among those who work full
time, non-INEM vocational graduates have a higher propensity to be
working full-time one year after graduating than do INEM vocational
graduates (about 34 and 28 percent, respectively). These percentages
change when those who both work and study are added to the graduates
who work full time. Vocational graduates from non-INEM schools are
more likely to be in some kind of employment (full- or part-time) one year
after secondary school than are INEM vocational graduates (about 44 and
39 percent, respectively).
d. Diversification leads to employment in the field of prevocational
specialization.
Even if diversification of the curriculum does not lead to a higher rate of
employment immediately after secondary school, it could be hypothesized
that those who do work will tend to be employed in occupations and
sectors that correspond with the curriculum followed. This hypothesis can
be tested by examining, first, the occupational groupings and, second, the
economic sectors into which INEM and control academic and vocational
graduates flow.
The wide range of occupations entered by graduates has been narrowed
down into six groupings: teaching, secretarial work, other white-collar
jobs, vending, agricultural work, and blue-collar jobs. Neither the INEM/
control nor the academic/vocational dichotomy shows any obvious pat-
tern, except that teaching is a much less likely occupation for control
graduates than for those of INEM schools. For a particular track and type
of school some differences emerge. For example, INEM academic graduates
are least likely to be in vending, control vocational graduates are least
likely to be blue-collar workers, and control academic students are least
likely to be in teaching. The occupational distribution has been further
disaggregated by vocational track, and again there is very little pattern.
Graduates who follow the commercial track are no more likely to work in
secretarial and other white-collar jobs than are graduates from other
tracks, and industrial students are no more likely to work in blue-collar
jobs. Nor are there major differences in the economic sectors in which
graduates of different school types or tracks are employed, except that
INEM graduates are more likely to work in personal services than are
non-INEM graduates.



220   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
Thus, one year after graduation there is no strong evidence that
academic and vocational students enter different kinds of jobs or work in
different sectors of the economy. Neither does the type of school attended
appear to influence the jobs helcl.
e. Diversification leads to a shorter period of job search after gradua-
tion.
This hypothesis assumes that graduates with some prevocational train-
ing will have some recognizable skills and more definite ideas of the jobs
they want. At the same time, eriployers may expect them to need less
on-the-job training and therefore will consider them to be more attractive
candidates than those without prevocational training. As a result, gradu-
ates of vocational curricula should have a shorter period of unemploy-
ment.
The data for Colombia do not substantiate this hypothesis. If anything,
they show the reverse: those who have taken vocational courses in either
INEM or control schools experience significantly longer periods of unem-
ployment before they land their fir-st job (26 weeks for vocational students
from both types of school compared with 21 weeks for INEM and 22 weeks
for non-INEM academic students). Some graduates, however, were still
unemployed one year after graduation. Levels of unemployment in late
1982 were slightly higher for INEM students, and in particular for those
who had followed the academic and agricultural curricula.
f. Diversification enables graduates to earn more initially than gradu-
ates of other programs.
If diversification provides prevocational training that either makes
graduates more productive or reeuces the amount of on-the-job training
they need, the initial earnings of diversified school graduates should be
higher than those of graduates who followed an academic curriculum.
The difference in average earnings between INEM and control graduates
is 1 percent in favor of the latter. Before any conclusions are drawn from
these figures, it should be repeated. that only 30 percent of the sample was
working, the range of earnings within each track and school type is very
wide, and differences in the means between tracks are not statistically
significant at the 0.05 level. At the very least, however, it can be said that
there is no evidence to support the hypothesis that the initial earnings of
prevocational students will be higher than those of academic students.
CORROBORATING EVIDENCE FROM THE 1978 COHORT
In the case of Colombia it was possible to obtain a sample of graduates
from an earlier cohort (1978) to get an extended view of the external
efficiency of diversified schools. By and large, the evidence from those who



SUMMARY AND CONCLUSIONS                                      221
have been out of school three years confirmed, and in some cases rein-
forced, the results obtained for the younger cohort.
* INEM graduates have a higher propensity than control school gradu-
ates to be studying three years after secondary school (54 and 49
percent, respectively).
* INEM graduates are less likely than control school graduates to partici-
pate in the labor force three years after graduation (67 and 73 percent,
respectively).
* School type and curriculum do not relate in any way to differences in
earnings among those employed.
* The length of time needed to find a first job is not related to the type of
school attended or the subjet studied.
* The incidence of unemployment three years after graduation is highest
among those who studied agricultural and academic subjects (in either
type of school) and those who studied industrial arts in INEM schools.
* The reservation wage of job seekers, especially those coming from
INEM academic schools, is substantially higher than the actual wage
rate of those employed.
In the case of the 1978 cohort we were able to estimate rough social
rates of return to investment in different types of secondary school offering
various curricula. No major differences were discovered in the returns to
education for those coming from the two types of school and various
tracks; the overall profitability to such investment was about 8 to 9
percent.
Tanzania
The tracer data on secondary school students one year after graduation
are difficult to interpret in the case of Tanzania. The government has
sought to expand secondary school enrollment strictly in line with fore-
casts of manpower requirements and itself hires nine out of ten graduates.
At the extreme, it can thus ensure a close fit between, say, graduates from a
technical school and jobs with a presumed technical content, regardless of
whether the graduates contribute to the national economy or whether
their education was cost-effective. Given the Tanzanian policy, anything
other than short periods of unemployment or job search and a close
correspondence between schooling, training, and work would indicate
that the government had failed to forecast manpower requirements cor-
rectly or to finance the absorption (in employment or future training) of
the planned output of graduates. The structure of earnings in Tanzania is
very much a creation of government policy: the private sector is small, and
a conscious attempt is made to narrow wage differentials in the public
sector. In these circumstances, significant differences in earnings across



222   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
curricula would be very unlikely, particularly in the first years of employ-
ment.
HYPOTHESES RELATING TO OUTCOMES
With these qualifications and expectations, the data from the follow-up
survey are discussed below in the f ramework of the stipulated hypotheses.
a. Diversification reduces individual demand for postsecondary
schooling.
The evidence from the Tanzania study does not support the view that
the incorporation of prevocational studies into secondary school will
decrease the demand for more schooling. When students in Form IV were
asked whether they would apply for Form V, 72 percent of agricultural, 71
percent of technical, 81 percent of commercial, and 82 percent of
academic students answered positively. Because this level of education is
severely rationed, only a minority af these students were admitted in 1982:
of those applying in each bias, 25 percent of agricultural, 33 percent of
technical, 34 percent of commercial, and 28 percent of academic students
gained places in Form V. Thus the very high private demand for postsec-
ondary schooling does not appear to be diminished by the introduction of
prevocational studies. Actual enrollment, however, is determined by the
effective supply of places.
b. Diversification leads to a closer correspondence between school
curriculum and the field of postschool training.
It is often argued by advocates of diversification that the additional
subjects are prevocational rather than vocational. But although one
should not expect graduates of these schools to be fully trained, there
should be a close correspondence between the curriculum followed in
school and the subject matter of subsequent training. The 1982 tracer
information reveals that teacher training clearly dominates postschool
training activities. The proportion of academic graduates who enter
teacher training does not appear to be out of line with the rest of the
graduates. One interesting result is the small proportion (6 percent) of
commercial graduates and the high proportion (21 percent) of technical
students who take further clerical courses. This implies that post-Form IV
clerical courses are not seen as adding anything to similar studies in
secondary schools, while a substantial proportion of technical students
who opt for postschool training anticipate openings in the clerical field.
Finally, fewer agricultural graduate, (5 percent) than other graduates take
postschool agricultural training courses.
Thus no close relationship has been demonstrated between school cur-
riculum and the content of subsequent training courses.



SUMMARY AND CONCLUSIONS                                    223
c. Diversification leads to a higher propensity to seek work after sec-
ondary school.
One year after leaving Form IV, only 14 percent of the graduates were
employed: 9 percent were in wage employment and 5 percent were self-
employed. Between 27 and 30 percent entered Form V or further training
immediately after Form IV (the proportion depends on the curriculum).
While students from technical schools showed an above-average propen-
sity to be employed (17 percent), those from the other two prevocational
curricula behaved no differently from academic students. There were no
significant differences by curriculum in the type of employment or in the
earnings received. Nor were there differences in the time taken to find a
job.
In addition to those already working or studying, 7 percent of the
sample was looking for immediate employment and an additional 8
percent were hoping to take additional training. This means that despite
the very tight control over secondary school expansion based on man-
power forecasts, one in seven graduates was not placed in a job or further
training one year after graduation. Technical students were underrepre-
sented in this group-10 percent of them were unoccupied, against an
overall average of 15 percent.
In short, diversification does not appear to lead to a higher propensity
for work.
d. Diversification leads to employment in the field of prevocational
specialization.
If the introduction of prevocational studies into secondary schools is
intended to orient students more toward the world of work in general,
graduates might be expected to move into occupations corresponding to
the curriculum followed. Again, any comments on this are necessarily
limited to only the 14 percent of the sample who were working for wages
or self-employed one year after leaving Form IV.
In general, it is not possible to observe any correspondence between bias
and job type. Of those who did obtain wage employment, 50 percent
found public sector clerical jobs, but commercial students were no more
likely to be working in clerical jobs than were academic or agricultural
students. For the self-employed, however, there is a difference by curricu-
lum. Although 21, 24, and 29 percent of the self-employed graduates from
commercial, agricultural, and academic studies, respectively, were work-
ing on farms, the corresponding figure for technical students was only 9
percent. Of those looking for immediate employment, only 8 percent of
the technical students expected to work in commerce, but for all other
curricula the figure was between 20 and 30 percent.
Although graduates from the academic, agricultural, and commercial



224   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
streams show no differences in their pattern of employment, technical
graduates obtained employment more in line with their specialization.
e. Diversification leads to a shorter period of job search after gradua-
tion.
If students are given some orientation toward specific forms of employ-
ment they ought to be more employable and more sure of the type of work
they wish to do. On this assumption, graduates who had followed agri-
cultural, technical, and commercial curricula would expect to experience a
shorter period of unemployment than those in purely academic subjects.
However, the data from Tanzania do not support this.
One year after graduation, 13 percent of academic students were still
looking for either work or training, while the percentages for technical,
commercial, and agricultural stuclents were 10, 16, and 16, respectively.
These academic graduates did not believe that they had any longer to wait
for placement than did other students. Nor did those in Form V or in
training courses show any difference by curriculum in the period they
expected to have to wait to find employment once they entered the labor
force.
f. Diversification enables graduates to earn more initially than gradu-
ates of other programs.
On the assumption that prevocational studies make a person more
employable and less in need of training on the job, it may be hypothesized
that at least the initial earnings of vocational graduates will be higher than
those of academic graduates. Because only 9 percent of Form IV students
were in salaried employment one year after graduating, any data presented
on earnings can be only approximate. The data give average monthly
earnings of 709, 710, 736, and 762 shillings for technical, commercial,
agricultural, and academic students, respectively, and an overall average
of 726 shillings a month. Those still looking for work put their "minimum
acceptable earnings" at a much higher average level of 947 shillings
(ranging from 857 shillings for academic students to 1,057 shillings for
commercial students).
Therefore, although only a very small proportion of school graduates
were working one year after leaving secondary Form IV, the data do not
support the hypothesis that initial earnings of graduates from prevoca-
tional biases will be higher than tiose received by other graduates.
THE RETURNS TO EDUCATION
It was argued previously that the external effects of introducing diver-
sified secondary schooling could in principle be measured in two ways:



SUMMARY AND CONCLUSIONS                                      225
first, by looking at the correspondence between type of school and occupa-
tion and at such measures as the amount and length of unemployment (a
manpower forecasting approach; second, by comparing the earnings of
graduates of different curricula with the costs of their education. The
problem with either approach in the case of Tanzania is that only a small
fraction of secondary Form IV leavers were employed one year after
graduation. Analyses pertaining to their labor market performance there-
fore need to be regarded as merely indicative.
Subject to this caveat mean earnings have been calculated and compared
with cost differences to give a rough approximation of social rates of
return. By any standard, the rates of return are low, ranging from 2 to 6
percent. The technical curriculum has the lowest return-a reflection of its
higher unit cost. Conversely, the academic stream exhibits the highest rate
of return. One should be cautious, however, in interpreting these rates as a
case against all forms of secondary schooling. Not only were the data
collected early in the careers of the graduates, but also the imperfections of
the labor market make it difficult to assess the results. Of those secondary
leavers who were employed, 88 percent were in the public and parastatal
sectors, and government policy was to minimize earnings differentials. In
these circumstances earnings differentials are unlikely to adequately reflect
differences in productivity, and hence rates of return are poor indicators of
the social profitability of educational investment. All that can be said at
this point is that the first indications do not corroborate the hypothesis
that the introduction of prevocational studies into secondary schooling
can be justified on the grounds that the economic payoff is greater for an
academic curriculum.
Concluding Remarks
The effectiveness of school reform is generally judged in two ways: first,
the extent to which access to education is increased and, second, the extent
to which knowledge is acquired and cognitive skills developed. In develop-
ing countries, however, the role of education in promoting economic
development has been emphasized. In particular, the responsibility of
schools for determining the labor market outcomes of graduates has been
stressed much more than in industrialized countries. Certainly, the move-
ment toward curriculum diversification in developing countries has cen-
tered on the hope that prevocational studies would ease students' transi-
tion from school to work by changing their attitudes and by teaching
employable skills.
Tnis study has attempted to assess the effects of curriculum diversifica-
tion by measuring any increases in the access to education, by measuring



226   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
and comparing students' attitudin al changes and learning achievement, by
comparing postschool experiences, and by comparing the costs of educa-
tion with its outcomes (graduates' activities and earnings). The last com-
parison is ultimately required for a thorough evaluation of the investment
made in curriculum diversification, but the data on this are the most
incomplete. Information on labo r market experience was collected only
one year after graduation. At that time many secondary school graduates
were continuing their studies, many of those who were employed could be
expected to change jobs, and the meaning of the unemployment figures
and the eventual outcomes of job search were difficult to determine. The
information on, and analysis of, postschool experiences therefore needs to
be treated with a great deal of caution.
The results of the attitudinal tests are also not definitive. Even if educa-
tion in prevocational subjects does change students' occupational aspira-
tions and expectations and instills attitudes toward work that increase
productivity, the effects of such changes (or even the direction) are not
always obvious and may not be felt for several years.
The most comprehensive data measuring the effects of diversification
are from the achievement tests administered to students at the end of their
secondary schooling. Any differences in the scores by school type or
curriculum could be meaningful irn two situations. First, if diversification
has already been accepted in principle and the question (as in Colombia) is
what type of school should implement it, then differences in learning
among students of the same specialization but in different types of school
are relevant. Second, if differences in learning achievement influence the
ability to find employment, then the knowledge and skills gained could
potentially raise worker productivity, provided the labor market is orga-
nized in such a way that they would be used.
It is always difficult, if not dangerous, to derive firm policy implications
on the basis of a single study of young graduates in two countries. Ideally,
one might have to repeat this study in several other countries and wait
years or decades until the young cohorts of graduates reach their full
earning capacity. But decisions mL st be made now to shape a secondary
educational system and provide curricula. It would be of little use to tell
the policymakers to postpone thesz decisions, even though they must be
based on intuition rather than on lfacts. It is probable, however, that the
longer the time perspective, the looser would be the relationship between
initial educational experience and subsequent occupation and earnings,
and this would weaken rather than strengthen the case for diversified or
prevocational curricula.
It has been argued that diversified curricula should be instituted on
purely educational grounds. Educators have often suggested that a good
general education should comprise both academic and manual compo-



SUMMARY AND CONCLUSIONS                                      227
nents in the interest of developing the "whole person." This argument has
merit, but a perusal of documents justifying such innovation suggests that
economic considerations were the predominant impetus for the develop-
ment of diversified secondary curricula, and the programs must stand or
fall on whether they meet these economic objectives. Diversified programs
cannot be justified on the basis of the null hypothesis that they are no
better and no worse than conventional curricula. Continued implementa-
tion or expansion can be justified only on the basis of evidence that the
internal or external efficiency of diversified schools is superior to that of
conventional schools. The data do not support this criterion.
The findings of this study that do have clear policy implications seem to
be the following:
* It must not be assumed that new structures and curricula will neces-
sarily broaden the social basis of recruitment and make secondary
schooling more accessible to the poor. Any type of school expansion
would increase the enrollment rate among the less privileged social
groups. The problem of equal access to educational opportunity
cannot be addressed through curricular reform alone.
* Curriculum diversification is expensive; the cost per student in a
diversified school can be double that in a conventional (academic)
school, mainly because of the specialized equipment used. Therefore,
policymakers should weigh this cost against the potential extra ben-
efits (monetary and nonmonetary) that such schools confer to the
students and to society at large.
* This study has failed to provide evidence that the measurable mone-
tary benefits of diversification are greater than those of conventional
education. All forms of secondary education increase the productivity
of the worker, but diversified schools have not yet proven better in this
respect than conventional schools.
* Another rationale for such schools-to improve the fit between the
school and the world of work-is not supported by our findings.
Graduates from both types of schools spend roughly the same amount
of time in finding employment.
* Neither has the argument that prevocational education provides bet-
ter preparation for further study been borne out by our data. Those
who have specialized in a particular curriculum (such as agriculture or
technical trades) in secondary school are often found to study a
completely different subject (such as liberal arts) at a university.
* Diversified curricula are difficult to implement. The need for new
instructional materials, teachers with new qualifications, and associ-
ated laboratories and equipment can present logistical problems that
cause the reform to fail at the outset. Unless a country has the



228   DIVERSIFIED SECONDARY EDUCATION AND DEVELOPMENT
necessary infrastructure to implement a diversified curriculum, the
reform may appear only on paper and never become a reality. To put
it differently, curriculum diversification requires more than new
buildings and equipment; ir. must be accompanied by a series of
software components that may well prove to be a bottleneck in
implementing the reform, if reform is in fact warranted.
Paradoxically, it seems that the lower the overall level of a country's
development, the weaker the case for introducing a diversified curriculum.
The more developed the country, the more it may be able to afford
diversification. This policy conclusion is exactly the opposite of what
actually happens: the poorer the country, the greater the pressure for
making the secondary school curriculum "more relevant to the world of
work."



References
Ahmed, Manzoor, and Philip H. Coombs, eds. 1975. Education for Rural De-
velopment: Case-Studies for Planners. Prepared by the International Council for
Educational Development under sponsorship of the World Bank and the United
Nation's Children's Fund. New York: Praeger.
Anderson, C. A. 1983. "Social Selection in Education and Economic Develop-
ment." Washington, D.C.: The World Bank, Education Department. Processed.
Armer, M., and Z. Youtz. 1971. "Formal Education and Individual Modernity in
an African Society." American Journal of Sociology, vol. 76, no.4, pp. 604-26.
Austen, R. A. 1968. Northwest Tanzania under German and British Rule. New
Haven, Conn.: Yale University Press.
Bellaby, P. 1977. The Sociology of Comprehensive Schooling. London: Methuen.
Benavot, A. 1983. "The Rise and Decline of Vocational Education." Sociology of
Education, vol. 56, no. 2, pp. 63-76.
Benoit, A. 1974. Changing the Educational System: A Colombian Case-Study.
Munich: Weltforum Verlag.
Biede, K. 1970. "The Pattern of Education and Economic Growth." Economic
Record, vol. 46 (September).
Blaug, Mark. 1973. Education and the Employment Problem in Developing
Countries. Geneva: International Labour Office.
. 1979. "The Quality of Population in Developing Countries, with Particu-
lar Reference to Education and Training." In P. Hauser, ed., World Population
and Development: Ch7allenges and Prospects. Syracuse, N.Y.: Syracuse Uni-
versity Press.
Blumenthal, I., and C. Benson. 1978. Educational Reform in the Soviet Union:
Implications for Developing Countries. World Bank Staff Working Paper
no. 288. Washington, D.C.
Boesen, J., B. Madsen, and T. Moody. 1977. "Ujamaa: Socialism from Above."
Uppsala, Sweden. Scandinavian Institute of African Studies. Processed.
Boudon, R. 1974. Education, Opportunity and Social Inequality: Changing Pros-
pects in Western Society. New York: Wiley.
Bowles, S. 1971. "Cuban Education and the Revolutionary Ideology." Harvard
Education Review, vol. 41, no. 4.
Breneman, D., and S. Nelson. 1981. Financing Community Colleges: An Eco-
nomic Perspective. Washington, D.C.: Brookings Institution.
Bukhari, N. 1968a. "Issues in Occupational Education: A Case Study in Jordan."
229



230                                                    REFERENCES
Stanford International Developmcnt Education Center, Studies on Content and
Methods of Education for Development. Stanford, Calif.: SIDEC.
-. 1968b. "Issues in Occupat onal Education: A Case Study in Tunisia."
Stanford International Development Education Center, Studies on Content and
Methods of Education for Development. Stanford, Calif.: SIDEC.
Cameron, J., and W. A. Dodd. 197C. Society, Schools and Progress in Tanzania.
Oxford, Eng.: Pergamon Press.
Carnoy, Martin, and J. Wharteim. 1976. Economic Change and Education Re-
form in Cuba. Baltimore, Md.: Johns Hopkins University Press.
-. 1977. "Socialist Ideology and the Transformation of Cuban Education."
In J. Karabel and A. H. Halsey, eds., Power and Ideology in Education. New
York: Oxford University Press.
Carroll, A. B., and L. A. Ihnen. 1967. "Cost and Returns for Two Years of
Post-Secondary Technical Schoolir.g: A Pilot Study." Journal of Political Econ-
omy, vol. 75 (December).
Castro, C. de Moura. 1976. Vocational Education and the Training of Industrial
Labour in Brazil. Report Studies C-20. Paris: Unesco.
-.     1983. "High Technology in Intermediate Countries? The Case of Brazil."
Brasilia: Instituto de Planejamentc Economico e Social. June. Processed.
Clark, D. H. 1983. How Secondary School Graduates Perform in the Labor
Market: A Study of Indonesia. World Bank Staff Working Paper no. 615.
Washington, D.C.
Cliffe, L. 1973. "The Policy of Ujamaa Vijijini and the Class Struggle in Tanzania."
In L. Cliffe and J. Saul, eds., Socialism in Tanzania. Vols. 1 and 2. Nairobi: East
African Publishing House.
Clignet, R., and P. Foster. 1966. The Fortunate Few: A Study of Secondary Schools
and Students in the Ivory Coast. Evanston, Ill.: Northwestern University Press.
Cohn, E. 1979. The Economics of Education. Cambridge, Mass.: Ballinger.
Comber, L., and J. Keeves. 1973. Science Education in Nineteen Countries: An
Empirical Study. Stockholm: Almqvist and Wicksell.
Conroy, W. G., Jr. 1979. "Some Historical Effects of Vocational Education at the
Secondary Level." Phi Delta Kappa, vol. 61, no. 4, pp. 267-71.
Corazzini, A. J. 1968. "The Decision to Invest in Vocational Education: An
Analysis of Costs and Benefits." Journal of Human Resources, vol. 3, Supple-
ment.
Court, D. 1972. Village Polytechnic Leavers: The Maseno Story. Working Paper
no. 70. Nairobi: University of Naitobi, Institute of Development Studies.
- 1976. "The Educational System as a Response to Inequality in Tanzania
and Kenya." Journal of Modern Ajrican Studies, vol. 14, no. 4, pp. 661-90.
Crossley, M. 1984a. "Strategies for Curriculum Change and the Question of
International Transfer." Journal of Curriculum Studies, vol. 16, no. 1,
pp. 75-88.
.1984b. "Relevance Education, Strategies for Curriculum Change and Pilot



REFERENCES                                                      231
Projects: A Cautionary Note." International Journal of Educational Develop-
ment, vol. 4, no. 3, pp. 245-50.
Cumming, D. E. 1984. "An Evaluation of Industrial Education in Kenya: The
Contribution of Cost Studies." Edinburgh: Moray House College of Education.
Processed.
DANE (Departamento Administrativo Nacional de Estadistica). 1976. XIV Censo
Nacional de Poblaci6n y III de Vivienda. Resultados Provisionales. Bogota,
Colombia.
Davis, G., and R. Lewis. 1975. Education and Employment: A Future Perspective
of Needs, Policies and Programs. Lexington, Mass.: D. C. Heath.
Diyasena, W. 1976. Pre-vocational Education in Sri Lanka. Paris: Unesco.
Dodd, William A. 1969. Education for Self-Reliance in Tanzania: A Study of Its
Vocational Aspects. New York: Teachers College Press.
Dore, R. 1975. The Diploma Disease. London: Allen and Unwin.
Dumont, R. 1974. Is Cuba Socialist? London: Andre Deutsch.
Eggleston, J., ed. 1983. Work Experience in Secondary Schools. London and
Boston: Routledge and Kegan Paul.
Evans, D. R., and G. L. Schimmel. 1970. The Impact of a Diversified Educational
Program on Career Goals at Tororo Girls High School in the Context of Girls
Education in Uganda. Amherst, Mass.: University of Massachusetts, Center for
International Education.
Fields, G., and P. Schultz. 1980. "Regional Inequality and Other Sources of Income
Variation in Colombia." Economic Development and Cultural Change, vol. 28,
no. 3 (April).
Figueroa, M., R. Prieto, and F. Gutierrez. 1974. "The Basic Secondary School in
the Country: An Educational Innovation in Cuba." Prepared for the Interna-
tional Bureau of Education. Paris.
Ford,J. 1969. Social Class and the Comprehensive School. London: Routledge and
Kegan Paul.
Foster, Philip J. 1965. "The Vocational School Fallacy in Development Planning."
In C. A. Anderson, and Mary J. Bowman, eds., Education and Economic
Development. Chicago: Aldine.
. 1966. Education and Social Change in Ghana. Chicago: University of
Chicago Press.
.1983. "Essays on Curriculum Diversification." Washington, D.C.: World
Bank, Education Department. Processed.
Freeman, Richard. 1974. "Occupational Training in Proprietary Schools and
Technical Institutes." Review of Economics and Statistics, vol. 56 (August)
pp. 310-18.
Fuller, William. 1976. "More Evidence Supporting the Demise of Pre-employment
Vocational Trade Training: A Case Study of a Factory in India." Comparative
Education Review, vol. 20, no. 1.
Furley, 0. W., and T. A. Watson. 1977. A History of Education in East Africa.
New York: NOK Publishers.



232                                                   REFERENCES
Godfrey, M. 1977. "Education, Prod activity and Income: A Kenyan Case Study."
Comparative Education Review, vol. 21, no. 1.
Gouveia, A. 1972. "Economic Development and Changes in the Composition of
the Teaching Staff of Secondary Schools in Brazil." In Thomas J. La Belle, ed.,
Education and Development: Latin America and the Caribbean. Los Angeles:
University of California, Latin America Center.
Grant, N. 1964. Soviet Education. London: University of London Press.
Gray, J. 1976. "Stalin, Mao and the Future of China." New Society, vol. 36,
no. 704, pp. 9-11.
Grubb, W. Norton. 1979. "The Phoenix of Vocational Education: Implications
for Evaluation." In The Planning Papers for the Vocational Education Study.
Washington, D. C.: National Institute of Education.
Haddad, W. 1979. "Diversified Secondary Curriculum Study: A Review of World
Bank Experience." Washington, D.C.: World Bank, Education Department.
Processed.
Hanushek, E., and J. Jackson. 1977. Statistical Methods for Social Scientists. New
York: Academic Press.
Heyneman, Stephen. 1984. "Diversifying Secondary School Curricula in Develop-
ing Countries: An Implementation History." Washington, D.C.: World Bank,
Economic Development Institute. Processed.
Hinchliffe, K. 1983. "Cost Structures of Secondary Schooling in Tanzania and
Colombia." Washington D.C.: World Bank, Education Department. Processed.
Holsinger, D. 1973. "The Elementary School as Modernizer: A Brazilian Study."
International Journal of Comparative Sociology, vol. 14, nos. 3, 4.
Hu, T., M. L. Lee, E. W. Stromsdorier, and J. J. Kaufman. 1971. "Economic
Returns to Vocational and Comprehensive High School Graduates."Journal of
Human Resources, vol. 6, pp. 25-50.
Husen, T. 1975. Social Influences on Educational Career. Paris: Organisation for
Economic Co-operation and Develcpment.
ICFES (Instituto Colombiano Para el Fomento de la Educaci6n Superior). 1979.
Historia Estadistica de la Educaci6n Superior Colombiana, 1960-1977.
Bogota.
. 1980. Estadisticas de la Educaci6n Superior. Bogota.
Inkeles, Alex. 1983. Exploring Individual Modernity. New York: Columbia Uni-
versity Press.
Inkeles, Alex, and D. Smith. 1974. Becoming Modern. Cambridge, Mass.: Har-
vard University Press.
Instituto Nacional de Educaci6n Media (INEM). 1973. "Principios Fundamentales
de la Educaci6n Media Diversificada en Colombia." Medellin.
Ishumi, A. G. M. 1974. "The Educated Elite: A Survey of East African Students
at a Higher Institution of Learning." Rural Africana, vol. 25, pp. 65-72.
Jamison, Dean T. and Lawrence J. Lau. 1982. Farmer Education and Farm
Efficiency. Baltimore, Md.: Johns Hopkins University Press.



REFERENCES                                                      233
Jolly, R., ed. 1969. Education in Africa: Research and Action. Nairobi: East
African Publishing House.
Jolly, R., E. de Kadt, H. Singer, and F. Wilson. 1973. Third World Employment:
Problems and Strategy. London: Penguin.
Kahl, J. 1968. The Measurement of Modernism: A Study of Values in Brazil and
Mexico. Austin: University of Texas Press.
Karweit, N., 1983. "Time on Task: A Research Review." Report no. 232. Balti-
more, Md.: Johns Hopkins University, Center for Social Organization of
Schools.
King, Kenneth. 1977. The African Artisan: Education and the Informal Sector in
Kenya. London: Heinemann.
. 1978. Education and Self-Employment. Paris: Unesco and International
Institute for Educational Planning.
Kipkorir, B. 1975. "Kenya: Development and Co-ordination of the Non-formal
Programs." In Manzoor Ahmed and Philip H. Coombs, eds., Education for
Rural Development: Case-Studies for Planners. New York: Praeger.
Klingelhofer, E. 1967. "Occupational Preference of Tanzanian Secondary School
Pupils." Journal of Social Psychology, vol. 72, no. 2, pp. 149-59.
Knight, J. B., and R. H. Sabot. 1981. "Correlates of Educational Attainment in
Tanzania." Discussion Paper no. 81-33. Washington, D.C.: World Bank,
Population, Health, and Nutrition Department.
Lauglo, J. 1983. "Concepts of 'General Education' and 'Vocational Education'
Curricula for Post-Compulsory Schooling in Western Industrialised Countries:
When Should the Twain Meet?" Comparative Education, vol. 19, no. 3.
Lecht, A. 1979. Evaluating Vocational Education: Policies and Plans for the
1970s. New York: Praeger.
Leiner, M. 1975. "Cuba: Combining Formal Schooling with Practical Experi-
ence." In Manzoor Ahmed and Philip H. Coombs, eds., Education for Rural
Development: Case-Studies for Planners. New York: Praeger.
Lema, A. A. 1978. Education for Self-Reliance: A Brief Survey of Self-Reliance
Activities in Some Tanzanian Schools and Colleges. Dar es Salaam: University of
Dar es Salaam, Institute of Education.
Lewin-Epstein, N. 1981. "Vocational Education." In J. Coleman, V. Bartot,
N. Lewin-Epstein, and L. Olson, eds., "High School and Beyond: Policy Issues
and Research Design." Washington, D.C.: National Center for Educational
Statistics.
Lewis, A. 1966. Development Planning. London: Allen and Unwin.
Lillis, K., and D. Hogan. 1983. "Attempts to Diversify Secondary School Curricula
in Developing Countries: A Literature Review and Some Additional Hvpotheses
Concerning Cases of Failure." Brighton, Eng.: University of Sussex, Centre for
Educational Technology. Processed.
Lofstedt, Jan-Ingvar. 1981. Chinese Educational Policy: Changes and Contradic-
tions, 1940-1979. Stockholm: Almqvist and Wicksell.



234                                                   REFERENCES
Lourie, S. 1978. Production School, and Rural Employment in Panama. Paris:
Unesco.
Low-Maus, R. 1971. Compendiurn of the Colombian Educational System.
Bogota: Editorial Andes.
Loxley, William. 1983. "A Comparison of Achievement Outcomes by Curriculum
Track in Selected OECD Countries: An Analysis of IEA Data." Washington, D.C.:
World Bank, Education Department. Processed.
Mbilinyi, M. 1974a. "Education for Rural Life or Education for Socialist Trans-
formation." Dar es Salaam: University of Dar es Salaam, Institute of Education.
1974b. "The Decision to Ecucate in Rural Tanzania." Dar es Salaam:
University of Dar es Salaam, Institute of Education.
McCarthy, F. D., J. A. Hanson, ancl S. Kwon. 1983. "Structural Change and
Sources of Growth in Colombia." Washington, D.C.: World Bank, Economic
Analysis and Projections Departmcnt. Processed.
McClelland, D. 1961. The Achieving Society. New York: Free Press.
McGinn, Noel F., and E. T. Balart. 1980. "Una Evaluaci6n de la Educaci6n Media
Tecnica en El Salvador." Revista Latinoamericana de Estudios Educativos,
vol. 10, no. 2.
McPherson, A., J. Gray, and D. Raffe. 1983. Reconstruction of Secondary Educa-
tion: Theory, Myth and Practice since the War. London and Boston: Routledge
and Kegan Paul.
Meyer, R. 1980. An Economic Analysis of Vocational Education. Columbus,
Ohio: National Center for Researchi in Vocational Education.
Meyer, J. 1979. "National Economic Development, 1950-1970." In John W.
Meyer and Michael T. Hannan, eds., National Development and the World
System. Chicago: University of Chicago Press.
Mincer, Jacob. 1974. Schooling, Experience and Earnings. New York: National
Bureau of Economic Research.
Ministerio de Educaci6n Nacional, Instituto Colombiano de Pedagogia (ICOLPE).
1974. La Evaluaci6n Institucional Aplicada a la Educaci6n. Bogota: Centro
Nacional de Informaci6n Pedag6gica (CENDIP).
Ministerio de Educaci6n Nacional, Banco Mundial, and Instituto SER. 1982.
Evaluaci6n del Rendimiento Externo del Bachillerato Colombiano: Cohorte
1978. Bogota.
Mmari, G. 1976a. Implementation of t6e Musoma Resolutions. Papers in Educa-
tion and Development no. 3. Dar es Salaam: University of Dar es Salaam.
December.
. 1976b. Directive on the Implementation of Education for Self-Reliance,
Dar es Salaam: Ministry of Nationa;. Education. May.
Moock, Peter R., and Joanne Leslie. Fo:-thcoming. "Childhood Malnutrition and
Schooling in the Terai Region of Nepal." Journal of Development Economics.
Moris, J. R. 1972. Agriculture in the Schools: The East African Experience.
Nairobi: East Africa Academy.



REFERENCES                                                      235
. 1966. "The Impact of Secondary Education upon Students' Attitudes
Towards Agriculture: Some Preliminary Considerations." East African Journal
of Rural Development, vol. 1, no. 1.
Narman, A., C. Hildesson, H. Sjolund, and J. Soderstrom. 1984. "What Happens
to Kenyan Secondary School Students with Industrial Education: A Tracer Study
One Year after the KCE-Exam 1983." Goteborg, Sweden: University of
Goteborg. Processed.
Ntuah, R. R. 1981. "Pros and Cons of the Diversification of Secondary Education
in Tanzania." In I. M. Omari and H. J. Mosha, eds., Decentralization of
Educational Research and Evaluation Capacities in Tanzania. The Hague:
Centre for the Study of Education in Developing Countries.
Nyerere, J. K. 1967. Education for Self-Reliance. Dar es Salaam: Government
Printer.
. 1977. The Arusha Declaration Ten Years After. Dar es Salaam: Govern-
ment Printer.
Olson, J. 1974. "Secondary Schools and Elites in Kenya." Comparative Education
Review, vol. 16, no. 1.
Omari, I. M. 1984. "The Predicament of Diversification and Vocationalization of
Secondary Education in Developing Countries: Experiences from Tanzania."
Washington, D.C.: World Bank, Education Department. Processed.
Orata, Pedro T. 1972. Self-Help Barrio High Schools: The Story of 250,000
Students, Earning their Education and Preparing Themselves for Life. Singa-
pore: Eastern Universities Press for Seameo Regional Center for Educational
Technology.
Oxtoby, Robert. 1977. "Vocational Education and Development Planning:
Emerging Issues in the Commonwealth Caribbean." Comparative Education,
vol. 13, no. 3.
Passow, A., and others. 1976. The National Case Study: An Empirical Study of
Twenty-one Educational Systems. New York: Wiley.
Perry, W. 1981. Training for Self-Reliance Project: A Formative Evaluation Prog-
ress Report. Maseru, Lesotho: Government of Lesotho/World Bank, Diver-
sification of Secondary Schools Curriculum Project.
Plowden, B. 1967. Children and Their Primary Schools: A Report of the Central
Advisory Council of England. London: Her Majesty's Stationary Office.
Price, R. F. 1973. "The Part-Work Principle in Chinese Education." Current Scene,
vol. 11, no. 9 (September).
Psacharopoulos, George. 1984. "The Contribution of Education to Economic
Growth: International Comparisons." InJ. Kendrick, ed., International Produc-
tivity Comparisons. Cambridge, Mass.: Ballinger.
. 1983. "Education and Private versus Public Sector Pay." Labour and
Society, vol. 8, no. 2 (April-June).
.198 la. "Lifetime Profiles of Earnings and Employment: A Survey." Social
Science Information, vol. 20, no. 4/S.



236                                                   REFERENCES
. 1981b. "Returns to Education: An Updated International Comparison."
Comparative Education, vol. 17, no. 3.
-. 1973. Returns to Education: An International Comparison. Amsterdam
and New York: Elsevier and Jossey-Bass.
Psacharopoulos, George, and Keith IHinchliffe. 1983. "Tracer Study Guidelines."
Washington, D.C.: World Bank, ELducation Department. Processed.
Psacharopoulos, George, and Williarn Loxley. 1984. "Diversified Secondary Cur-
riculum Study (DiSCuS)-Survey Instruments." Washington, D.C.: World
Bank, Education Department. Processed.
Psacharopoulos, George, E. Velez, and A. Zabalza. 1985. Una Evaluaci6n de la
Educaci6n Media Diversificada en Colombia. Bogota: World Bank and Instituto
SER de Investigaci6n.
Psacharopoulos, George, and A. Zabalza. 1984a. The Destination and Early
Career Performance of Secondary School Graduates in Colombia: Findings
from the 1978 Cohort. World Bank Staff Working Paper no. 653. Washington,
D.C.
-. 1984b. "The Effect of Diversified Schools in Employment Status and
Earnings in Colombia." Economics of Education Review, vol. 3, no. 4,
pp. 315-31.
Puryear, Jeffrey M. 1979. "Vocational Training and Earnings in Colombia: Does a
SENA Effect Exist?" Comparative Education Review, vol. 23, no. 2.
Roca, 5. 1977. "Cuban Economic P'olicy in the 1970s: The Trodden Paths."
Studies in Comparative International Development, vol. 12.
Rosenbaum, J. B. 1976. Making Inequality: The Hidden Curriculum of High
School Tracking. New York: Wiley.
Ruddell, D. 1979. "Vocationalising Ghana's Schools: Purpose and Product."
Ph.D., dissertation. Birmingham, Eng.: University of Birmingham, Faculty of
Commerce.
Rumberger, R. 1983. "Dropping Out of High School: The Influence of Race, Sex,
and Family Background." American Educational Research Journal, vol. 20,
no. 2.
Sanyal, B. C., and M. J. Kinunda. 1977. Higher Education for Self-Reliance: The
Tanzanian Experience. Paris: Unesco and International Institute for Educa-
tional Planning.
Sato, Kuniu. 1974. An Alternative Approach to Vocational Education. Bangkok:
International Labour Organisation, Asian Regional Team for Employment
Promotion.
Schiefelbein, Ernesto. 1979. Education! and Employment in Latin America. Report
Studies. Paris: Unesco.
Schnaiberg, A. 1971. "The Modernizing Impact of Urbanization: A Causal Analy-
sis." Economic Development and Cultural Change, vol. 20, no. 1, pp. 80-104.
Silberman, H. 1978. "The Relationship of Education and Work." New Directions
for Education and Work. Spring.



REFERENCES                                                     237
Silvey, J. 1969. "Unwillingly to School: The Occupational Attitudes of Secondary
School Leavers in Uganda." In R. Jolly, ed., Education in Africa: Research and
Action. Nairobi: East African Publishing House.
Squire, Lyn. 1981. Employment Policy in Developing Countries. New York:
Oxford University Press.
Sullivan, G. 1981. "Secondary School Leavers and the Labour Market in Swazi-
land." International Journal of Educational Development, vol. 1 (September).
Tanzania Ministry of National Education. 1979a. "Estimated Unit Cost per
Student and Total Cost of Running Secondary Schools and Colleges for the Year
1978/79." Dar es Salaam.
. 1979b. "Guide to Headmasters (Kiongozi cha Mkuu wa Shule ya Sekon-
dari)." Dar es Salaam.
1980. Basic Facts about Education in Tanzania. Dar es Salaam.
1984. Educational System in Tanzania Towards the Year 2000. Dar es
Salaam.
Taussig, M. K. 1968. "An Economic Analysis of Vocational Education in the New
York City High Schools." Journal of Human Resources, vol. 3 (Supplement).
Tchen, Y. 1977. "Education and Productive Work in China." Prospects, vol. 7,
no. 3.
Thailand Ministry of Education and Faculty of Education of the University of
Alberta. 1980. "Final Report of the Thai-Alberta Cooperative Assessment
Project Concerning Thirty-two Rural Secondary Schools." Bangkok and
Edmonton. Processed.
Unger, J. 1980. "Bending the School Ladder: The Failure of Chinese Education
Reform in the 1960's." Comparative Education Review, vol. 24, no. 2 June).
Urevbu, A. 0. 1984. "Vocational Education in Nigeria: A Preliminary Appraisal."
International Journal of Educational Development, vol. 4, no. 3, pp. 223-29.
Van Rensburg, P. 1974. Report from Swaneng Hill: Education and Employment
in an African Country. Stockholm: Dag Hammarskjold Foundation.
Varner G. 1965. Educaci6n Secundaria en Colombia. Informe a la Agencia para
Desarrollo Internacional de los Estados Unidos, la Asociaci6n Nacional de
Educaci6n de los Estados Unidos y al Ministro de Educaci6n Nacional de la
Republica de Colombia. Bogota.
Velez, E., and C. Rojas. 1983. "Evaluaci6n del Rendimiento Externo del Bachil-
lerato Colombiano Que Saben y Cuales son Las Expectativas Educacionales y
Ocupacionales del Bachiller Colombiano." Bogota: Instituto SER de Investiga-
cion.
Vulliamy, G. 1980. SSCEP and High School Outstations: A Case Study. Education
Research Unit Research Report no.33. Port Moresby: University of Papua New
Guinea.
Walters, P., and R. Rubinson. 1981. "Educational Expansion and Economic
Growth in the U.S., 1870-1970." Paper presented at the Annual Meeting of the
American Sociological Association, Toronto. August.



238                                                   REFERENCES
Wanasinghe, J. 1982. "A Critical Examination of the Failure of the Junior Second-
ary School Curriculum and Pre-Vocational Studies in Sri Lanka." International
Journal of Educational Developrnent, vol. 2, no. 1, pp. 61-72.
Watson, J. 0. 1977. "Estimated Rates of Return to Investment in Public and
Proprietary Post-Secondary Vocational Education and the Desirability of Sub-
sidizing Proprietary Institutions." Paper presented to the 1977 Western Eco-
nomic Association Conference, Anaheim, Calif.
Weeks, S. 1978. The Foster Fallacy mn Educational Planning. Education Research
Unit Research Report no. 32. Port Moresby: University of Papua New Guinea.
Weisberg, A. 1983. "What Research Has to Say about Vocational Education in the
High Schools." Phi Delta Kappa, vol. 33, pp. 335-59.
Wijemanna, E. L., and G. H. F. Welkala. 1975. "Sri Lanka: Non-formal Education
for Rural Youth." In Manzoor Ahmed and Philip H. Coombs, eds., Education
for Rural Development: Case-Studies for Planners. New York: Praeger.
Wiley, D., and A. Harnischfeger. 1980. High School Learning, Vocational Track-
ing and What Then? Final Report prepared for National Center for Educational
Statistics, Contract no. 300-78-05*16. Evanston, Ill.: Central Midwest Regional
Educational Laboratory.
Williamson, Bill. 1979. Education, Social Structure and Development. New York:
Holmes and Meier.
Woods, E. M., and W. Haney. 1981. Does Vocational Education Make a Differ-
ence? Final report prepared for the National Institute of Education. Contract
no. 400-79-0026. Cambridge, Mass.: Huron Institute.
Zachariah, M., and A. Hoffman. 1984. "Gandhi and Mao on Manual Labor in the
School: A Retrospective Analysis. " Paper presented at the Fifth World Congress
of Comparative Education, Paris, July 2-6, 1984. University of Calgary, Depart-
ment of Educational Policy and Administrative Studies. Processed.
Zymelman, Manuel. 1976. The Economic Evaluation of Vocational Training
Programs. Baltimore, Md.: Johns Hopkins University Press.



Index
(Page numbers in italics indicate figures or tables.)
Academic performance: Colombia (1981     Centros Auxiliares de Servicios Docentes
cohort), 66-67, 71, 72, 73, 76; Tanza-  (CASD), 53
nia (1981 cohort), 157; of vocational vs.  Chile, 29
academic students, 26, 27, 156; testing  China, 16, 25, 31
of, 48                                 Clark, D. H., 14, 28
Academic programs: INEM and non-INEM     Class time: achievement and (Colombia),
graduates in, 98; prevocational studies  77-78, 210-11; achievement and (Tan-
added to, 12; quality of (Colombia), 59;  zania), 155, 158-59,160, 161, 162,163
rate of return to (Colombia), 125; in  Cognitive achievement: in Colombia, 66-
Tanzania, 158-59; vocational subjects    67, 68; cost-effectiveness of (Colombia),
in, 30                                  89, 91, 213-14; cost-effectiveness of
Achievement scores: course time and, 77-  (Tanzania), 176-77,215; differences in,
78, 155, 158-59, 210-11; curriculum     25-27; in Tanzania, 154-55. See also
and, 26-27; gains in (Tanzania), 168,   Achievement scores
169; INEM program and gains in (Co-   Colombia, 6, 13, 28; benefits from voca-
lombia), 75, 77; influence of home and  tional schooling in, 29; diversification
school on, 211; by school type and cur-  in, 17-20, 35; economy of, 54; educa-
riculum (Colombia), 66, 67, 68, 71, 74,  tional system in, 55, 57, 59, 60
75, 76; by student characteristics (Co-  Colombia case study: background of, 54-
lombia), 70; by student characteristics  63; class of 1981 in, 92-105; class of
(Tanzania), 155; of Tanzanian students,  1978 in, 111-32; curriculum costs and
154-55                                  outcomes in, 85-92; decision to work or
Africa: student expectations in, 24-25;   study and, 105-11; full-time students
vocational education in, 11-12          (1982) in, 96-98; graduates looking for
Angola, 16                                work in, 102-03; internal efficiency of
Apprenticeship programs, 8, 13, 29, 115   schools and, 65-85, 214-15; sampling
Armer, M., 49                             1981 cohort in, 60-62; sampling 1978
Asia and the Pacific, vocational education  cohort in, 111, 113; socioeconomic pro-
in, 13-14                               file of 1981 cohort in, 63-65; summary
Attitudes. See Modernity; Student ex-     of findings on 1981 cohort in, 103-05;
pectations                              summary of 1982 tracer study in, 110-
11
Balart, E. T., 30                       Corazzini, A. J., 29
Barbados, 30                            Costa Rica, 13
Benavot, A., 32                         Court, D., 12
Biede, K., 32                           Cuba, 16, 25
Brazil, 13, 25, 30, 49                  Curricula: cognitive achievement and, 26-
Bukhari, N., 29                           27; cost-benefit studies of, 28; in INEM
schools, 87; noncognitive outcomes of,
Cameroon, 24                              48-50; unit costs of (Colombia), 85-88,
Career options, diversification and, 9, 13  212-13; unit costs of (Tanzania), 175-
Castro, C. de M., 30                      76, 214-15
239



240                                                                    INDEX
Curriculum costs: cognitive achievement   El Salvador, 13, 30
and  (Colombia), 89-91; cognitive      Employment: factors influencing (Co-
achievement and (Tanzania), 176-79;       lombia), 118-19, 122, 123; family con-
per student (Colombia), 85-87; per stu-   nections and (Colombia), 100; govern-
dent (Tanzania), 175-76. See also Social  ment (Tanzania), 191,199, 200; rates of
costs of education                        (Colombia), 118-19, 121; salaried
(Tanzania), 195, 198; search for (Co-
Development. See Economic development      lombia, 1982), 102-03, 104; search for
Diversification: assumptions related to, 3-  (Tanzania, 1982), 191-93; students'
6, 9, 216-24; career options and, 13;    plans to seek (Colombia), 82, 83. See
debate on, 8-11; economic argumeiit      also Job search; Labor market; Unem-
for, 9-10; effects of, 22, 37-38; failure  ployment
of (El Salvador), 30; justification for  England, 26
continuing, 227; in Latin America, 13;  Ethiopia, 16
objectives of, 18-21; political objectiv,es  Europe, secondary education in, 15-16,
and, 209; self-reliance and (Tanzania),  23-24, 25
139-40; sociopolitical considerations
and, 10-11; types of, 16-17            Factory schools (U.S.S.R.), 16
Diversified schools, 3; economic impact of,  Family characteristics: by school type and
216; economic impact of (Colombia),       curriculum (Colombia), 65; in Tanza-
217-21; economic impact of (Tanza-        nia, 145, 146, 147, 148. See also
nia),221-24; in El Salvador, 30; evalua-  Socioeconomic background
tion of, 21; in Tanzania, 144; targer-  Father's occupation, income, and educa-
control group comparison of, 40, 41;      tion: Colombia study, 63-64; students'
teachers for, 35                          destinations and (Colombia), 96.
Dodd, W. A., 49-50                        Finland, 26
Dominican Republic, 13                    Foster, P. J., 24
France, 26
Earnings: anticipated (Colombia), 84-85,
97; anticipated (Tanzania), 173, 175,   Germany, 26
191, 194; determinants of, 110, 112,   Ghana, 24
119,123, 124,202,203; forgone, 212--   Gouveia, A., 25
13; school type and cognitive skills anm  Graduates, secondary school: destinations
(Colombia), 100, 102, 107, 109-10        of (Colombia), 93-96; destinations of
Economic development: diversification      (Tanzania), 181-84; earnings determi-
and, 3-4, 225, 227-28; improvements       nantsfor (Colombia, 1978 cohort), 119,
in labor force (Colombia) and, 54; tradi  124; earnings determinants for (Co-
tional patterns and values and, 48; voca  lombia, 1981 cohort), 110, 112; earn-
tional education and, 11-16, 31-33        ings expectations of (Colombia), 97;
Economic efficiency, 9-10                   earnings of vocational and academic,
Education: acquisition of modern atti       31, 100, 102, 119, 123, 124, 194; em-
tudes and, 78-81; costsof, 212; demand    ployment characteristics of (Colombia),
for free, 10; economically valuable, 33:  119,121,122,123,124; employment of
inadequacies of traditional, 9; measures  (Tanzania), 188-91; enrollment in
of value of, 36-37; reason for not con-  Form V (Tanzania), 184-87; occupa-
tinuing, 99; returns to (Tanzania), 204-  tions of (Colombia), 100, 101; opinions
05, 224-25; goals of (Tanzania), 138-    about current (post-Form IV) studies
41, 144. See also Postsecondary studies;  (Tanzania), 188; opinions about current
Vocational education                      (postsecondary) studies  (Colombia),
Educational aspirations. See Student ex-    96-97; opportunities for Form IV (Tan-
pectations                                zania), 180-81; part-time occupations



INDEX                                                                     241
of (Colombia), 10 1-02; predicting post-  Kipkorir, B., 12
secondary study of (Colombia), 107,     Knight, J. B., 175
114-15, 117; response rate of (Tanza-  Korea, 14
nia), 179-80; still seeking employment
(Tanzania), 199, 201; target occupa-   Labor: meeting demands for skilled, 9;
tions of (Colombia), 103, 105; unem-      productivity of (Colombia), 54; supply
ployment of, 100-01, 102-03,191-92,       (Tanzania), 138-40
193, 194. See also Students            Labor market: in Africa, 25; in Colombia,
Grant, N., 49                               217; education and, 3, 14; INEM gradu-
Greece, 29                                  ates entering, 118-19; for secondary
Guinea Bissau, 16                          school graduates, 8; in Tanzania, 217
Guyana, 13                                Languages: courses in (Tanzania), 159;
need for (Tanzania), 202
Hinchliffe, K., 85                        Latin America: diversification in, 6; voca-
Hoffman, A., 25                             tional education in, 12-13
Holsinger, D., 49                        Lewis, A., 78
Hungary, 26                               Lickert scale, 168
Hu, T., 28                               Loxley, W., 24, 26
iEA study of science achievement, 26-27  McClelland, D., 48
India , 30, 31                           McGinn, N. F., 30
Indonesia, 14, 28                        Methodology. See Study (present)
INEMS. See Institutos Nacionales de Educa-  Mexico, 49
ci6n Media (INEMS)                     Meyer, J., 32
Inkeles, A., 49, 78                      Modernity attitudes: in Colombia, 78-81;
Instituto Interamericano de Capacitaci6n    in Tanzania, 168-70; tests for, 48-50,
Agropecuaria (IcA), 13                    226
Instituto Nacional de Cooperaci6n Educa-        b       1
tiva (INCE), 13
Institutos Nacionales de Educaci6n Media  Nigeria: psychological modernization in,
(INEMS) (Colombia), 17; academic         49; vocational education in, 31
knowledge and, 91-92; academic sub-    Nyerere, J., 138
jects in, 67; curricula of, 85; distribution
of sample in, 113; employment of        Olson, J., 24
graduates of, 118-19; equipment in, 91;
graduates' postsecondary studies, 131,  Papua New Guinea, 13
132; math and science in vocational    Part-time students and workers (Co-
programs of, 67; recurrent costs of, 212;  lombia), 101-02; characteristics of (Co-
six-year program of, 58                  lombia, 1981 cohort), 103
Israel, 29                               Peru, 13
Italy, 26                                Philippines, 28
International Association for the Evalua-  Postsecondary studies: in Colombia, 114-
tion of Educational Achievement (IEA),   15, 116, 117, 118; determinants of spe-
23, 26-27                                cialization in (Colombia), 118, 120;
Jamaica, 13                                perceived value of, 97, 99; type of secon-
Job search (Colombia): 1978 cohort, 126-   dary school and opportunities for (Co-
27,128,129,130;1981cohort,99-100         lombia), 104-05, 115, 118
Jordan, 29                               Prevocational studies, 19; in Sri Lanka, 25
Proprietary schools, 29
Kahl, J., 49                             Psacharopoulos, G., 32, 85
Kenya, 12, 24, 31                        Puryear, J. M., 29



242                                                                    INDEX
Reservation wages. See Wages (minimurn    Servicio Nacional de Aprendizaje Indus-
acceptable)                              trial (SENAI), 13
Response bias, 179                        Skills, 9-10; marketable, 27; testing of
Rubinson, R., 32                            manual, 46n
Smith, D., 49, 78
Sabot, R. H., 175                         Socialist countries, secondary education
Salaries. See Wages                         in, 15-16, 25
Sao Paulo, Brazil, 25                     Social costs of education, 85, 175,212-13,
Schiefelbein, E., 29                        214
Schnaiberg, A., 49                        Social rate of return: defined, 6n; to educa-
Schools: comprehensive, 15, 30, 53; cort-   tion (Tanzania), 204-05
trol (Colombia study), 60-61; earnings,  Socioeconomic background, 23, 25; post-
cognitive skills, and (Colombia), 100,    secondary school activities and (Co-
102, 107, 109-10; effectiveness of rt-   lombia), 95-96; of Colombian students,
form of, 30, 225; external efficiency of  113, 114; of Tanzanian students, 145,
(Colombia), 217-21; external efficienc'   147, 148. See also Family characteris-
of (Tanzania), 221-25; influence of      tics; Father's occupation, income, and
home and (Colombia), 211-12; in-          education
fluence of home and (Tanzania), 214;    Squire, L., 14
influence of quality of (Colombia), 71,  Sri Lanka, 13, 25, 30
75, 77; influence of quality of (Tanza-  SSCEP. See Secondary Schools Community
nia), 165, 168; internal efficiency of    Extension Project (SSCEP)
(Colombia), 65-85, 210-15; internal    Student expectations: educational, 24-25,
efficiency of (Tanzania) 154-55, 158--    45; of job and earnings, 81-85, 171-73
59,165,168-73,214-15;publicvs.pri..    Students: ability measures of, 152-54;
vate (Tanzania), 151-52, 153. See alsc   achievement in different curriculum
Diversified schools; Education; Second-   tracks, 26-27; assessing performance of
ary schools; Vocational education         diversified school, 40; career choices of
Science achievement, curriculum and, 26-    (Tanzania), 172, 174; characteristics of
27                                       (Tanzania), 145, 146; distribution in
Secondary Schools Community Extension       1983 training course of (Colombia,
Project (sscap), 13-14                    1981 cohort), 98; distribution by sex of
Secondary schools: costs of (Tanzania),    (Tanzania), 150-51; educational back-
173, 175-79; enrollment breakdown of     ground by type of school of (Colombia),
(Tanzania), 150-52; manpower re-          81; experiences after schooling of, 27-
quirements and, 4; rate of return studies  31; Form V (Tanzania), 184-87; longi-
of, 29, 125-26, 127; results of different  tudinal information on, 38-41; post-
types of (Colombia), 217-21; results of   secondary school opportunities for, 36;
different types of (Tanzania), 221-24;    psychological modernity of, 48-50,78-
selection for (Tanzania), 149-50; in      81, 168-70, 226; restrictions on occu-
Tanzania, 139, 140-41, 142, 143, 144.     pational choices of, 11; school experi-
See also Diversified schools; Vocational  ence of (Colombia), 81-85, 170; school
schools                                   experience  of (Tanzania), 170-73;
Self-employment: training for (Kenya), 12;  strategy to get into Form V of (Tanza-
training for (Sri Lanka), 14              nia), 202. See also Graduates
Self-reliance, education for, 6, 12, 138,  Study (present): assessment of school qual-
139-40, 209                              ity and, 46; cognitive achievement tests
Servicio Nacional de Adiestramiento en      in, 46-47; general ability tests in, 47-
Trabajo Industrial (SENATI), 13           48; hypothesis testingin, 41-43; metho-
Servicio Nacional de Aprendizaje (sENA),    dology of, 6; policy implications of,
13, 29                                    227-28; research design of, 37-44;



INDEX                                                                     243
scope of, 5-7; social background indica-  Unemployment, 8; of graduates (Colom-
tors and, 45; survey instruments for,     bia), 100-01, 102-03, 127; of gradu-
43-44; test reliability and, 50-51; vari-  ates (Tanzania), 191-92, 193, 194. See
ables in, 44-50, 51-52. See also Co-      also Employment; Job search; Labor
lombia case study; Tanzania case study    market
Swaziland, 30                             United States: science achievement levels
Sweden, 26                                  and curriculum in, 26-27; vocational
education in, 14-15
University of Dar es Salaam, National In-
Tanzania, 6, 16; diversification in, 17-21,  stitute of Education at, 154
35, 139; economy of, 138; educational   University education: excessive demand
goals in, 138-39, 140, 144; educational   for, 4; qualifying for (Tanzania), 141;
system in, 139-41, 142, 144; labor        students planning to seek (Colombia),
supply of, 138-40                         83. See also Postsecondary studies
Tanzania case study: background of, 138-  Urevbu, A. O., 31
41, 144-45; class of 1981 in, 179-94;   U.S.S.R., 15
conclusion of, 205, 207, 209; curricu-
lum costs and outcomes in, 173, 174-    Varner, G., 20
79; internal efficiency of schools in,  Venezuela, 13
154-55, 158-59, 165, 168-73, 214-      Vocational education, 9; benefits from,
15; predicting post-Form V destinations   29-31; class time spent in, 18; cost of,
in, 194-95, 199, 202-05; sample dis-      10, 177; curriculum in (Tanzania), 141,
tribution  by  curriculum  in, 145;       158; difficulty in finding skilled teachers
socioeconomic profile of Form IV stu-     for, 30; economic development and, 11-
dents in, 145, 147, 148; variables in,    16, 31-33; economic value of, 28; in-
206-07                                    fluences on student achievement in, 164;
Taussig, M. K., 29                          jobs and, 30; postsecondary, 29; rate of
Teachers: for diversified secondary educa-  return to, 28-29, 125-26; schools for,
tion, 35; student achievement and (Co-    19; social influences on enrollment in,
lombia), 71; student achievement and      23; stigma on, 28
(Tanzania), 165, 168; vocational train-
ing, 30                                 Wages: industry's reluctance to pay ex-
Thailand, 14, 28                            pected, 30; minimum acceptable (Co-
Tobago, 13                                 lombia), 102, 104, 129, 131. See also
Tracking in schools, 9                      Earnings
Training: access to (Tanzania), 195, 197;  Walters, P., 32
hands-on, 53; on-the-job, 9; post-Form  Wanasinghe, J., 14, 25
IV (Tanzania), 187-88; teacher (Tanza-  World Bank: diversified schools and, 3, 4;
nia), 187, 195. See also Apprenticeship;  projects of, 33-35
Vocational education
Trinidad, 13                              Youtz, Z., 49
Tunisia, 29
Turkey, 28, 49                            Zachariah, M., 25



The most recent World Bank publications are described in the annual
spring and fall lists. The latest -dition is available free of charge from
Publications Sales Unit, Department B, The World Bank, Washington, D.C.
20433, U.S.A.
George Psacharopoulos is chief of the Research Division in the Education
and Training Department of the World Bank. William Loxley, a former
consultant to the World Bank, i, now with the Ford Foundation in In-
donesia.



0 8018 3119 9