~~~~~~~~~~~~~~~~~~~I4 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~ - m 1 E I (I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~L ts 1gs s48-'S!i' >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~L | F7 )2> t.< t = "t2~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~L pE i;|E~~~~~~~~C DiRECTiONS IN DEVELOi;MENT Understanding and a a, n 0 -1 - ri "C lVlLbl& tL111 Lb Social Capital A Multidisciplinary Tool for Practitioners Christiaan Grootaert and Thierry van Bastelaer Editors THE WORLD BANK Washington, D.C. © 2002 The International Bank for Reconstruction and Development/THE WORLD BANK i818 H Street N.Wvv. Washington, D.C. 20433, USA All rights reserved Manufactured in the United States of America First printing June 2002 1 2 3 4 05 04 03 02 The findings, interpretations, and conclusions expressed in this book are entirely those nf the alithors and should not be attributed in anv manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the coun- tries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in UUi VIUlIIC UV dLUL iIImply oUL LtLC Fart L UhUIC 'VYUIU Bak. .V.Fu a-y jud.-r.ent on thL legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. The World Bank encourages dis- semination of its work and normally will promptly grant permission to reproduce por- tions of the work. Permission to photocopy items for internal or personal use, for the internal or per- sonal use of specific clients, or for educational classroom use is granted by the World Barb, provided -h-at th_e a-ppr-opriat-e fee ;.-ai direct-ly to the. CopyrighClearnc Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone 978-750-8400, fax 978-750-4470. Please contact the Copyright Clearance Center before photocopying items. For permission to reprint individual articles or chapters, please fax a request with complete information to tne Republication Department, Copyrignt Clearance Center, fax 978-750-4470. All other queries nn rights and lirenses shouild be addressed to the Office of the Publisher, World Bank, at the address above or faxed to 202-522-2422. Cover photograph from Grameen Bank, copyright Grameen Bank. ISBN 0-8213-5068-4 Library of Congress Cataloging-in-Publication Data has been applied for. Contents Foreword ............................................ vii Acknowledgments ............................................ ix About the Authors ............................................. xi 1. Social Capital: From Definition to Measurement ................................1 The Search for a Definition of Social Capital ...........................................2 The Forms and Scope of Social Capital ............................................. 3 Interdisciplinary Concepts ..............................................5 Rigorous Economic Terminology or Value as an Interdisciplinary Tool? .............................................7 How Does Social Capital Affect Economic Development? ...................8 Approaches to Measuring Social Capital ............................................. 9 Structure of the Book ............................................ 10 Notes ............................................. 12 References ............................................ 13 2. The Social Capital Assessment Tool: Design and Implementation ............................................. 17 Minimum Requirements for an Acceptable Measurement Tool ............................................. 19 Designing the Social Capital Assessment Tool (SOCAT) ..................... 23 Structure and Implementation of the SOCAT ....................................... 24 Notes ............................................. 37 References ............................................. 38 3. Quantitative Analysis of Social Capital Data ..................................... 41 Tabulations of Indicators of Social Capital ............................................ 43 Multivariate Analysis of Social Capital Data ........................................ 57 iii iv UNDERSTANDING SOCIAL CAPITAL Caveats ................................................... 76 Notes ................................................... 79 References ................................................... 81 4. Qualitative Analysis of Social Capital: The Case of Agricultural Extension in Mali ................................................... 85 The Extension System in Mali .................................................... 86 Results and Analysis ................................................... 92 Conclusions and Recommendations ................................................... 101 Appendix 4A: Village Interview Guide ................................................ 104 Appendix 4B: Description of Villages ................................................... 106 Notes ................................................... 107 5. Qualitative Analysis of Social Capital: The Case of Community Development in Coal Mining Areas in Orissa, India ................................................... 108 Database and Methodology ................................................... 109 Context: Overview of Orissa ................................................... 112 Profile of the Study Areas ................................................... 112 Analysis ................................................... 118 Implications for Future Research ................................................... 136 Conclusions: Specific Implications for Community Development ................................................... 139 Notes ................................................... 144 References ................................................... 146 Annexes ................................................... 152 Annex 1. Instruments of the Social Capital Assessment Tool ........... 152 Annex 2. Does Social Capital Facilitate the Poor's Access to Credit? .................................................... 237 Annex 3. Does Social Capital Matter in the Delivery of Water and Sanitation? ................................................... 265 Index ................................................... 297 Tables 3.1 The Social Capital Assessment Tool ................................................... 42 3.2 Active Memberships in Local Organizations in Indonesia, by Province ................................................... 46 3.3 Percentage of Households Participating in Local Organizations in Indonesia, by Province .................................................... 46 (CNTPNTS 3.4 The 10 Most Important Local Organizations in Bolivia .................. 47 3.5 Distribution of Associational Memberships, by Gender, in Burkina Faso ................................................ 48 3.6 Dimensions of Structural Social Capital, by Province and Household Characteristics, [ndonesia ............................................... 49 3.7 Solidarity in Times of Crisis, in Burkina Faso .................................. 52 3.8 Factor Analysis of Trust Variables for Ghana and Uganda ............ 54 3.9 Particination in Collective Action (Gotong Rovong) in Indonesia, by Province ................................................ 56 3.10 Structural and Cognitive Social Canital in Raiasthan. India: Factor Analysis ................................................ 57 3.11 Fartor Analvsis of Multiple Social Capital Variablesq Ghana and Uganda ................................................ 58 3.19 H4oiusehold WAfnne and Social Canital: The Index Model . 61 3.13 Household Welfare and Social Capital: Disaggregating the Social Capital lndex 62................................................ 3.14 Determinants of Total Sales of Trading Firms in Mndr4ic0ier'qr AA in Ma dga ca . . ... . . . . . . .. . . . . . . . ................................................. 6 3.15 The Effect of Social Capital on Household Health in Central Java,> Indonesiaf; 3.16 Household Probability of Adopting Improved Agricultural Practices n Tnnr7ari 70 3.17 Ownership of Assets, by Quintile of Household Expenditure per Capita, in Blnhvia .............. 71 3.18 Poverty and Social Capital: Quantile Regression Results ............... 73 3.19 OLS Regression of vte Determ.inanlts1 of 'JIJA0uJSeoLv-A_T -I. Social Capital in Rajasthan, India .............................................. 75 3.20 OLS Regression of the DeterM,inants nf Sc-ial rapit-a n Dhaka, Bangladesh .............................................. 77 A 1 1MI1--n Tc-1-4-0A4- Ci-.A- S , - . - -lage Tnluedinth Suy ............................................... ...... -- 4.2 Distribution of Interviews ............................................... 88 4.3J LJJiCJCIenLC IJCtVVCeI Vfllages vithLI 1High ar.d Low/I Le~vels of Social Cohesion .............................................. 90 ~~~~~~~- -4.4 Exenio Agen-VlaeItrctomn Otes ........................................9 -2.t _t M2LI1tO1UIL ~rCl5I Lt V IIIC15C LI LC1 a %_LItflL L.LILUIL3.. 5.1 Total Population of Samaleswari, by Category ............................. 125 .2 ill I u ian Lo inga, by .......................................... .I... 5.3 Community-Based Groups in Samaleswari .................................... 129 5.4 Co..ou ruiiu y-Baseud Groups U in Kalng11ac ............................................ 1 5.5 Composition Profile of Village Working Groups in Samaleswari ...............................................13 L 5.6 Degrees of Generalized Interpersonal Trust ................................... 133 VI UNDERJTANDING SOCIAL CAPITAL 5.7 Attitudes toward Management of Common Property Resources and v.illac I..fr ctnih.r. 1.. . .............................................. 134 A3.1 Community-Managed Irrigation: Selected Factors Influencing Performrmanrce of Community Managemnnf .270 A3.2 Community-Based Rural Drinking Water Delivery: Selected Factors I.nfluencing Performance of Community Management .............................................. 280 Figure 1.1 The Forms and Scope of Social Capital ..............................................A Foreword Tlme. ltoftance ofsocial capitaI fr ssanbedeeo,.n i elrc litC 1_11tjJLL(L LI ~UI ~aIt ILl I u~I DUaiilault: UC:VVVFi~lI1tKL ID VVVII IK:%L ognized by now. Anthropologists, sociologists, political scientists, and econo,m.iLsts h-ave i.tei w v-y demons-ra-e h r'ia role of irst; ~LUuUul L~ ILcit I LILCIU IJVVIL VVIyn U liULI LU LItC -LllLILal IUl tJi ftLl tutions, networks, and their supporting norms and values for the success of deuveiuprnet in fLventilonsI The SoCIicald CaplLC.l IJILidLIVe at tUL, VVUIIU Bank has endeavored to contribute to this understanding by focusing on ho-w U to rieasure social capital and its r inpact. liTe SoLial Capital Assessment Tool presented in this book is one of the important products of uhi-s T_r-iuauve 01 uus 1iitiative. The value of the tool goes well beyond its role as a measurement uevice. Building social capital is a core element in tne empowerment pil- lar of the poverty reduction strategy put forth by the World Development Report 200012001. in this context, the nvWVorld Bank has begun to support direct investments in social capital in countries in Africa, Eastern Europe, and Latin America. It is critical that we know now to evaluate tne out- comes of these efforts. Is sustainable social capital being created, and what project components make the critical contributions to tnat outcome?, This book presents a unique opportunity to have a monitoring and eval- uation tool available at the outset of such an investment program. The Social Capital Assessment Tool draws on a rich, multidisciplinary empirical experience, and its application can provide project managers with valuable baseline and monitoring information about social capital in its different dimensions. I hope that this book will faciiitate this applica- tion and in that way contribute to imnproving our knowledge of how to invest effectively in social capital. Steen Lau Jorgensen Director Social Development Department SOCIAL CAPITAL INITIATIVE vii Acknowledgments We would like to express our thanks to all the researchers who partici- pateA srlthe Sc-al Cptal 4 ni;a&ive. Terscsfu olp.;nofhe studies of the Initiative has produced a unique body of evidence on the has led to a significantly improved ability to measure social capital and its :impcts. We also thank our colleagues on the Social Capital Team for their many con,it-ibut,Lonts Susant IAssaf, Gracie O.Ljieng, Gi-`aI` ikOLk , lute iRussing Feldman, and Casper Sorensen. ve acknowledge gratefully the guidance of the Steering Committee and the leadership of Ian Johnson, Vice-President, Environmentally and Socially Sustainable Developmenr, and Steen Jorgensen, Director, Social Development Department. We would like to pay tribute to Ismail Serageidin, whose vision iec ro tne creation or tne Social Capital Initiative. We also honor the memory of Mancur Olson, whose ideas and contributions snaped tne early pnases of tne initiative. The Social Capital Initiative received generous funding from the gov- ernment of Denmark. Christiaan Grootaert T nierry van Basteiaer ix About the Authors Christiaan Grootaert is Lead Economist in the Social Development T_)innrtnm,onf nt thsz- Wnrlr1 R>nnl nnso M>nnncr reSFh Cno; Cn 4.+1 Initiative. His research centers on the measurement and analysis of poverhy, risk, and m11n1rn ili y;A edcao,-tn a- 1a- .marke,.ts hi1- a1or-; and the role of institutions and social capital in development. He has worked'c. in A f-ron A.;. +l,ct 1lAiAA1. T~,-4 -A T-L I-- -,1-~ . in Africa A-ia, hie Middle East, and Eastern Europe. He ht s pub= lished numerous articles and is co-author of Poverty and Social Assistance in Tr-asitiOn Cou-trie, The Pi-y An-a-i of Cild4 TMLabor A Compraiv tL.,IS&UI- la -Y Lt&WL JO -J ttt t L ".dU 1. "I¼.tl1111 4 .Lt Study, and World Development Report 2000-01: Attacking Poverty. Satu Kahkonen is a Senior Economist at the World Bank. Prior to joining 4the W-A A rld Bar, sh wa ssociate Elirector of the Center for Inst;.^-t-onal LILt V tt].Itttin] het O .waS Aij n uIa L~ICL UL U LILC ..CLC jul AILI L ULIiUl al Reform and the Informal Sector (IRIS) at the University of Maryland. In addi.o;on to conductlg eerho nttuinlapcso eC-30ornoic ,LttAtILIUI LOu t.J LA. L.LJ.LrLb IC,Ca%_tO t.l i tU ILL LI LLA ULIVU Il [ cli Ct ~ L~U L1FulI development, she has undertaken fieldwork in a number of developing aLn taLaniLuILn econoinuies. SLh is ULsh aLULILUI thLo a c-eiUto1 Uof A Nt-so-,U:s- mal Science: A Broader View of Ecotnomies and Societies, A New Institutional (pIrUflI LU LtLo E UfL UVLUVtlopr], et, andU Instit-u,io,,s, IretUi-tues urtu IZLUrTUfTiLL Reforms in India. Anirudh Krishna teaches at Duke University in the public policy and political scilence dUepar.lterLts. He hlodus a PhI.D. Hi guveriunleil stulies from Cornell University and master's degrees in international develop- mentSH allU CCUfl,)llJ 1101 lUli t_tlnell unilversity allU Inle LJetiu ZCiIU01 01 Economics. Before taking up an academic career, he served for a number of years in the Indian Admluinistrative Service. Enrique Pantoja is an urban planner and sociologist. He has worked in several international organizations on development policy and programs in Latin America and Souhn Asia, witn special focus on urban develop- ment, community-driven development, and social risk management. Mr. Pantoja has carried out research related to social capital, identity politics, collective action, and microfinance. He is currently working at the World Bank as Country Officer for Bangladesh and Bhutan. xi xii UNDERSTANDiNG SOCIAL CAPITAL Catherine Reid is a consultant with experience in participation method- oninay, tradring, -a ngricture, eiihr ro1nenbt, and -0,men'sA developen- issues. She has worked on a wide range of projects for the World Bank, theo UTTS Agoenc for Inern De-ve-I lotTJmeTnto (USAID), -nA the .. - "1 1 lb1- i *..ti Vt.SJ.lt-lt \tJJ k-JL' p Lii itt L International Center for Research on Women, and has conducted benefi- c.i-y assessments and PnrojeCt evaluatio-Sn in 4-he a,-144, eA-uain, and- public infrastructure sectors throughout Africa. She was the team leader of anLI iLovatLve pilotagiculLtura-l extenion project fo - pll f-- tllLCt, Lin the Democratic Republic of Congo. Ms. Reid has an M.S. in development rlngmttl ntL-o te h- 0lliL Lt1. ill TVV V lLL 4aL4lial - J-- V-t-C iA 1lClIt University. Lawrence E Salmen is Lead Social Scientist in the Social Development LiCyaLillilVLLUt fLiC t VhIe B-tLICSBA.. IVMI. SalIit llan h2lt IVIaJPX diiU cl I 1P.1J. In urban planning from Columbia University. He is the architect of two par- ticipatory researchL approaches used widely at the Bank, bene.Iciary assessment and participatory poverty assessment, designed to bring the voices of stakeholders in.to projec LIIanagerr.ent daIU poliLy lornation, respectively. He is the author of Listen to the People and numerous articles. I rIoL LU jUiliLrg uth Baik, Mvii. Salllein -wdas VICUe PreslUdenl 01 the Cooperative Housing Foundation and Director of Research and Evaluat-iUont uLhe UItLeI-tllricalLi ro-uiuatiouin. Elizabeth Shrader is an iternationai aeveiopment consultant speciailz- ing in gender violence, civil society strengthening, and reproductive health. She holds a master's degree in public healtn from the University of California at Los Angeles. Ms. Shrader has lived and worked through- out tne Americas. Tnierry van Basteiaer is Director of the Integrated Financial Services Team at the IRIS Center at the University of Maryland, and Study Coordinator for the Social Capital initiative. After receiving his Ph.D. in economics from the University of Maryland, he worked on private sector policy reform, legal and regulatory reform for microfinance, policy prior- ities for electronic commerce, small enterprise development, and corrup- tion. His published research has focused on the political economy of com- modity pricing, social capital, and microfinance. Social Capital: From Definition to Measurement Christiaan Grootaert and Thierry van Bastelaer Particular villages on the Indonesian island of Java build and maintain complex water deliveny systems that require collaboration and coordination, while other villages rely on simple individual wells. Residents in apparently similar Tanzanian villages enjoy very different levels of income due to differences in their abilities to engage in collective action. Households in Russia rely on informal networks to gain access to health services, housing, education, and income security. Some neighborhoods of Dhaka organize for local trash collection, while others allow garbage to accumulate on the streets. Hutu militias rely on fast networks of information and high levels of mutual trust to carry out a terrifyingly efficient genocide in Rwanda. Despite their geographical and sectoral diversity, these five examples have something in common: they all testify to the ability of social struc- tures and underlying attitudes to increase the efficiency of collective action.! In one form or another, these examples demonstrate the critical role of social interaction, trust, and reciprocity, as elements of social cap- ital, in producing collective outcomes, both beneficial and harmful. Readers of this book most likely can think of their own personal exam- ples of the impact of social capital on the development activities they engage in or observe. Development practitioners who have witnessed social capital forces at work in their field programs are looking for com- mon lessons from different areas and sectors. Academicians are searching for innovative tools with which to measure and analyze this new socio- economic element. And governments and donor organizations are hop- ing to find ways to identify the social settings in which their scarce funds will be used most productively. Social capital is assuming an increasingly important role in the World Bank's poverty reduction strategy. The World Development Report 2 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 2000/2001 identifies three pillars to that strategy: promoting opportunity, facilitating empowerment, and enhancing security. Building social capi- tal is at the core of the empowerment agenda, together with promoting pro-poor institutional reform and removing social barriers. However, social capital is also a critical asset for creating opportunities that enhance well-being and for achieving greater security and reduced vulnerability (World Bank 2001). This book, which is based on the results of the Social Capital Initiative at the World Bank, aims to provide concepts, measurement tools, and lit- erature reviews in a form readily usable by public and private actors interested in the nature and impact of social capital.2 The book discusses the respective value of quantitative and qualitative approaches to the analysis of social capital and provides specific examples of research based on these two approaches. It presents new field-tested tools designed to measure social capital in urban and rural settings, along with the ques- tionnaires and interview guides to apply these instruments. Finally, it reviews how the concept of social capital has been used in important areas of economic development. The Search for a Definition of Social Capital Research on social capital is relatively recent. Although the concept in its current form can be traced to the first half of the 20th century, and early applications to the 19th century, only in the last 20 years has it captured the attention of practitioners and academicians from different back- grounds.3 Not surprisingly, the lack of an agreed-upon and established definition of social capital, combined with its multidisciplinary appeal, has led to the spontaneous growth of different interpretations of the concept. The resulting definitions, which fortunately are more often complementary than contradictory, have been used in a growing number of research pro- jects and field activities to try to capture the essence and development potential of the concept. It is perhaps a testimony to the seriousness of these activities that the lack of agreement on a precise definition of social capital has not inhibited empirical and applied work. By clearly delineat- ing the concept they are using and developing methodologies adapted to it, most researchers have shown that solid and replicable results regard- ing the impact of social capital on development can be produced without a prerequisite fieldwide agreement on a specific definition. This lack of agreement, and the reluctance to impose a narrow defini- tion on a still-evolving conceptual debate. has led us to define social cap- ital broadly as the institutions, relationships, attitudes, and values that govern interactions among people and contribute to economic and social develonment.4 SOCIAL CAPITAL: FROM DEFINITION TO MEASUREMENT 3 The Forms and Scope of Social Capital The broad general definition implicitly distinguishes two elements, or forms, of social capital. The first, which Uphoff (2000) called "structural social capital," refers to relatively objective and externally observable social structures, such as networks, associations, and institutions, and the rules and procedures they embody. Athletic and musical groups, water user committees, and neighborhood associations are all examples of this form of social capital. The second form, known as "cognitive social capi- tal," comprises more subjective and intangible elements such as general- ly accepted attitudes and norms of behavior, shared values, reciprocity, and trust. Although these two forms of social capital are mutually reinforcing, one can exist without the other. Government-mandated orgyanizations represent structural social capital in which the cognitive element is not neressarily nresent. SimilarlyI many relations of mutual trust persist without being formalized in organizations. This description of social cap- ital according to its forms has provpn onuitp useful as a basis for emnirical analysis.5 A spcond distincrion that allows researchers to isolate the elements of social capital is based on its' scope, or the breadth of its unit of observa- .ii-n qSocial ranita! ran he observed at tfe fimiro level in the form of hor- izontal networks of individuals and households and the associated normc and vnilues t-hat underlie these networks. The choral groins high- lighted by Robert Putnam in his 1993 study of civic associations in Italy are rapidly becomnming a favorite examnple of t.his uinit of observationn in large part due to the impact of Putnam's study. Thoe eso level of observatiron, wAThich captures horizontal and vertircal relations among groups (in other words, at a level situated between indi- viduals ar.d society asa.h le) has ben 1,icusrnateA byw regionnal roup-1 ings of local associations, such as the Andean poor people's organizations described by Bebbington aand Carrol (2000). Finally, in the broader use of the concept, social capital can be observed aUt thle .acro level, in the foJ..J o the in-. ;tu.L;onati ar. poli."-nal er-t -i.inrtn rent that serves as a backdrop for all economic and social activity, and the qual- ity of 4LI. govem.ance arrangem.Lentt. TJI'L- -.-r.iir.ts, VViLV tet. ftous of inquiries by Olson (1982) and North (1990) into the sources of develop- ,merntar.d growvtLl-, put thie conceptLof soc~ialcapit-al wi&uLln sL.e real ..of nst-, II~I UIU6U VU,yS i I.ULty Li] UI.CI .c1ILaI VILIUI.L uIte, i l ti _,si - tutional economics, which posits that the quality of incentives and institu- iions (sucLIL db LIM LU rl o laW, UIe JLIULIdIcU bybLlIL, VI URt. qUaHLY VI LVItLICILL enforcement) is a major determinant of economic growth. Inc U1scUSSIUl 01 Socldl cdpiLdi dacrUIng, IU llb itUIIIb dIIlU bLU}U) Is illustrated in figure 1.1, where specific concepts of structural and cogni- T-TfCnC-A KIfT~T'.I A I'Tfl NNE ACT TDXC CllA ADT 4 UNER11NDN AND -ESRING SOCIAL CAPITAL tive social capital are presented along a continuum from the micro to the macrn dimensions= The breadth of the cnncept of social canital; as ilius- trated by figure 1.1, has been seen as a sign of both strength and weak- ness of the concepnhIal debate so far The strength of such a broad concept is that the inclusion of micro, meso, and macrn levels of social canitnal allows for imnortant effects of complementarity and substitution between the three levels. These vari- ous levels of social capnita can con-nPlemen- eahrl other, as A^hen national institutions provide an enabling environment in which local associations can. develon. HoTw-ever, local forr^s of social canital can developnn a result of both "good" and "bad" government. Tendler's 1997 study of decen- thralioa4in in Brazil, fonr eavnn.ple sho.virs l-ow 1.f1 inchnn,n govern- ment at the central level affects the success of local programs. Similarly, C pml /1Q95 csho,wc that o^ver timne in a T-Tnl-+A Ci-a+-c 1ir-1 r-rr'-nc .AitIfrAtI \ a .1/ 01 LU V OV tLI; flAt I ss[ LII I t. L L L . 111lICi IOLILi.0 I'.JCLAI have benefited from the strength of state institutions. Conversely, infor- mal. ne1.,-orks can be createndrespondftoaAurcf,inn4Analrstate, anhp pened in Russia (Rose 1998). In turn, micro- and meso-level social capital car, contriu-te to -ither in--ove or weal-eneA uc^on ftesae XV IJLILULC LU CALL-CI.. AAALf-T CI M.31V LI.AC LICA.A14 '3 It LL. On the positive side, local associations can sustain regional and national 11 LA LUL LIL1 C.RA I L LCAI AC LA L y a_ t it aLa -Ly. A I C A AC LI Ve side, excessive ethnic identification in local associations can impede suc- LuC poUlCie Uo CeVCeL ICCU LUa VtUCAL-C e LC(at 1777, %UICoLLia anditi ACL 2000). Figure 1.1 The Forms and Scope of Social Capital Macro Institutions of the state, Governance rule of law Structural | Cognitive reo Local institutions, ITrust, iocai norms, networks values Micro SOCIAL CAPITAL: FROM DEFINITTION TO MEASUREMENT 5 A certain level of substitution is also present among the levels of social capital. For examole, communities in develoDingy countries often rely on social pressure and reputation to enforce agreements between individuals or groups. When institutional development strengthens the rule of law and the court system, local informal arrangements for dis- nute resolution become less relevant and may lead to the weakening of the social ties that supported them. An exclusively micro-level focus on this transformation may exaggerate its imnact on social canital by ignoring the positive effects of a stronger rule of law on society as a whnole As expYnrPscld hv Stigtlitz (20nnn n pi;.) "[ars the mndern canitalist ---- -_ --- - - - r - .- --/ -------- -r------ state matures, representative forms of governance with a clear hierar- chilal striictiire and a system of laws rules, and regculations enforced by traditions replace the 'community' as the guardian of social, business, and nprsonal rcontracts " Hence a cnncept of social canital tihat encom- passes the micro, meso, and macro dimensions will be better able to rcatiiro the counit-erbalanrinog efects of s-triirciitres and attituiides at all levels of society. The potential weakness of using such a comprehensive concept of social capital in research and development programs is that it may be too bro-4ad too dra.v specific conclusions habout t1he role of attituideis, beh,aviors, or structures. In other words, by trying to be a catch-all concept, social capital may end up capturing nothing. Our experience . 4--h the Social Capital Initiative suggests, however, that when cautious researchers develop mrethodologies and indicators that match the spcfcconcep~t that they choose, they can produce solid and verifiable results. Although the emlpirical. stuidiess of the Social-I Ca-pital T dltative examline different forms and levels of social capital, covering all quadrants of figure 1.1, or qualitative-has helped ensure relevant and analytically focused lTte defir.iL.o CandU classificaf.ion of scS0ia-1l-C1F-1 .aitl gvten abouvte paralle5ls other conceptual attempts from a variety of disciplines of social sciences. ror example, as syndthesized by -vVoUlucock adIU Narayd1t (0uu), hte soLIU- logical literature on social capital has produced a different but compati- Vie categorization oti social capital. -l Il His1 - apprua n- gories, which largely overlap those in figure 1.1. The communitarian perspective describes sociai capitai in rerms of local organizations and groups, such as those that make up the micro/structural quadrant of figure 1.1, with a focus on productive social capital. Although this approach has helped focus the attention of 6 UNDERSTANDING AND MEASURING SOCIAL CAPITAL development practitioners on the role of social relations in the fight ~again, per'arh- if i9 nc,f Aot desOA 4. n tocap,re the nf social capital (such as sectarian interest groups, crime syndicates, or eth- nic -obs). It also ignores the fact that a large nu-ber of local organiza- tions, in itself, can be of little use for development if the groups lack external links and finflulence or if th1esLe structlures are not con-.plern.ented with strong cognitive elements, such as common values, norms, and n-utal LIrustL. The networks view encompasses relationships between and within hor- izonta erticA -s as such i ajects elemen ILU-iCLi aCLLk VC.L LILU GotLaif D LIL- a DCCLII -IJLLLO a ILC3L iI L LIL i this largely structural category. This view has led to the "bridging/ bond- II LAMIL IALLM;Y, VV I LIsL LL1Z5L Lu15 IDisI LtCi%tcllULU iLy Urom L-L dLacorinrLnu- nity ties.8 This perspective also takes into account productive conse- quences>v a:s well a:> udet_lll-:ltu -Ul:IUCitU: U1T_llLllLl 1Hl LjUCILLC c1~ WCi Li 11UCIIICLLCLI LUI t4L`tCI.jU-IILD Ut DJLIal L1FLaI.JiLL CLII .aLLU- lar, it draws attention to the benefits that social capital can provide to membuers of' a conUiILLiniity, as we'll as Uto the Uosts oI IUIILiIVIII1iLLJbehav- ior. Similarly, this approach recognizes that while social capital can unite LItC IILCIItL)tI of a cLUitiurLitLy itt 1i uofte dos stL the Lost ut ofexLudir.g nonmembers, as is, for example, the case with sects. -.q A. A .. . 1 _1 -- _ ilite inbuiuiuwriut VieW ptJSlb UldL Ull Politiald, legal, adlU llbtltUtluundl environments are the main determinants of the strength of community networks. This approach reflects tne ifiluence oi worKs by Olson (1982) and North (1990), in that it suggests that the ability of social groups to act in their collective interest is affected by tne quality of overarching formal institutions. Reliable, transparent, and predictable operations of the state, low levels of corruption, an independent judiciary, and strongly enforced property rights are but a few of the attributes of societies in which indi- vidual and collective initiatives can be successful. As such, this approach closely echoes the macro scope of social capital illustrated in the top half of figure 1.1. Finally, the synergy view aims to integrate the micro, meso, and macro concerns of the networks and institutional approaches, in effect cover-mg all four quadrants of figure 1.1. Based on the assumption that none of the development actors (states, businesses, and communities) have access on their own to the resources necessary to create sustainable and equitable growth, this approach focuses on the relationships between and within governments and civil society.9 The similarities and overlaps between the two sets of definitions in the above discussion underscore the unique multidisciplinary aspects of social capital research. Although the relative novelty of the concept man- ifests itself through a healthy proliferation of definitions, most of them fall within the boundaries described above. 0 SOCIAL CAPITAL: FROM DEFINITION TO MEASUREMENT 7 Rigorous Economic Terminology or Value as an Because noneconomists are usually credited with coining the term social capital, it should probably not come as a surprise that many economists feel that the use of the word capital to describe the concepts discussed here is not consistent with the use of the term in traditional economic the- ory. Social capital shares a number of features with the more traditional forms of capital. For example, social capital, like physical capital, accu- mulates as a stock that produces a stream of benefits (in the form of infor- mation sharing, and collective decisionmaking and action, as described below). Like physical capital, this stock requires an initial investment and regular maintenance, in the form of repeated social interaction or trust- building behavior. Social capital can take years to build and is more eas- ily destroyed than built or rebuilt. Social capital also exhibits several features that set it apart from phys- ical and human capital. First, and by definition, social capital, unlike human capital, cannot be built individually. Second, unlike physical cap- ital (but like human capital), the stock of social capital does not decrease-and can actually increase-as a result of its use. As Ostrom (2000, p. 179) points out, "social capital does not wear out with use but rather with disuse." On the basis of these and other similar concerns, respected economists have expressed serious misgivings about the use of the term capital to describe the concepts discussed in these pages, although they do not question their importance for development. Arrow (2000, p. 4), for exam- ple, has indeed urged the abandonment of the concept of social capital, largely on the grounds that it does not meet the definition of capital used by economists, in particular the aspect of "deliberate sacrifice in the pre- sent for future benefit." For his part, Solow (2000, p. 7) suggests that the difficulties in the measurement of this capital as "a cumulation of past flows of investment, with past flows of depreciation netted out" funda- mentally invalidates the underlying concept. As economists, we are sym- pathetic to the concern for terminological rigor that undergirds these arguments. We are also aware, however, of the unique opportunities for interdisciplinary research and development activities that are opened by the use of a common-if maybe incompletely focused-term by sociolo- gists, economists, anthropologists, practitioners, and others. As Woolcock (2001, p. 15) puts it, "one of the primary benefits of the idea of social cap- ital is that it is allowing scholars, policy makers and practitioners from different disciplines to enjoy an unprecedented level of cooperation and 8 UNDERSTANDING AND MEASURING SOCIAL CAPITAL dialogue." In short, the unique horizons opened by this interdisciplinary collaboration may well outweigh, for now, the benefits of a more rigorous economic definition. How Does Social Capital Affect Economic Development? Since it first entered conceptual and empirical debates, social capital has captured the imagination of development researchers and practitioners as a particularly promising tool for alleviating poverty. Indeed, because social capital is often the only asset that the poor have access to, devel- opment programs that build on it have rapidly become a priority for donors and governments. Social capital affects economic development mainly by facilitating transactions among individuals, households, and groups in developing countries. This effect can take three forms. First, participation by individ- uals in social networks increases the availability of information and low- ers its cost. This information, especially if it relates to such matters as crop prices, location of new markets, sources of credit, or treatments for live- stock disease, can play a critical role in increasing the returns from agri- culture and trading. For example, research under the Social Capital Initiative shows that better-connected Malagasy traders have access to more accurate information on prices and credibility of clients, resulting in higher profit margins (Fafchamps and Minten 1999). Lack of access to credit by the poor in developing countries is largely a consequence of lim- ited information about potential borrowers' credit risk, loan use, and truthfulness in reporting repayment ability. Several informal arrange- ments, drawing on social capital between participants and within their communities, have been developed to address these information imper- fections. These arrangements, which include rotating credit and savings associations, money lenders, trade credit, and microfinance, are described in annex 2. Second, participation in local networks and attitudes of mutual trust make it easier for any group to reach collective decisions and imnlement collective action. Since property rights are often imperfectly developed and applied in developing coiintries, collective derisions on how to man- age common resources are critical to maximizing their use and yield. Krishna and Uphoff (1999) describe how.^ farmer groups in the Indian state of Rajasthan use local structural and cognitive social capital to build consensus on the use of watershed land, result4ng in more productieve use of these lands, as well as improved outcomes for broader development {,acManagj-l rnll-or ve ndtior is alen onF-l f- 2D¢curilng -rssf-DC+ water and sanitation services, such as irrigation, drinking water, and u.-ban2f A2f w astenisosa! Arr.ex, 34 discusses the liera.r t-hat desc,ril,bes how .lSOCI.AL CA PI.lT A L.: FR OM.. D EVFI IN i TI LOLN T O M EAS'a UR E M EN T 9 social capital can help in the provision and maintenance of these essential services. Finally, networks and attitudes reduce opportunistic behavior by com- minimu mem.hcrc rn cpHsingsz wAT]erea c re-rtain 1-eaiobrr s avepcteaA fron individuals for the benefit of the group, social pressures and fear of exclu- cion rcn inducre thsoc indlivid1a to Arr,xiAl1a l vr,or+aA o l. example, farmers have resorted to these networks and attitudes and exerted mu';ual pressures to prevent individ-al diversion o4 irriglaion water. This issue is also discussed at length in annex 3. Approaches to Measuring Social Capital The main goal of this volume is to discuss the merits of various approach- es to mLLeasuring aCld -L Layt -Ib -he na -e and aiL_tj-_ of -o-I capital atI the micro and meso levels and to provide field-tested measurement tools to Ao -o Growin ---liria evid14ence iniae that- -ocia caia s b-est CLI tJ DU. 1.iii.VVV"L6 -ILLyIIiCLCI vIUtu ml%tt.L~LC3 ULaL D`L`al %- LCyLLa I~ LI L measured using a variety of qualitative and quantitative instruments. TnAeed Ah conep A_,.tb o.peeddstitywti h cnr. lI LL %Al. L L UIC%VLL~C;F t-LIULUL VIC %LUul1tVCiCLIUCkA 3LIILLIY VVILILilt Lilt CL_U1LL11Il ic paradigm, using quantitative methods. Neither can it be investigated solely tlihug6h anthropological or sociologial case s'udie. A corr.bina tion of both methods can be successful at isolating the role of social capi- tal ILLn incrIati lncor.tLes and fadilli.a.'itr dLaccb Lu bseViLes.-- The Social Capital Assessment Tool (SOCAT), described and made d adlecIUI Ut UIT LIULM, lILLCrIdL%fb LjUdIILILdLIVCt dUULIU 4UIILLIV IlILSLUIlLtelLLb. availalir Libokiner.sqatt.v an q-ualtative isrrlns SOCAT's community and organizational profiles each contain a quantita- tive aad a qualitaive compolient, while the household survey is a quanti- tative instrument. For collecting the data, structured questionnaires are I .- .1_I J . ..1 1 AI. ., useu along withl open-enueu participatory mernous. As a result, trne SOCAT data can support a wide range of quantitative and qualitative analyses. `Tis book provides examples or each type or analysis. Ideally the measurement of social capital should capture all four quad- rants of figure i.i, but in practice most experience has been gained with measurement at the micro and meso levels. The SOCAT is geared toward measuring social capital at these levels. Three units of data collection and analysis are considered: the household, the community, and the organiza- tion. i-nus, the SOCAT instruments consist of a househoid survey, a com- munity questionnaire and interview guide, and an organizational interview guide and scoresheet. Each of these instruments attempts to capture the rel- evant structural and cognitive dimensions of social capital. The importance of obtaining information on both these dimensions has been well estab- lished in the empirical literature and is discussed at length in chapter 2. In terms of the conceptual framework represented in figure 1.1, the Social Capital Assessment Tool covers the entire lower half of the diagram. 10 UNDERSTANDING AND MEASURING SOCIAL CAPITAL The data derived from the SOCAT can be analyzed on their own, when L.he ULJeLtVC 1E LU 11istoitLULY eAxisL,I. soLAcI Lcapital, Lt liLaF Litt= U1dLLULJL1U1t of social capital across areas or socioeconomic groups, or to gain a better understandIr.g 0I Uth fUn'iotLLnIIiLIng Ut key org1aizLdL11ti. Moe oUlLtfeit liart not, the objective will also be to relate social capital to outcome variables suct as houselo10Uld VI OcoUiltluitty welfare, access 10 services, or general development indicators. In such cases, the SOCAT questionnaires need to be combinled -wiult oner survey modules capturing income or expendi- tures and use of services. Indeed, most of the analytic examples in this 1 . , . 1. .~ . . . - . .1 DooK COlIme rromn studies tnat combinea a social caplial measurement instrument with other surveys. Most countries have experience with sur- veys that capture income and expenditures data, labor market participa- tion, or integrated socioeconomic information. The World Bank has sev- eral prototype surveys avaiiabie, such as the Living Standards Measurement Surveys (Grosh and Glewwe 2000) and the Social Dimensions of Adjustment integrated Surveys (Delaine and others 1991 ), which have been used extensively in many countries. The Local Level Institutions Study, from which several examples in chapter 3 are drawn, combined a social capital instrument with a household expenditure mod- ule (World Bank 1998). The SOCAT is a flexible instrument that can be implemented at the national level and at the project level. At the national level, an application in conjunction with a Living Standards Measurement Survey or house- hold income or expenditure survey would yield a data base suitable for studying the relation between social capital and poverty or for analyzing the impact of policy reform on social capital. At the project level, an early application of the SOCAT can provide baseline information on the differ- ent dimensions of social capital while the project is still in the develop- ment stage. This is valuable both for projects that aim to stimulate social capital directly and for sectoral interventions (agriculture, education, water and sanitation, infrastructure, and the like) that need to draw on local social capital resources for successful implementation. Repeat appli- cation throughout the project cycle can be used for monitoring and eval- uating the effect of the project on social capital. Structure of the Book In chapter 2, Anirudh Krishna and Elizabeth Shrader explore the relative advantages of different approaches to measuring social capital, argue the importance of locally and contextually relevant measurement tools, and discuss how to integrate qualitative and quantitative methodologies in developing these tools. This is followed by a detailed description of the Social Capital Assessment Tool. The community profile integrates partic- SOCIAL CAPITAL: FROM DEFINITION TO MEASUREMENT 11 ipatory qualitative methods with a structured community survey to iden- Lily fea,ures associated wi.i social capital wilfnl a specific communiry. The household survey explores participation in local organizations and attituues of trust and cooperation among individuals and nousenolds. Finally, the organizational profile examines the relationships and net- '-.1 . - . { . . . worKs tnar exlst among rormal and mrormal, organizations in tne com- munity. Together, these three instruments constitute a tool for multisec- toral and multidisciplinary application in diverse communities and coun- tries. The tool was developed on the basis of a large set of existing research instruments and has been successfully pilot-tested in Panama and India. In chapter 3, Christiaan Grootaert discusses basic quantitative meth- ods for analyzing the data collected by the SOCAT and similar measure- ment instruments. Three types of proxy indicators for social capital are recommended: membership in local associations and networks (structur- al social capital), indicators of trust and adherence to norms (cognitive social capital), and indicators of collective action (an output measure). Each of these indicators can be cross-tabulated with relevant socioeco- nomic data to provide a picture of the distribution of social capital. Various methods are discussed to aggregate the data through suitable rescaling and factor analysis. Multivariate models can be estimated to assess the contribution made by social capital to household welfare (including its monetary and nonmonetary dimensions) and the manage- ment of risk by households through better access to credit. The link between social capital and poverty receives special attention, especially the question of whether the investment that poor households make in social capital provides an adequate return and helps them escape from poverty. Finally, the chapter draws attention to critical methodological issues, such as the possible endogeneity of social capital in multivariate analysis, and suggests various solutions. Shifting focus to the qualitative measurement of social capital, the sec- ond part of the book suggests that this approach to measurement is best implemented through case studies. Two specific examples are presented: an agricultural sector case study from Africa, and an industrial sector case study from Asia. The case study presented in chapter 4 by Catherine Reid and Lawrence Salmen posits that trust is a key determinant of the success of agricultur- al extension in Mali. The study identifies three equally important aspects of trust: trust among farmers, trust between farmers and extension work- ers, and the relationship between extension workers and their national organizations. Using the beneficiary assessment methodology, which capitalizes on the establishment of trust for deriving insights, the study documents the importance of preexisting social cohesion. In that context, 12 UNDERSTANDING AND MEASURING SOCIAL CAPITAL women and their associations are found to be consistent diffusers of information and technology, able to tap into and generate social capital. The most important practical finding of the case study is that extension workers and development agencies in general need to gain an opera- tionally relevant understanding of the social and institutional fabric in places where they work. In chapter 5, Enrique Pantoja describes a qualitative case study that explores the role of social capital in the rehabilitation of coal mining areas in the Indian state of Orissa. Through the use of interviews and focus groups, the author finds that different forms of social capital (family and kinship, intracommunity, and intercommunity) interact to produce a mixture of positive and negative results for the rehabilitation process. The same strong ties that help members of a group work together can also be used to exclude other community members from the benefits of col- lective action. Although mutual trust exists in abundance around the mining sites, it is highly fragmented by gender, caste, and class, and results in closed groups with high entry costs and a considerable lack of horizontal links. The practical conclusion of the chapter is that lack of social cohesion at the village level can be a major impediment to com- munity involvement in the rehabilitation process and to community- based development in general. The full text of the SOCAT-the community profile interview guide and questionnaire, the household survey questionnaire, and the organiza- tional profile interview guides and scoresheet-is reproduced in annex 1 as well as on the CD-ROM enclosed with this book. As mentioned above, annexes 2 and 3 discuss how social capital facilitates the delivery of ser- vices such as credit, and water and sanitation. Notes 1. These examples are excerpted from the following papers, respectively: Isham and Kahk6nen (1999); Narayan and Pritchett (1999); Rose (1999); Pargal, Huq, and Gilligan (1999); and Colletta and Cullen (2000). 2. The Social Capital Initiative, a large-scale research project funded by the government of Denmark, produced an extensive body of original data on social capital. It drew on these data to provide strong empirical evidence of the perva- sive role of social capital in a wide variety of development processes and projects. For a description of the Initiative and its main findings, see Grootaert and van Bastelaer (2001). 3. Woolcock (1998) identifies Lyda J. Hanifan, superintendent of schools in West Virvinia in 1916. as the first oromoter of the concent of social canital. a concept that was resurrected in the 1950s and 1960s. The concept gained new attention in the early 1990s (Ptitnam 1993). Woolcock also depcrihes the conditions that led to the SOCITAL CAPITAL: FROA DEFINITION TO MtEASUREAMENTT "first deliberate attempt to establish financial institutions with the poor in devel- nnincg Pcnnnmips on thei hbsi of thpir qociia rather thin thpir matPria! rPeou1rces" (p. 95), referring to the "People's Banks," as the credit cooperatives that Frederick Raiffeisen created in mid-19th-ceni-rni Germqany were collectively called. 4. Grootaert and van Bastelaer (2001, p. 4). The last elements of this definition rcn be in+trpreted broadly, to nllolnw fnr the Pyic*tenr. of d aoaina or harmful social capital. 5. See Krishna anrd Uphoff (1999) for a parti.cul,,ly persuasive em.pirical use of this classification. 6. See Grootaert and van Bastelaer (2002) for a selecti:on of these stdies. 7. For an authoritative review and discussion of the sociological and develop- mrent fiterature on sodal capital, see l,V.oolcock (19,98). 8. Although the relevance of the bridging/bonding distinction is somewhat ;r.him-d by the r..alleable an --d. dtyn.i a-I of Li.e Loncept o_f L Ul ULU (at what point does successful bridging social capital turn into bonding social capi- tal?), it h-asben a u3sItnn basis for ItLe dns.3siio sectarian Uti-re-t groups -d4 exclusion mechanisms. ( T A- fL I - .. * A_/1(1c] ]\z__ ~L _ A. 7. jnani'i alitl. I t1Y2YJ jJICaCILL alL SnLLjLI-aI DlUL. U1 LL 11 lilFVl Lail tc 01 "coproduction" between local authorities and communities in designing and antaIIIL1irn, WatLI UdeiveIy sysintnls 0n the islanlU 01 Java. 10. A small number of definitions of social capital differ from the general approac,'t dUescribe heeU by- t oc__ n on - subst- of the ele.lt discussed Iabove. d~jJUdi.I UCL1~UU IttfI1 uy LULLUW11r, uit a Z~UL'bt:L 01 Llit IC LM~LCILM UIbLUZICU ilUUVC:. For example, Robison, Schmid, and Si les (forthcoming) define social capital as "a pro' or gru' syp_i - vLWalr _ ___L- L __ , yz1bU1I b 01 blUL1p b byiipdiily LOWOIU Ut )lUltll t:UlUII Ut rIUUF LlIdt lilay pIUUULL a potential benefit, advantage, and preferential treatment for another person or group of persons beyond that expected in an exchange relationship." Accordingly, their definition roughly corresponds to the micro/cognitive quad- ranrt of figure 1.1, witi.h an added unilateral feature. 11. See Grootaert and van Bastelaer (2002) for a collection of empirical studies that draw on oth quantrirative and qualirarive analytical metnods. Rxeferernces Arrow, Kenneth. 2000. "Observations on Social Capital." in Partna Dasgupta and Ismail Serageldin, eds., Social Capital: A Multifaceted - Tiny C' - - In lIf a Perspecrive. 'v'vasnington, D.C.: vvoria Dank. Bates, Robert. 1999. "Etnnicity, Capital Formation, and Conflict." Social Capital Initiative Working Paper 12. World Bank, Social Development Department, vvashington, D.C. Bebbington, Anthony J., and Thomas F. Carroll. 2000. "Induced Social Capital and Federations of the Rural Poor." Social Capital Initiative 14 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Working Paper 19. World Bank, Social Development Department, l 3ahn t T , X.. ,1 T C.olletta, NYa. J., and MiV.hIe'lle L. Cullen. 2000l. "Theteusb'.-e Violent Conflict, Social Capital, and Social Cohesion: Case Studies from Camnbodia and R-w-anda." Social Capi'tal Initiative WvAIorking Paper 23. World Bank, Social Development Department, Washington, I- I - T 1f -- f U f . 2 Dielail-ne, Gh.Aislaine, Lioln LJCDemer-y, Jedia-Luc D-ubis, DIdiKo GUJiL, Christiaan Grootaert, Christopher Hill, Timothy Marchant, Andrew McKay, Jeffrey Round, and nnristopher Scot. 199;. ine ;ocial Dimensions of Adjustment Integrated Survey: A Survey to Measure Poverty and Understand tne Effects of Policy Change on Households." Social Dimensions of Adjustment in Sub-Saharan Africa Working Paper 14. 'vvorld Bank, Africa Region, vvashington, D.C. Fafchamps, Marcel, and Bart Minten. 1999. "Social Capital ana tne rlrm: Evidence from Agricultural Trade." Social Capital Initiative Working Paper 17. W'orld Bank, Social Development Department, vvashington, D.C. Grootaert, Christiaan, and Thierry van Bastelaer. 2001. "Understanding and Measuring Social Capital: A Synthesis and Findings from the Social Capital Initiative." Social Capital Initiative Working Paper 24. World Bank, Social Development Department, Washington, D.C. eds. 2002. The Role of Social Capital in Development: An Empirical Assessment. Cambridge, U.K.: Cambridge University Press. Grosh, Margaret, and Paul Glewwe, eds. 2000. Designing Household Survey Questionnaires for Developing Countries-Lessons from 15 Years of the Living Standards Measurement Study. Washington, D.C.: World Bank. Isham, Jonathan, and Satu Kahkonen. 1999. "What Determines the Effectiveness of Community-Based Water Projects? Evidence from Central Java, Indonesia, on Demand Responsiveness, Service Rules, and Social Capital." Social Capital Initiative Working Paper 14. World Bank, Social Development Department, Washington, D.C. Krishna, Anirudh, and Norman Uphoff. 1999. "Mapping and Measuring Social Capital: A Conceptual and Empirical Study of Collective Action SOCIAL CAPITAL: FROM DEFINITION TO MEASUREMENT 15 for Conserving and Developing Watersheds in Rajasthan, India." Social Capital Initiative Working Paper 13. World Bank, Social Development Department, Washington, D.C. Narayan, Deepa, and Lant Pritchett. 1999. "Cents and Sociability: Household Income and Social Capital in Rural Tanzania." Economic Development and Cultural Change 47 (4): 871-97. North, Douglass C. 1990. Institutions, Institutional Change, and Economic Performance. New York: Cambridge University Press. Olson, Mancur. 1982. The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities. New Haven, Conn.: Yale University Press. Ostrom. Elinor. 2000. "Social Canital: A Fad or a Fundamental Concept?" In Partha Dasgupta and Ismail Serageldin, eds., Social Capital: A Multifarcted Perenertive. Washinixton. D.C.: World Bank. Pargal, Sheoi, Mainul HuqP and Daniel Gilligan. 1999. "Social Capital in Solid Waste Management: Evidence from Dhaka, Bangladesh." Social C'anita! Initiativp Working Paner 16. World Bank- Social Develonnmnt Department, Washington, D.C. Putnam, Robert D., with Robert Leonardi and Raffaela Nanetti. 1993. Manling, fpmnrnrri, l Wnrk: C'i7ic Traditinn¸s in Mndern Italu Princeton; N.I I Princeton University Press. Robison, Lindon, Allan Schmid, and Marcelo Siles. Forthcoming. "Is Soc;ia C pital Ra!yCpia? R07)i]1 of' Socinl Ecofnomy. 1P-.o Richar.A 1Q998 "(2t4+hncr Thingac FDno in an, Anfi-M Ac.rn C.neiphr Social Capital Networks in Russia." Social Capital Initiative Working Paper 6. WAktorld Bark Socia! i.Dertfmer .VinQco+,nr) C. 1999. "What Does Social Capital Add to -ndividual _AIelfare?" Social Capital Initiative Working Paper 15. World Bank, Social D e v e1o pm. te nt Depar-l-.ent, ltAlaShi-ngtor D.C. Skoc.pUl, 11TLhe1. 1995P. I IUrLL,g oUIeIrs) tilt iviuti D. I Mtes I UPotLLL ticalOIrii of Social Policy in the United States. Cambridge, Mass.: Harvard Ulies. T Press. ;6 UNDERSTANDINGxTr AND ME-ATURI- CSOI AT CA T AIT Solow, Robert. 2000. "Notes on Social Capital and Economic PPrfnrmanrce" In Partha Dasgnpta and Tsnil Serageldin; eid wAonrial Capital: A Multifaceted Perspective. Washington, D.C.: World Bank. Stiglitz, Joseph E. 2000. "Formal and Informal Institutions." In Partha Dsufnio nrl Tcm:zil qPr:;iolrjin Prti qncinl Cnnifnl A Mijltifnrptpd Vur -O~~~~--- -- ---- -r-- -_ 1_-^7,___ Perspective. Washington, D.C.: World Bank. Tendler, Judith. 1997. Good Government in the Tropics. Baltimore, Md.: oh-ns l-oi-c 1 TUni-e,rcit Proes Uphoff, Njorman. 200O. "Understanding mb cia Captal Y'*r-- Analysis and Experience of Participation." In Partha Dasgupta and Is,ail Serageldin, eds., Social Capital: A Multifaceted Perpecive Washington, D.C.: World Bank. Woolcock, Michael. 1998. "Social Capital and Economic Development: .-woar. Aa TI'.eore.LIcal Syn -hesis -an ATl.-y Ar-L.JIfl. Theoy -4rd Society 27 (2): 151-208. . 2001. "The Place of Social Capital in Understanding Social and Ecconorric ut-'.comres." Isuma I (-1)\: 11-I17. VvvIolcock, MiVch-atel, Cand Deep.a Naryan." 20 0.'"Scia CaFpital: Implications for Development Theory, Research and Policy." World Dan'k- R'esearchr- ObUserver 15 (2-): 2 T1 AS fl- 1 I Anf 1lO 1 T 1 Y * * l2 af l . . v^voril Bank. i99o. "The Local Leve! Institutons Stuuy: rrogram Description and Prototype Questionnaires." Local Level Institutions 'vvorking Paper 2. 'v'World Bank, Social Development Department, Washington, D.C. . 2001. World Development Report 2000/2001: Attacking Poverty. New fork: Oxford University Press. 2 The Social Capital Assessment Tool: Design and Implementation Anirudh Krishna and Elizabeth Shrader ,s Tne concept or social capital nas traveieu' oey onu its semnudi dpplica- tion in Italy and analysts have extended it to other countries and regions, new ana dirrerent measurement tools nave emerge'. These tools dirfer substantially from those originally developed by Putnam (1993). This raises a key question: Must the measurement of social capitai necessarily vary by national, regional, or ethnic setting? Can some common method of measurement be developed tnat can be applied uniformly across adi- ferent countries, regions, and cultures of the world? To some extent, the existence of different measurement concepts is jus- tified because empirical correlates of social capital vary from one context to another. Recognizing that context matters, however, does not imply the need to have countless measurement tools unrelated to each other by a unifying concept. The number of useful categories for examining sociai capital may well remain fairly constant, even as inquiry shifts from one context to another. It would be desirable, therefore, to have a toolkit for measuring social capital that subscribes to a commonly agreed definition of social capital but that is broad enough and flexible enough to be applied in diverse cultural contexts. As discussed in chapter 1, the broad definition of social capital used in this volume embraces micro (individual), meso (community), and macro (national) phenomena. Examples of social capital at work have been doc- umented at all three levels. The tool proposed in this chapter focuses mainly on social capital at the micro and meso levels. As defined in this book, the social capital of a society consists of the institutions, relationships, attitudes, and values that govern interactions among people and contribute to economic and social development. Two sets of empirical questions emerge from this definition: * What types of networks are most commonly associated with social trust and with norms that promote coordination and cooperation for 17 18 UNDERSTANDING AND MEASURING SOCIAL CAPITAL mutual benefit? Do all networks or only some types of networks need to be aggregated into a measure of social capital? Should rhe same ner- work types be considered in all settings, or should there be some vari- ation across culrurai contexts? * Can norms and trust be assessed directly? Can they be graded in terms or their effectiveness for facilitating mutuaily beneficiai coliective action? What instruments enable a researcher to identify and assess these measures across diverse cultural contexts? Empirical studies of sociai capitai differ in the manner in which they have addressed these issues. While some studies have assessed social capital solely in terms of network density, others have reiied purely on a measure of trust. Yet other studies combine a measure of network densi- ty witn some proxies tor evaluating tne strengtn ot the relevant norms. We argue that neither an exclusively networks-based nor an entirely norms-dependent measure suffices for scaling social capital. Sociologists generally agree that the shape of any network-horizon- tal or vertical, homogeneous or heterogeneous, formal or informal-does not by itself indicate much about the nature of human relationships with- in that network. What sorts of norms are related to which type of net- works cannot be assumed a priori but must be investigated indepen- dently for each context. Granovetter (1985, p. 487) offers an oft-quoted view: Actors do not behave or decide as atoms outside a social context, nor do they adhere slavishly to a script written for them by the particu- lar intersection of social categories that they happen to occupy... While social relations may indeed often be a necessary condition for trust and trustworthiness, they are not sufficient to guarantee these and may even provide occasion and means for malfeasance and con- flict on a scale larger than in their absence. Social capital in one context can be unsocial capital in another. Organized religion that supports peace in one context becomes a forum for armed militancy in another. Unions that may promote coordination and cooperation with the state in a corporatist context can wage bitter confrontation in another context. It is hardly surprising, therefore, that analysts studying social capital in different contexts have found different network forms to be associated with social capital formation. The hori- zontal or vertical nature of an organization may matter a lot in Italy but little in Panama or rural north India. Heterogeneous organization may be more valuable in some countries and less so in others. Context matters, because norms of behavior associated with different network tvDes are THE SOCIAL CAPITAL ASSESSMENT TOOL 19 not constant across cultures. Different types of networks are associated with cooperative norms in different cultural contexts. Minimum Requirements for an Acceptable Measurement Tool A tool for measuring social capital must recognize and be sensitive to cul- tural variation, but it must also provide a common conceptual framework that helps unify the different dimensions of social capital. Flexibility is essential, but it must be contained within what Peters and Waterman (1982) have referred to in a different context as a "loose-tight" frame- work: loose, or flexible, in the details but tight on the essential concepts. The types of organizations to be included in the measure of social capital can be determined through careful examination of the existing local structure of organizations, suggesting that organizational-level analysis should accompanv assessments of social canital among communities and households. (The organizational profile presented later in this chapter is heInful in this regard.) - r c- - Tncorrorating Different Dimensions of Social Canital Scrial capital was originally defined and measured in terms that related entirely to the density of horizontally organized social networks. I.;iihsenint investigations have incrrPeasd thp rnmnlpyitv of thp connrnt by adding a number of dimensions. STRUCTURAL VERSUS COGNMVE SOCIAL CAPITAL. The structural elements of sncial canital (relatincg tn nPhvonr1ol rnles riles and nreredents') must he assessed separately from cognitive elements (relating to norms, values, attitudes, ar.d beliefs). While cognitivp elements nredispose peop toward mutually beneficial collective action, structural elements of social capit1a-! facjiitate schr action (Krishna 2000, TTphoff 1999). Bothl structralr and cognitive dimensions matter; they must be combined to represent the aggre-ate -otenti;al for mn.utuall- beneficia 2oll0ecti-ve aitin ti-- exis within a community. As, HGter.i1 (1987) has tren4har,t1y argue,,A group, Cn1ida21i4 is dilfficult t-o verify with reference to norms alone. By the same token, the type of net- wolrk is notl a reli-able indicattor of the t^y-pe of Imar ;-4erp-44o --cc--ring within a group: network types that support cooperation and coordination it Uoin cnULLxAL IlLay LIVIXLULote cor.11LeLLttiLoU aLcit in another. Ethnographic examination by Eastis (1998) of two otherwise similar choir gi-oups conclfudtes 'U.tatt rr.ere nrI rILo Uo anLULICte Lt:ru1y of Vul- untary association is too crude a measure to capture empirically the com- 20 UNDERSTANDING AND MEASURING SOCIAL CAPITAL plex experience of membership. Members of both choir groups reported very extensive participation, yet came away from the experience with a rather different mix of human, cultural, or social capital. Such variation owes much to the characteristics of the groups and the structure of rela- tions between their members-not to participation per se nor to the types of groups per se. From his study of groups and associations in contempo- rary Russia, Rose (1998) concludes that trust is not associated with all types of networks, even those that are horizontally organized or that have a het- erogeneous group of members. In the Russian context, some (but not all) informal networks are associated with trust and trustworthy behavior. To understand the separate importance of norms and networks, con- sider the following examnle. Someone's house or barn burns down at night. Neighbors come together the next day to help the family rebuild the structure. This kind of collective action can be found not iust in devel- oping country villages, but also among diverse social groups in all parts of the world. What is interesting to examine are the factors that lead neo- ple to behave in this way. TwO alternative scenarios can he envisioned. It is nossible that there is well-recognized leadership within the neighborhood. Receiving informa- tion about the unfortunate eventf commiunity leaders direct villagers to collect at the site, bringing along whatever tools and building material they may possess. Alternatively, no clear roles for organizing such action exist in the community. Motivated instead by norms of what is appropri- ate behavior, people collect spontaneously to assist with the rebuilding.l Thus the same cooperative outcome can come about in two different .vays. Tn the first case, the structural dimensions of sorial rcnital (roles rules, networks) facilitate cooperation and coordinated action. In the sec- ond case, collective action is based on norms and beliefs-that is, ithas a cognitive, not a structural, basis. Considering only networks neglects mutually beneficial collective action that has a cognitive basis. Considering only norms underestimates social capital by ignoring its structural1 A..mensions. To b-e vaidA and accurate, a to0olC f-or tmfl-uuning social capital must account for both its structural and its cognitive dimen- siVors. Bo)Lth norrnms and ne.hvvork's mustL be assessed, and" LILC llteaaure of social capital must represent the aggregate potential for mutually benefi- Lial collecfi;ve ac'lon. Specific dimensions have also been added to the concept of structural social capital, m ti'te forImI ofl ute follUW-ir. UibLULctions among tLypes oUL organizations: HORIZONTAL VERSUS VERTICAL ORGANIZATIONS. In his seminal 1993 work, Putnam argues tnat horizontally organized networks contribute to social TH4E SOCCIA CAPITAl ASrPSSMEPNT TOOT capital formation, while vertical relationships inhibit it. Later analysts of social capital have challenged Putnam's preference for horizontal over vertical and hierarchical organizations. The Boy Scouts (Berman 1997, p. 567), for example, are a hierarchically oreanized group, vet thev seem to be favorably regarded by most social capital analysts. Empirical investi- gzations carried out in a varietv of countries indicate that horizontal net- works do not necessarily reveal the presence of greater social capital. Studying variations in economic growth in 29 countries, Knack and Keefer (1997) conclude that trust and civic cooperatioh are associated with stronger econonmic performance hbut that associational activity is unrelated to trust. HETEROGENEOUS VERSUS HOMOGENEOUS ORGANIZATIONS. That the internal heterogeneity of groups matters for both socinl ranita! and cnnnomir IATwel- fare has been independently verified by Narayan and Pritchett's 1999 Shid nf Ta3n7znian x7ill!agcs and nil 19QQQ slidy nf Tnrlnnod2n vi!- lages. In each case, a household-level index of social capital-combining thb niimbor nf nccn%rian.an namomhorclxnc tbh int-rn,lh bterogeneiy, of associations, and the span of activities-was found to be positively and consistentl, rea0teo ton houshold - e r lfare. Ar,nI-brngroup of scholars derives the opposite conclusions. Drawing on data gathered from five U.S. cifkes, Portn,ey and Berry (1097) conclude that- con-rnpA with social, service, self-help, and issue-based organizations, "participa- ,on in neighbor-hood associations is More strongly associated with a high sense of community" and civic engagement. All else equal, more homo- geneLous nighborhloods are- rr.ore. likely to.1V hv rr1.ore L Lff--v __ighIbor- hood associations. What is a researcher to make of these seemingly con- Cbr4r, ,,4.-EA,,;A e9 CAAI-ln IA 10a0aQ\ r.,1.An, i-k . A-U i-,.- nCss A- _. .,,A_ fliLtiLg viewsV St-LJL111. \1919.J) c£JI1L1.ACOd tLtI Lthe LypJe Of 5i'JAu IFLLIi3' more in some contexts than in others, suggesting that selecting an appro- priate ne.vvork mreasure for sociall capitLall depends on the-C LcourLtry VI cul&- ture being studied. FORMAL VERSUS INFORMAL ORGANIZATIONS. Several other issues are also relate-o4 le.gaprpit type -- f oraia.os Should --- ICcLC1. wU ZC:1C:LItf6 CtFyiuFiCatLc LYFUC2 13L uI6C11LlaLlulLb. JItVUIU Vitt: include only formally organized groups, as Putnam does? Or should ilnfoIIimiii oupsal Ute con1Li1Urt:U, etsetldly silltc, ds INe-wItUl (1992, p. 582) suggests, "the socialization role of creating 'habits of the heart' is m-ore likely to be played... by informal groups"? Should only small face- to-face groups be considered, or are large multiregional and multination- al organizations also instrumental in promoting coordination and coop- eration for mutual benefit (Minkoff 1997, Oliver and Marwell 1988)? Are strong associational ties better than weak ones, or vice versa? 22 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Observing the Appropriate Collective Activities Social capital exists "in the relations among persons" (Coleman 1988, pp. S10-S1l0); thus only act,v; 4-t r n a 4.,-" ; --, out collectively rather than individually should be considered for assessing social capital. WhaAL,t 4-p __tL4-i;- are v aries i It- -eo x to UIIt.i. Crop diseases, but not crop harvesting, are usually dealt with collec- LIVC:Xy LLL I Ulal oI thJi LIn dLtia. A HUose L.oIr0 sLI U Li is an ULAVIiLdal eILLteFise in Indian villages and a collective one in Somalia (Farah 1992). li g CLI6actoIr cUoirraring suoIciai caFiLai amoinLUgLL JULLa l%i cJIILItULLnities should therefore look at networks that build houses. In rural India, how- ever, thitey mList loo0k at sociall bVelhavior for LIual"r YIgiLI crop 1disease:. Past practice and future expectations of community action are likely to vay conslUielrably acIro1ssUo LUI1UILUILiLIU. MV11iIrbers o1 LU1lUIILUMiCZe ULat have acted collectively in the past expect that the community will contin- ue to cooperate regitflarly. 1ithbes VdaldLiailb Ut UtC LULbJeVdItLt 01 bildItaU collective norms and practices provide the locus for measuring social capidal WIltlhn any giVenl cUIUrdl space. A tool for measuring social capital must build primarily on those activ- ities that cournufuity members consider appropriate for collective execu- tion. What these activities are in any given region can be determined only mrougn on-The-ground investigation of community organization (the community profile presented later is helpful for this purpose). Integrating Quantitative and Qualitative Methodologies Increasingly, social science research, including economic research, inte- grates both quantitative and qualitativ.e methods.2 Integrating compie- mentary methodologies allows researchers to confirm and corroborate results, elaborate or develop analysis, provide richer detail, and initiate new lines of thinking by studying surprising results or paradoxes (Rossmian and Wilson 1985, 1994). integrating complementary data col- lection techniques is especially important when trying to analyze a com- plex and innovative concept such as social capital. A tool for measuring social capital that integrates both quantitative and qualitative methods is likely to be more useful and reliable than mea- sures based on only one type of research methodology. Coupled with results from validated survey data presented in the form of scaled items, qualitative indicators can provide a deeper understanding of what indi- viduals, households, and communities regard as social capital. Survey data generate a broad overview of the institutional framework in a par- ticular community. Institutional mapping, focus groups, and other quali- tative techniques provide a more nuanced understanding of institutional THE SOCIAL CAPITAL ASSESSMENT TOOL 23 characteristics. Applied together, qualitative and quantitative techniques can provide a more comrrplete and corn. ncn, ng-4 r -enr.n of 4-b- local ne4- tutional landscape. Designing the Social Capital Assessment Tool (SOCAT) The preceding discussion makes clear that a tool for measuring social ita l m s LLI. L IlilLi the II. f lVVVLi L5 ILLU Li L L- M.i.Leia. 1.It& must rcbizardb es.;etcuualvar-a.-on whifle atth 1.IL I IU % i_U61L1LLC LTILU UC LDUI,LtVU LV %LUILUICdI c1LI LiUi Wtii CL Li Lt: same time provide a unifying conceptual framework. 2.It.T rr.tust -ake acout of_ --_e -_u.ua as 1 -_l as 3.he con.v ien- sions of social capital. Both networks and norms must be assessed to obtain a valid estiImate of t*he aggregate potential for mIutually benefi- cial collective action. 3. It must build priniarily on those activities local people consider appro- priate for collective execution. 4. it should be constructed usimg Dorn qualtative and quantitative methods. These four criteria are not only theoretically valid, they are also help- ful in practice for resolving measurement issues. Whether to include only horizontal organizations or both vertical and horizontal organizations can be resolved in terms of the third criterion, for instance. How are mutually beneficial collective activities commonly undertaken in the con- text being studied? 'vWvhat types of organizations do people use to carry out undertakings valued by the community? Horizontal organizations may be the common carriers of collective action in some settings, but ver- tical ones may be more appropriate in others. Similarly, heterogeneous organizations may serve valued collective purposes in some cultural con- texts, but more homogeneous membership patterns may be associated with cooperative norms in other cultural settings. The design and field testing of the Social Capital Assessment Tool rep- resents a first step toward the development of a uniform measure of social capital. The tool-described in detail below-accounts for cogni- tive and structural social capital, integrates qualitative and quantitative measures developed concurrently and iteratively, and is valid and reli- able across a wide range of community, household, and institutional con- texts. It is applicable at all levels of project design. In the earliest exploratory phase, an abbreviated version of the SOCAT can be used to assess quickly levels of social capital as well as community needs and assets. In the preparation phase, the SOCAT can be used to collect base- line data on norms and networks to determine the extent and types of 24 UNDERSTANDING AND MEASURING SOCIAL CAPITAL social capital present in the project area. The flexibility and rapid appli- cation of the SOUAT allows for the collection of monitoring data through- out the project cycle, providing quantitative and qualitative supervision benchmarks. Project evaluation will benefit enormously from a full SOCAT conducted upon project completion, ideally comparing the results with data from control communities that did not participate in the project. This information can provide insight regarding project sustain- ability and the role of social capital in future development interventions. A key objective of the SOCAT is to contribute to the understanding of how community-, household-, and organization-level measures of social capital interact with other development indicators and thus to assess whether social capital contributes to or erodes economic and social devel- opment. Pertinent development indicators-including measures of poverty, education, health, infrastructure, crime and violence, public sec- tor reform, and labor-can be incorporated at all levels of project inter- vention, including initial preparation, implementation, and evaluation. Associating these development indicators with measures of social capital may begin to explain important, perhaps previously overlooked, aspects of successful and replicable community development. Structure and Implementation of the SOCAT The Social Capital Assessment Tool consists of three instruments: the com- munity profile, the household survey, and the organizational profile. These instruments were developed after studying a large set of research instru- ments applied in more than 25 studies conducted in 15 different countries. The SOCAT was pilot-tested among urban, rural, and indigenous popula- tions in Panama and India, and refined based on the results.3 Community Profile A SOCAT exercise should begin with the administration of the commu- nity profile, for several reasons. First, meeting with community members in groups enables the research team to familiarize itself quickly with important community characteristics, which should be known before the other instruments can be applied. Group interviews help establish boundaries for the "community" within which the research will take place. This understanding of community boundaries will be used throughout the exercise for selecting household members to be inter- viewed and for identifying organizations that will be studied with the help of the organizational profile. Second, social capital needs to be investigated with reference to activ- ities that are commonly undertaken collectively within the cultural set- THE sOCIAI. CAPITAT ASSESSMENT TOOL ting being studied. The nature of such activities varies from one culture to another. Group discussions help identify activities that are commonly executed collectively in the community in question. Third, and perhaps most important, are the intangible benefits that arise as the investigating team and members of the local community come to know each other better in the course of these onen discussion sessions. Misgivings are dispelled as community members become famil- iar with the nurposes and nronosed activit-ips of the research team Everyone is welcome to participate at these sessions, which are publi- cized widelv in advance. A dive,rsitv of viewc ic rpnre-c,ntedi anc th,^- research team gleans a preliminary sense of the various characteristics of interaction salient for assessing social capnial in the commnunity. The com- munity maps, diagrams, and fielcl notes generated from this exercise serve as primary sources of information on organizations and relation- ships within the community. Researchers hepin thfe cornmmnibt nrofilini exercise wAith anLnifi.al open-ended discussion with the community, based on a participatory interview giidep Thpe fkhan nAminhi-ar a cstand rdjizrA ed 1 str+re1 qs-c- tionnaire, which will help to compare responses across communities. This nh,et,cnnirainxTanfnrcrlc I-ncir ;nf,,,,~~~r~cf,,-o -,a - qusfon.ar invnto:e bai--r.mlnyi -atuture ar. iden-;fi-es available resources and development needs and priorities. OPEN-ENDED COMMUNrTy DISCUSSIONS. Before convening a community reetinf and admnnistering this nt-rulnt.-, the re.s.arr- teamn. should spend several hours to one full day walking through the community to gan a sense of he spe-ial characteristics that rright inluence filw interactions and logistical planning. Ideally, this community "transect n^.3alk" shoufldbe gLhit-id by a IICIy cor.rrr it lea -vho acts as a kin of gatekeeper of knowledge to facilitate the researchers' access to commu- ritLy-level iinforr.mation. tnifta.all contact's _n_ade A-ur_ U;it transect wvalal Lar serve to identify key informants, establish convenient times and venues 4Air comni ui ni ty It ILt-iii 1 n0,. at r ts tatttl la1[C 1_ [ a... _t L LM 1% UCL VVW CCI L LUll UtljUl IJtLY members and researchers. A convenient public space slh-ould' be selece stevnefrec P-x LJlVCLCL UUJ 1C1 IUIU LI CIcteLCU C1~ LIt%C VtC.ILUC 1LUl tcdtLI group meeting, which may be either spontaneous or planned. Planned m^eetings ------vtn e ifrlats(rhvigakyinoral tLC11 IIIVUIVC 111VILt11% I.t.Cy UL1UMIt1c11Lb IkU 1 tdVItI6, i iStfy IIUVLlIltdtL invite other key informants). These informants are particularly useful when thet- VI"-VWUItI ofd acetaiLdUn utfiltulrplUL rI`U-yutul, wUoliell, laborers-is of interest. In societies in which divisions based on gender, ethlu-icity, or some otner factor are likely to be important, it is often useful to conduct separate meetings with different subgroups. Two to eight group interviews can be conducted in eacn community, witn a larger number of interviews required in communities that appear to be more 26 UNDERSTANDING AND MEASURING SOCIAL CAPITAL deeply divided. Generally, however, at least two group interviews should be carried ollu vi'h TAomen and men separate Mixed groups crnn also be interviewed to assess levels of consensus, but these interviews should be in additi-on to the gender-based goup interriews. Spontaneous n er- views take place at any given moment in a public place. These interviews should bIe aiAAn aiti toil telanr.ie ---et;k1L6g, -Vi-, soulA Ile announced and publicized widely in advance and held at times conve- Inint fo lkarge r.L.bers of Colm-LllU Lty m rerlrbers. Each group interview should have a moderator and two observers. Ther;.deatr' role is to- fcltediusor,pbekyisseslii IJ IclLtJ,LtICILVJl 0 JJk 1 LV IILLk-LQLC: U1%L ADDLUOO 1t, FIVLVC 11.Cy IDDLAIZ;, C71%ILI comments from all participants, and focus the discussion on issues of iLntCresL VVILiLUhL ouCtViXLU s-i t lterU groupC l.yiiakuL. The t observers role is to take notes on the content of the discussion, the group dynamic process, ar.d th_e comrposition of the gop oti-prat h bevr pkULc~0, cLL L I Lk Lj~LkUi L VI tL ~tU rULLjJ. IVIU~ L III LFUI Lai LL, LI MLI UDZ V 1=Z record the substance of the discussion that takes place while the commu- iUty Udvelops iLs nl-c.p, vziu1 uar1a'dlr., adLU Ut.LILtLlUitCU lIUVVLILcULt. Materials needed for each community interview include a copy of the Intilev1W wud -Ull paperi us uwnmalrileu Ull HUCA cdlUb), ofub u notepaper for recording field notes, pens, pencils and pencil sharpeners, flip-chiari paper, mdrkers (several colors), colored paper, masking tape, and scissors. The participatory interview guide, presented in annex iA, assisrs the flow of discussion at these group meetings. The guide has six sections: 1. Definition of community boundaries and identification of community assets. The group interview begins with a mapping exercise. Setting out iarge sheets of paper and distributing markers, the facilitator asks the group to draw the geographic boundaries of the community. The assembied group also identifies the community's principal assets, resources, and important iocai landmarks on the map.4 This mapping exercise usual- ly generates open discussions, with participants speaking freely about items such as the location of drinking water supplies, roads, and school buildings, and residential patterns. This exercise also opens the way to discussion of other issues. 2. Discussion of case study of community collective action. We have found it useful to focus discussions of collective action and solidarity on a spe- cific case in which the community worked collectively to resolve an issue, whether or not the outcome was positive. The facilitator probes a specific instance of collective action undertaken at some time within the past three years. 3. Discussion of community governance and decisionmaking process. The com- munity informants identify leaders within the community and describe the processes of leadership selection and community deci- sionmaking. Facilitators probe specific instances that illustrate these feaulres of commmunity-level social capital. 4. Identification of local organizations. Both formal and informal organiza- tions are assessed in terms of accessibilihy by different community7 members, involvement of different groups in the community, and eXtent of inusiolln and exclusion. A brief history of each organiztio- n is recorded. For the most important organizations, this information will be suppIlmnented 1-laer w,^.ith data cOlleCfd for th-e Orgariza,.nnal profiles. 5. Asses-srren of the rdolni;0.,hu;rS betwveer 0oirgrzn;Zt;0ns anr the co4lummur.ity This assessment is done by means of a Venn diagram exercise. Organi-a.^;o-s are-assessed A ter- .s of 1b.heir cotrbuin, to ditferent community purposes and their accessibility by different community mtemILbers. This UO7eni diagramt exer.e is oit-iplt to conAuclt anA it Las an immediate visual appeal. The exercise sparked a great deal of inter- est andlit vely tdeateLC ti IaIaALCta anitd Iti"a. 6. Institutional network-s. Another visual exercise, institutional mapping, also serves tL spairk airurIated. discUssIions. IiLLIUAWLCaI ICIa.IuiVitIi`F between pairs of organizations are mapped using a flowchart. All sorts of organlizad-ons working VVwit.inL L,he com l-.uiULy are Lonsidered,:U including local government departments, nongovernmental organiza- Lions (NGOs), co)rL-IIILfULity grotups, and other civil society actols. tL first sigL, Lith pd1LCiidLpatoy Hti eview guide nmtay appear to involve long and complex discussions between researchers and community mem-bers. In practice, however, during the pilot tests in India and Panama these interviews were usually completed within one to three htouur. They generdteu d gredt ueal of userul inufrmation and nelped the research teams gain a clearer understanding of the community and its component parts. Upon completion of each group interview, observers write up detailed field notes, allowing about as much time for the write-up as for tne inter- view itself. When all group interviews for a single community are com- pleted, tne team writes up a summary assessment of the community pro- file. The focus should be on issues salient to the measurement of social capital, such as the collective activities and identified organizations. Issues related to social and economic development in the community are also noted, but these will be covered in greater detail in the second and more structured part of the community profile. The team also writes up at least one case study in which some community-level effort was under- taken in the past three years. The following outputs are obtained from this process of open-ended community interviews: 28 UNDERSTANDING AND MEASURING SOCIAL CAPITAL * Community maps indicating the location of community assets and ser- vices. * Observation notes of group process and summary of issues discussed. * L1iL VI tXTLI1t (ad1U UdesiUre) cfL11t,ltLLu dbbtsLs dItU seIVices. * List of all formal and informal community organizations. * Case study of previous collective action. * Venn diagram indicating the accessibility of different local organiza- tions. * Institutional (web) diagram of relationships among local organizations and service providers. These community exercises constitute a set or qualitative data about the extent and nature of social capital in the community. These data are complemented with the more quantitative information collected with the help of the structured community interviews and the household profile. STRUCTURED COMMUNITY INTERVIEWS. The community questionnaire (annex 1B) inventories basic community characteristics and provides an initial identification of community needs and assets. A subset of ques- tionnaire items (questions 1.16 through 1.19 and section 8) are directly related to measures of social capital. The remaining items record sector- specific information, useful for identifying local needs and formulating programs and projects with a social capital component. Annex 1B focuses on several different types of community assets and services: 1. General community characteristics. These includes age, size, main eco- nomic activities, geographic boundaries, and means of reaching and making contact with the outside world. This information can be used to assess the community's capacity to create and maintain networks within and beyond its boundaries. This section also includes measures of cognitive social capital-perceptions of trust, solidarity, and quality of life-comparable to items included in the household survey. 2. Principal services. The availability and quality of services are assessed: electricity, public lighting, potable water, communication services, waste disposal, public markets, transportation, recreation, and public safety and security. This information can be used for assessing levels of infrastructure development and identifying community needs for projects to improve access to services. The effective provision of many of these services requires some form of collective action by the com- munity. Communication services are indicative of the community's ability to maintain social networks. Taken over time, these data may allow for an analysis of the relationship between levels of physical cap- ital and snocial canital 3. Labor migration. Assessing levels of migration both into and out of the communituy allows greator inderstrldinc of flip efpnf to which com- munity boundaries are fixed or flexible and how receptive a commu- nih, ma-r ho 1n nowAr idoac nnl rolanf-inchine 4. Education. This module describes available school facilities, teaching csnff and rnmm iinihr arocc sn tboeo rocniirroe thrniioug crlinnl nffan- dance. These data may allow for an analysis of the relationship betvveen levels of human capital and social capital. 5. Health. This section assesses principal health problems affecting men, *VorrA '.en, UandALALALI. chldrn avit lJableA LLeLalAI LA.h failties ar.d level of seriLe.V 6. Environmental issues. This unit covers the presence of environmental hazards and percep.i;ons of overall en,iror--ntal nual-it and deterio- ration. These data may be useful in evaluating the relationship Ile-w en levels oC na.lral resource capital doil -caItaI. 7. Community support. This module identifies grassroots groups, NGOs, and governrnent orgaria-.ions *.hat enbecoi.rit re^.est CLI5UVA CILI C LL U51~ ULLUI 13 LI IalL Ci LatIC LUIUllLUIILI Y) II ICItUI I/C LV.' deal with their common problems. Questions are included to assess patterr.s of commur;ur.iy deiso.-.ki. ----- peceve 1levelso risk f'OLLCIII3 Ut LUIILIILLLIUL)' LICLI3AUILULCU'titt6 OILU jJCILCIVCU ICVCI2 Ut 1131'.~~~~~~~~~~~~~~~~~~~~~~~uv%izuiuttu-.LLr,alLA CI-C:VtA l;VIZ and vulnerability. This information indicates the extent to which com- mon solu',;ons tLLu COUILUI- .IUUptIoblems are IIund wiUhi a LUIIlUUrL and it inventories types of organizations that may be studied with the ltiU Ult Utof *erCIIILzIUILol plUtie. IIle cor-ULIULunILy questionUllire tLUeCLts Uasic tiL U)IIL kIU no -wILUnIL LU most adult members of the community and is therefore designed for applicatiUll HIn 1d rUUp btLLLItI8. lilt rIUUF IILLtIVItW blIUUIU UtC UIIUULtU with between 4 and 10 adult community members and their responses shouiii be coueu udiectly on he instrumient. TIhe unerview timie general- ly does not exceed one hour. Not all sections may be relevant for every investigation. The team conuucting the field investigation should select those parts that are likely to be most useful for preparing projects in the given setting. Sections can also be modified, and additional sections added, to reflect the economic or social context relevant to the research. Household Survey Once both parts of the community profile have been completed, work should begin on the second SOCAT instrument, the household survey. This survey should be administered to a random sample of households in the community. it is therefore necessary to obtain or develop a list of all households in the community and then to draw a random sample from 30 UNDERSTANDING AND MEASURING SOCIAL CAPITAL that list. Sometimes a list of households may be available in advance (IfrlI. VoLtII6 Is trLal01t recod orU pIr evious suL rveys, for1 eAal LFI ). If a list is not available, the investigation team must compile one.5 rneU ho-isehoUlU survtey k(al11WX 1C.) lb ULte prlUl.ay ,stbULUmLUIL Uuse for generating quantified indicators of social capital. It has five main sec- tions. 1. introduction. inis section identriies the seiecteu househoiu, registerb the time and date of the interview, and lists the names of the inter- viewer and field supervisor. 2. Household characteristics. This section contains a roster of all household members and their demograpnic and socioeconoric imnormtanon (age, gender, marital status, occupation, education). The module also col- lects basic information aDout tne nousmg umr (type, construction material) and the available amenities (water, electricity, sanitation). 3. Genogram. Tnis module involves the participatory development of a genogram, or family tree, which shows at a glance the relationships both within the household and between household members and the larger community. At a minimum, the genogram systematically iden- tifies the respondent's siblings, parents, children, cousins, aunts, uncles, nieces, and nephews. It distinguishes between family members living within and outside the research community, a potentially impor- tant link for the construction of social networks outside one's immedi- ate sphere. Membership in local institutions is also recorded for each member of the household. Annex 1C includes an example of a genogram. 4. Structural dimensions of social capital. The structural dimensions of social capital relate to the networks and organizations to which the household belongs. Structural social capital is further investigated with the help of questions about past experience, present practice, and future expectations regarding collective action and mutual support. Triangulating information in this manner enables the investigating team to derive more reliable assessments of the structural aspects of social capital. This section of the household profile has four subsec- tions: Organizational density and characteristics. Questions in this subsection are intended to elicit information about the household's current membership in different types of local organizations. A typology of organizations is provided (see code box at the bottom of question 4A.1), and respondents are asked whether they or any other house- hold members participate in the activities of any of these organiza- tions. The quality of participation is assessed by asking whether the household member is a leader, an active member, or merely an occa- THE SOCIAL CAPITAL ASSESSOMENT TOOL 31 sional member with irregular participation. The relative importance to the household of groups and organizations attended by house- hold members is also assessed. The internal characteristics of the characteristics (homogeneous versus heterogeneous membership), deCisiornmakinmg proces-sels and lea-dership patterns (inclusiv-e erefsusl exclusive), and the opportunities provided by organization mem- bership for learning new skills and acquiring valuable informa.on The idea is to obtain information about organizational density, the quality of- ya;.;cip4,;-44- -.-A 4.he ir te.sity of IlLe..Ilers' in-olvement in order to assess the functioning of social networks more accurate- ly (NTraya a Prit--ArchetI 1000 Grootar-- 1 999) ly I\- 7daiay all ln I. 1L.LLL A.', ~l1... ft1 Expectations regarding networks and mutual support. This section pre- sert-LLs a series 0of hyoheia qus-A;oA intnde __ gaueeson aoiec VI iLYFu LIlL%LaI %JLLCOLIIJtI3 XXLLC.L"LLC% LU 5a6ClL ICOytnr- dents' expectations about how community members would act in sibtlafions such- as crop failure, urbar violence, -_ a teGhr lng ~1LaLUi~ ~A.L it LIU LdhU1C, UIUa.IL V1LUIt1l%t, Ui al LVOLICIi D 1U1%- term absence from the classroom, where resolution of the problem would require Lco11lc.ive adtLIoL. LDJiLIeeLL 0ypes of daiv.LiVeL Lcr U included within this subsection; the selection of activities depends Lit *hsecfc corntext. 'skiing respondents to thlillk*hog es on eii specui -O1L~L t~5'i1 I OIUCIL LULIU LILIOUriI LIISCb situations helps researchers understand expectations with respect to coIIecEive ac,ioII dIU solidarity ir LIIt lutUre. If sociaI capital is hig1n, it will be reflected not only in high participation in organizations at the present ilne but also in people's positive expectations about others' behavior in future. * Exclusion. Questions in this sectioni are imtenueu to ninu out who in the community feels excluded from collective action or access to ser- vices, and to ldentiy Tne perceiveu grounus for sucn exciusion. Respondents are asked questions about the community generally ("Wvno in 'tis community is excluded?") as well as about tneir own specific situation ("Are you and your family excluded?"). Understanding who is excluded helps researchers identify the appropriate unit of collective action and whether social capital should be assessed for the entire community or for particular sub- groups. * Previous collective action. information about present organization membership and future expectations is supplemented by asking respondents about their participation in coliective activities under- taken in the past. Social capital has been found to be high in com- mumties in which large numbers of people have participated in pre- vious collective activities (Hirschman 1984). Both current member- ship in organizations and expectations of future collective action should be high in these communities. Answers to the different mod- 32 UN-DERSTANDING AND MEASURING SOCIAL CAPITAL ules on structural social capital thus complement one another, and 4uie mL.easure of socal .capital is approalLeu from different anles. This provides a more complete picture about how the community behlaves as a collec.Live e-;ty ar.d -- w wel pariLcula household is integrated in community efforts.6 ,J. LCUXTLUC Ulf!imteb-Uflo U] sULLU LUapiLUL. Litt: Wo;IULVte UdiletrIsiorU ofU social capital assesses norms, values, attitudes, and beliefs related to solidar- ity, trust, andU reciprocity. Decause theese diliien1sionis cLdIU1t Ue observed directly-people carry them inside their heads-researchers probe indirectly, asking questions about hypoutetical situations that people in a community might typically face. The specific situations addressed in tlese questions willi vary from context to context; the questions included in the annex are illustrative situations that we believe may apply widely, in both agrarian and urban contexts. Wvve encourage researchers to fine-tune these questions to relate them more closely to the communities they study. T'nis section of tne household survey is divided into tnree suDsections: * Solidarity. The hypotheticai situations depicted in this part of the questionnaire relate to situations, such as the death of a parent or the failure of a crop, in which members of a community might need assistance from others. Do members of this community typically assist one another in times of need, or do individuals in need usual- ly fend for themselves? Do people expect that they will be assisted by others and, reciprocally, have the expectation that they will ren- der assistance to others? How widespread is the network of mutual assistance? Do only neighbors or family members assist each other, or can the entire community be relied upon to assist any of its mem- bers when they are in trouble? Communities with high scores on other dimensions of social capital tend to report greater solidarity. * Trust and cooperation. We found that a question commonly used to investigate trust-"Do you think most people can be trusted?"-is flawed in that respondents do not usually know which people are being referred to by the investigator.7 If different people consider different reference groups in answering this question, their respons- es are not comparable. Consequently, we decided to focus the trust questions on specific types of individuals ("people in your village") or on transactions requiring trust ("matters of lending and borrow- ing"). These trust items were found to be consistently useful and valid during the pilot tests conducted across different parts of India and Panama. People in both settings could readily relate these situ- ations to their own notions of how trust functions in their commu- THE SOCiAL CAPITAL ASSESSMENT TOUL 33 nity. In addition, the trust questions are complemented by questions on ----ple's beliefs~ 4, ,-. _ 4_ --Au cope-to and -eir.. 4n sui - .om ~ -- ULL F~F U XLCI3 aL~lJLL LUUtJCIaL1UIL ajit%. L 1LU 1 _11U ULCIL L-UIll munities. * JIL/tLL "IIL"L LIJILJLLkA ICOMl"tUft. rA,L ir.JILpILtL aspctL Ut ofocal capital relates to the dynamics that create conflict and the mechanisms available frrsligconfit mlln olllu,t remr.bers. aaiIC Ul ul ICJCIUVI16 ILL0 L11WIIILr, LUILULIUlLYL Where a large number of interpersonal or intergroup conflicts exists or whier LLUU1IULies are unable Lo eUsolve sULII coUlLHLL Fceaeful- ly and efficiently, social capital may be adversely affected. Trust amuiong corLUtmurtiLy memuers mllay suflter, and' bUondus ofU rec1ipLocLity may be weakened. Mechanisms for resolving conflict and its effects-assessed in termiis of hiarmoLUious relationships witl-ll1 the community-are addressed within this subsection.8 Organizational Profile The organizational profile (annex ID) seeks to assess the internal charac- teristics of specific local organizations and to delineate the relationships and networks they have with other organizations. Both formal and infor- mal organizations can be studlea using this instrument. The organizational profile uses data collected in a series of semistruc- tured interviews with organization leaders, members, and nonmembers. Key information sets relate to the organization's origins and history; quality of membership (why people join, exclusion and inclusion of par- ticular subgroups); institutional capacity (quality of leadership, partici- pation, organizational culture, and organizational capacity); and institu- tional linkages (extent and nature of exchange with other governmental and nongovernmental agencies). Separate interviews with leaders, mem- bers, and nonmembers are recorded as field notes. Characteristics of organizations that are amenable to quantification can be scored using the score sheet (annex 1E), and a comparative index of organizational scores can be generated. Three to six organizations per community are usually profiled, depending on the size of the community. The organizations are identified through the community profile, the household survey, or both. It helps to make appointments with key informants, especially with the leaders of these organizations. Individual interviews can be con- ducted with up to three leaders per organization. Key informants include the executive director (or functional equivalent), a member or members of the board of directors or other recognized leaders, and senior staff. These interviews should preferably be conducted face-to- face, but a self-administered written questionnaire may be substituted. Interviews generally last 45 minutes to an hour. One or two members of the research team can be trained to conduct these interviews. It is useful 34 UNDERSTANDING AND MEASURING SOCIAL CAPITAL to use a two-member team, with one person acting as interviewer and tne otner as note taker. Focus group interviews are carried out separately with members and nonmembers. Obtaining information from leaders, members, ana non- members separately enables the team to assemble a more reliable assess- ment of the organization and its capacity. Depending on tne size and diversity of the organization's membership, one to four focus groups should be conducted. Effort should be made to conduct at least two sep- arate focus groups among nonmembers, one for nonmembers who wish to become members and another for nonmembers who have no interest in becoming members. These focus group interviews generally take about 15 to 30 minutes and include 5 to 12 participants per group. Each focus group should have a moderator and one or two observers. The moderator's role is to facilitate the discussion, probe key issues, elic- it comments from all participants, and focus the discussion on issues of interest. The observers' role is to take notes on the content of the discus- sion and the process of group decisionmaking. Upon completing the focus group interview, the moderator and observers should meet to refine the interview notes and discuss preliminary findings. When the series of interviews required to profile any particular orga- nization is completed, the research team should draft a summary memo of the key characteristics of the organization, its strengfhs and weakness- es, and its relationship with the broader community and with other local- level institutions. This summary serves as an initial analysis of the orga- nization. The research team should complete the organizational score- sheets using all the information at hand. In order to ensure consistency in scoring, we recommend that two or more interviewers work together for assigning numerical scores. Lessons Learnedfrom the Pilot Tests in Panama and India The SOCAT instruments were pilot-tested in 1999 in 14 communities in Panama (including 3 urban, 6 rural nonindigenous, and 5 rural indige- nous communities) and in 9 communities in Rajasthan, India (6 rural and 3 urban). The instruments were translated into the relevant dominant lan- guage (Spanish or Hindi) and, as needed, bilingual researchers orally translated the interview questions into local languages. The objective of this exercise was not simply to translate the SOCAT from the English ver- sion but also to contextualize the inquiry and to situate it within the spe- cific social context being examined. In making this adaptation, it is impor- tant to keep the questions as simple as possible for investigators to admin- ister and as easy as possible for respondents to follow. Any questions in which respondents might feel under some kind of social pressure to pro- THE SOCIAL CAPITAL ASSESSMENT TOOL 35 vide the "right" response should be reworded or omitted. Translated instruments were first applied in Panama, and further modifications were made before the instruments were applied in the field in India. STAFF TRAINING. In each country, a team of local investigators was select- ed and trained to administer the SOCAT. We found that engaging verv experienced or highly educated persons to administer the SOCAT instru- mpnt- is tLnnecessarv: a middle school education is sufficient. The house- hold survey requires the largest commitment of time from the investigat- ing team, as many houephn1ds need to be interviewped However, the household profile is also the simplest of the three SOCAT instruments to adminictpr Tle rnmomminitv prnfilp and esnpecii11v the nro'qfni7ztifnlfn nro- file require more sophisticated skills to administer and record; it is prefer- able that at least one member of each team is a college graduate (see annex 1F for selection criteria and terms of reference for interviewers). Trtainin ic crritial. Orirnnallyx r%lannn.e fCor hATO ridysc the ftrainingo fnr the Panamanian team was expanded to four days. The training focused on the definition and concepts of social capital, a revie. of the study objectives, the development of qualitative and quantitative interview techn-ique, a roVie.,A nf the reaonrch- - imcnr.nnfc 2n fad Oir applicationf and field exercises for practicing the techniques involved. (See annex 1G for the detailed lesson plan used for traini.g. TIn India, t ri las ed five days. A full se. ,-of revisd4 ;i-ru,--nts had been circula--ed ahead of t.m among all investigators and supervisors so that they could study and session. Team members were expected to study these instruments care- < .11- -- A j. 4 A A_ Lb _ L4 4-- ------A A 4 2;-14-1 - -l A ; A juu1Ay altu LU 1lAIlLIy JILCIL .L11L ayy.0X.. i.illtLI L V31 ttllhh"ItaLtL administer. Each of the three SOCAT instruments was discussed fully during Litaining, VVIit speciOal aLLCeLLIVIL gi1VCeL Lt qJUesLtioL LILCth LIM ILVes- tigators had identified as being difficult or complex. Team members were providUue wiLIL an opport JJ IULYnt LU auirisuI UtL:LeU IILLtIIlenLt i Lrcticela sessions organized among both rural and urban inhabitants and, in the cas oftheorgniat'ona prfi]s,with a COr-niurni-n- orgarization. caseoC 01 LICtg an-L/dUcLi1VtCtl Flu"]~ Wb,ItIt c UIILUIULy VCU1dL1L/CL1U1L. Training staff had a chance to see how the process was going and to incor- porate finlc refilenleents before fiueld LebUilg LUllelltet. rIELDWORK t. lAfe-… .1:....:11- er 2]- ..] rIELLL)VKK. After Iraumiiig, Ele nivesLugatrs were uividu Ue iLt edlili, which were sent into different communities. Each team was jointly responsibie for administering tne entire set or SOCAT instruments in a particular community. The field staff in Panama was divided into four gender-balanced teams, each made up of a supervisor and three inter- viewers. In India, each field team consisted of three female and three 36 UNDERSTANDING AND MEASURING SOCIAL CAPITAL male interviewers, along with one male and one female supervisor. Interviewers had primary responsibility for administerinthe h ousehol questionnaires, while supervisors were responsible mainly for the com- rnu i- and orgarnizatuonal -rofiles. Depending upon the size of the community, a research team spent two to fivNe days- in a, crro.-runit;, to conduct the courn.mit, inten.v,-- te household survey, and the organizational interviews. Sufficient time was allottedLi duing1 --L ------- fo; w ...i. up ar.d CIar.ciIiL Ci fied notes and coordinating field activities. Administering the community profile was the fr.st exercise upon entering a town or village. It provided the community with a useful intro- du.A o to the_ pupoes an it - hepe tamr merr*ers 4or-rl general UAULLILUIL LU) LILt: FL`uIyU:z, aWLU IL IM-1F C.Lt LMCu IiLILUCI.D CUJICII r:IL CI impressions about collective action and social organization in the com- IiLU ULy. S.UIIUIUHLYit I.LVI4VWS were usully o16garLizeU ealy inL Litt morning or late at night, times that were most convenient for communi- ty mel-ltIIeLs. 'In PardItdIa, Lthe piIUL LtesL LUtJ plIa.eI UUIULng LiLth higlly pUItLI- cized weeks preceding national elections, which required the research L 1 I 1 I C I. _ _ . . _ tearims to 10e especi-dml-y senitive t re- ui3uemus SUSIL_ULIS d__U ld31 u -IC trust. In India, the pilot test was organized during a relatively quiet peri- od in the agricuirural cycle, when rural residents were nut especially busy with activities on the farm. Participants in both countries showed considerable interest in the pro- file process, particularly while drawing village maps and constructing Venn diagrams and flowcnarts. Community memDers partcipared active- ly in these visually appealing exercises. Venn diagrams and flowcharts also helped focus respondents' thoughts and heiped researchers understand the importance of local organizations, the psychological distance that peo- pie in the community experience with respect to different iocai organiza- tions, and the nature of interorganizational linkages. Useful information was derived about collective efforts undertaken in the past by the commu- nity and about the extent to which these actions had succeeded or failed. Researchers compiled a list of organizations active in the community that could be studied further with the help of the organizational profile. The household survey was administered after completing the communi- ty profile. In Panama, household surveys took between 15 and 90 minutes to complete; in households where no members belonged to a community organization, interview time did not exceed 15 minutes. In India, household questionnaires took between one-and-a-half and two hours to complete. Because some respondents will not agree to be interviewed, the random sample must contain a reserve list of households that can be drawn on. Fewer than 2 percent of respondents in India and Panama refused to be interviewed, however, so very few replacements were required. THEW CCIC A T A DIT ATI A CCSCSCAMNT' TT 37 No difficulty was encountered in administering any questionnaire itpm; Pven among less-edurated respondents. In India, lonwvpr the questionnaire was found to be too lengthy. Ideally, the household survey should take no mnore than one hour to administer, since respondents tend to lose interest if the interview drags on too long. Some survey questions mighif need to beh Amif-ttd in nrder to redciir the length oif flhe ineor1iiolAl Survey items are precoded, so data entry is a straightforward task that can be comnpleted relatively rapidly. It is useful to verify 5-10 percent of the completed questionnaires through repeat interviews before the data nro ntrd n h2 rnr~ci,~~,- f- s,l ] n( -iAri-,Ar -, 1- n''rbe'-b erk of understand-n can b detected and rectified at an early stage. The organiza^tiona1 profiles took the field teams three to four hours to administer and another three to four hours to write up. Several different formal and irforrmal organization s. - 4om ;t A The scori ;indA vided a standardized sumrnary assessment of organizational characteris- tics useful f-or baseine-, iL J- flotoi.ngl_, andV A J evlato 11-;9_. Notes 1. This hypothetical example is discussed along with other real-world cases in T_Is. _ 1.- nnnn\ KIs'-Iftrl (2000V). 2. For lack of more precise terminology, quantitative methods refers here to mnethods t'lat maximize representativity and generalizabili y .o the larger study population, such as surveys based on random sampling, structured individual intervie-ws, and the statistical analyses gelIerated thereby. Quulitutive metnods refers to a wide range of data collection and analysis techniques whose nonran- dom sampling criterion or "saturation of information" allows for in-deptn analy- sis of social phenomena. 3. Tne Ilist or tnese researcn instruments can be found in Krishna and Snrader (2000), which also contains detailed information on the pilot tests. 4. vve found it useful to begin the SC)CAT community profile exercise by focus- ing on what communities have (assets) rather than on what they do not have (needs). Tnis approach avoids creating the impression that the research team has come to the community to distribute tangible benefits. 5. Even if a list is available, it is advisable to verify with community leaders and/or in a meeting with community members that the list is comprehensive and current. 6. In some research settings, it will be useful to supplement this module with questions about participation in sector-specific collective action (irrigation, road maintenance, education, and so forth). This information can be used in conjunction with the sector-specific data from the community profile in the design of projects. 38 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 7. We had included this question in our original questionnaire, and as an exer- cIse, we fUoUlVlw UF Utis tJUlLIJIL by asaL, ng 4CDFpLLUndLet il one I-iaL-LIL villagc who (that is, which "people") they were thinking about in answering this ques- LLVnL. FULCCLL IrUespLondts1Lb sda they IIadu UCIZLL UtILL6 vi of'Cpeopl III uLeiU v1.1a6e, another 13 said they had thought of people in a group of villages (their own vil- lag a :_:_:_ adjo :11_n ___ O _A ____ I _ - L__U _r -__: _____ A_ A LdUC dULU dUJUIUIIb VUldaCb), 0 bdiU LlCly lIdU LILUULILL UL )ILL I rultCudl, dllU ' said they had been thinking of people with whom they had business dealings. 8. Although conflict is often the result of a failure of trust (that is, a cogniuive social capital issue), the mechanisms for conflict resolution, such as courts or counciis of elders, are often part of StruCiural social capital. Thus, the theme of conflict and conflict resolution combines issues of structural and cognitive social capital. R1eferences Berman, Sheri. 1997. "Civil Society and Political Institutionalization." American Behavioral Scientist 40 (5): 562-74. Coleman, james. 1988. "Sociai Capital in the Creation of Human Capital." American Journal of Sociology 94 (Supplement): S95-S120. Eastis, Carla M. 1998. "Organizational Diversity and the Production of Social Capital." American Behavioral Scientist 42 (1): 66-77. Farah, Nuruddin. 1992. Close Sesame. Saint Paul, Minn.: Graywoif Press. Granovetter, Marc. 1985. "Social Structures and Economic Action: The Problem of Embeddedness." American Journal of Sociology 91 (3): 481-510. Grootaert, Christiaan. 1999. "Social Capital, Household Welfare, and Poverty in Indonesia." Local Level Institutions Working Paper 6. World Bank, Social Development Department, Washington, D.C. Hechter, Michael. 1987. Principles of Group Solidarity. Berkeley: University of California Press. Hirschman, Albert. 1984. Getting Ahead Collectively: Grassroots Experiences in Latin America. New York: Pergamon Press. THE SOCIAL CAPITAL ASSESSMENT TOOL 39 Knack, Stephen, and Philip Keefer. 1997. "Does Social Capital Have an Economic Payoff? A Cross-Country Investigation." Quarterly Journal of Economics 52 (4): 1251-87. Krishna, Anirudh. 2000. "Creating and Harnessing Social Canital." ln Partha Dasgupta and Ismail Serageldin, eds., Social Capital: A Multi- faceted Perspective. Washington. D.C.: World Bank. Krishna. Anirudh. and Elizabeth Shrader 2000 "Croqs-C1ilh1ral Measures of Social Capital: A Tool and Results from India and Panama." Social Canital Tnitiative Wnrking Paner 21 Wnrld Bank, Social Development Department, Washington, D.C. Minkoff, Debra. 1997. "Producing Social Capital: National Social Movements and Civil Sorietvy" Amprican Beha7vinrel F1rion-iqn 40 (5): 606-19. NIairauan r na vnn n T i nt Priftc.hctit 1QQ Q "(gtntc :rnA n,C r2i_hi v Household Income and Social Capital in Rural Tanzania." Economic le7velopmenf and Cu-ltuX.ral Change 47 (4): 871-97. Newton., Kenneth. 1997. " American Behavioral Scientist 40 (5): 575-86. Oliver, Pamela E., and Gerard Marwell. 1988. "The Paradox of Group Size in Collective Action: A T-hr o ri-.-1 Ma TT" A- I Coliect. I _ I.011. A __L1Y ..1 ICtCl II31I. £ICL Lf Sociological Review 53 (1): 1-8. Peters, Thomas J., and Robert H. Waterman. 1982. In Search of Excellence: Lessons,41,0m Americal' BCst 1u.u C.ompar,ies. NTew vYork: YAJi -at Booksua. Portney, Kent E., ar. Tefry- . IJCIy. 1007. "oilizir. Minorit; Communities: Social Capital and Participation in Urban Neighborhoods. Amercar Behaviora --SIcie,,tist 40~ An) 632-44.A Xje; g- ..LVV-a iC.C MILLCIt UC,1tULWLUI t JL L~~ ± ~) J.± ULLILaXIL, IXRobertL LD., VVILIL IXUUtIL LRb V1LLeU1 CaIIU Iad IIRaffal NC.t.iU. 1979. Making Democracy Work: Civic Traditions in Modern Italy. Princeton, N.J.: I ILM.lLUIL UJIIVCIILY I l:bb. RXose, Rithard. 1998. "Getting Things Done in an Anti-Modern Society: Social Capital and Networks in Russia." Social Capital Initiative 40 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Working Paper 6. World Bank, Social Development Department, vvashiington, D.C. T] _ _r 1 _T >___ T AT 1 _ Aln r ItN2T___1 __… _ 1T*t2 Rossmari, Gretcnen D., and Druce L. VVnlsoUl. 1905. INUnIIbers anul VVUrUs. Combining Quantitative and Qualitative Methods in a Single Large- Scale Evaluation Srucy. Evatuariun Review 9 (5); 6273 1994. "Numbers ana vvoras Revisited: Being 'Shamelessly Eclectic."' Quality and Quantity 28: 315-27. Stolle, Dietlind. 1998. "Making Associations Work: Group Characteristics, Mrembership, and Generalized Trust." Paper presented at tne annuai meeting of the American Political Science Association, September 3-6, Boston. Uphoff, Norman. i999. ' Understanding Social Capitai: Learning from the Analysis and Experience of Participation." In Partha Dasgupta and Ismail Serageidin, eds., Social Capital: A Multifaceted Perspective. Washington D.C.: World Bank. 3 Quantitative Analysis of Social Capital Data Christiaan Grootaert As explained in the previous chapter, the Social Capitai Assessment Tooi (SOCAT) is a multifaceted instrument designed to collect social capital data at the nousehold, community, and institutional levels. Structured questionnaires are used as well as open-ended participatory methods. As a result, a wide range of quantitative and quailtative analyses are possi- ble using multiple units of analysis (table 3.1). It is beyond the scope of this book to cover all possible uses of the data collected with the SOCAT. The most novel feature is probablv the detailed information about struc- tural and cognitive social capitai at the ievei of the househoid, and this chapter focuses on the analysis of these data. Since these data are obtained through a household survey, the analysis is predominantly quantitative. As a complement, the next two chapters illustrate qualita- tive analysis of social capital data, with a focus on the community as the unit of analysis. The first section of this chapter presents a tabular analysis of social capital data, centered on three basic sets of indicators of social capital: membership in associations and networks (structural social capital), trust and adherence to norms (cognitive social capital), and collective action (an output measure). Tabular analysis is a simple and convenient way to organize data and to extract the basic messages that the data contain. The basic messages of the SOCAT household survey pertain to the extent social capital is observed across different types of households and the main characteristics or dimensions of this social capital. As explained in chapter 1, the typology used here is anchored in the distinction between structural and cognitive social capital. The household information can also be aggregated at the level of the community and cross-tabulated by different characteristics of the community. The main limitation of tabular analysis is that only a few variables can be tabulated at once, making it difficult to discern social capital's contribution to the welfare of the household. The second section of this chapter therefore turns to econometric analysis, in particular the estimation of multivariate models of household 41 42 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Table 3.1 The Social Capital Assessment Tool instrument Data collection method Unit of analysis Type of analysis Community Focus groups, Community, Qualitative profile inter- community mapping, institution view guide institutional diagram Community Key respondent Community Quantitative characteristics interviews, focus and services groups questionnaire Household Household survey Household, Quantitative questionnaire individual Organizational Interviews with leaders, Institution Qualitative profile interview focus groups with guide members and non- members Organizational Scoring by field Institution Quantitative profile score team sheet welfare. Such models aim to identify the contribution of social capital to monetary and nonmonetary aspects of household welfare (consumption of goods, health, and education) in relation to other household assets (land, human and physical capital). A key question in this context is the direction of causality: is it social capital that makes higher household wel- fare possible, or is it existing higher household welfare that allows the acquisition of more social capital? Several ways are proposed for dealing with this problem of "endogeneity" of social capital, as it is generally called. This section also discusses how to analyze the links between social capital and poverty reduction, that is, whether social capital is of special importance to the poor, considering that poor households typically are unable to accumulate other forms of capital. Finally, the section address- es whether it is possible to estimate models of the determinants of social capital, that is, models that can explain why some households or com- munities have more social capital than others. By necessity, the section dealing with the econometric analysis of social capital data is technically more complex than the rest of this book. It is geared toward readers with an interest in and a working knowledge of econometric analysis. However, the section can be skipped without loss of continuity. QUANTiTATIVE ANALYSIS OF SOCIAL CAPITAL DATA 43 Throughout the chapter, examples illustrate the proposed analysis. Most of these are ,a--en from 1the sA dies un.dertlaken in the conteCx Af th World Bank's Social Capital Initiative and Local Level Institutions Study, -0c this boo a...sI to "disserru..ate Lthe: resuAlts VI LthaL VVN. lLThere are, VI course, many other empirical studies on social capital in the literature, an A 1, Aps_er 1 arL 1: sAL_ I.& - -_, - -, - - - - - '_ _.._A 'i_ _C A]1_ :AL_ _1 _ clILU ILU~ I ".L LILC jItUIatUIC: 1CVLVV~ UIL d.aItMACD L ClILU . Ul LIUb LUVUIS contain a wide range of citations.2 Tabulations of Indicators of Social Capital Recent experience in measuring social capital and its impact, including the studies thlat constitute the Social Capita ituautive, nas snown quite clear- ly that empirical analysis must look at both structural and cognitive social capitaU (Grootaert and van Basteiaer 2001). vvnether and now networks and associations function depends on trust and adherence to norms. Likewise, Ihe potential benefits or trust are often fully realized only wnen interactions are leveraged through networks or associations. Thus, a basic set of indicators or social capital must include those refiecting structurai social capital as well as those reflecting cognitive social capital. Some skepticism has been voiced about the measurement of social cap- ital on the grounds that social capital really refers to an underlying social force that eludes measurement and that proposed indicators are at best imperfect proxies. There is some validity to this point of view. Indeed, one must be careful not to equate the measurement variabies with the underlying social capital. Not every association or collective action reflects the presence of social capital; take, for example, associations or collective actions that are the result of government force. However, the tact that proxy indicators are being used to measure social capital does not, in our view, detract from the validity of the exercise. Human capital provides a useful analogy. This theory, developed some 40 years ago, claims that human capital embodied in individuals increases their ability to earn income over their lifetimes. TWo convenient proxies were pro- posed to measure this ability: years of schooling and years of work expe- rience. No one confused these proxy indicators with human capital per se. Rather, the proxies are input measures that capture the two most important ways in which human capital is acquired. Even 40 years after the development of the human capital model, measuring human capital directly (through performance or aptitude tests) remains very difficult. But this difficulty has not prevented the empirical literature on human capital from blossoming and leading to many extremely useful results for developing and implementing education policy. The social capital model may currently be at the same early stage that human capital theory was 30 to 40 years ago. 44 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Based on extensive empirical work, several useful proxies have been lUenUfiedllLU for 1[LCsnU11U6 oLcial capi.Lal lit cl fJUlLy-tlevCUL lltCllLule (Grootaert and van Bastelaer 2001). Specifically, we propose three types of proxy nU icators: mlle,lmbershl lUpsin l0caUld dasocLadLlUt on LU l1tfLW-lks, indicators of trust and adherence to norms, and indicators of collective action: * iviemuersrhps in tucal ubssctiuouris una nerworks. 11ms IIcUaodL0E oI struc- tural social capital is based on the density of associations and the inci- dence oi nousenold members'nips. 'Various aspects of membership (such as internal diversity) and institutional functioning (such as the extent or democratic decisionnaking) are also relevant indicators. Which associations to include in the indicators is culture-specific: agrarian syndicates couid be relevant in one country, rotating credit and savings associations in another, parent-teacher associations in yet another. in the case of networks, which are iess formai, the key infor- mation is the scope of the network and the internal diversity of mem- bership. * Indicators of trust and adherence to norms. Measuring trust and adher- ence to norms (cognitive social capital) requires asking respondents about their expectations about and experiences with behavior requir- ing trust. Key questions reiate to the extent to which househoilds received or would receive assistance from members of their commu- nity or network in case of various emergencies (loss of income, ill- ness). * Indicators of collective action. rhe provision of many services requires collective action by a group of individuals. The extent to which this collective action occurs can be measured and is an indicator of under- lying social capital (at least to the extent that the cooperation is not imposed by an external force, such as the government). As proxies, these three types of indicators measure social capital from different vantage points. Membership in local associations and networks is clearly an input indicator, since the associations and networks are the vehicles through which social capital can be accumulated. This indicator resembles perhaps most closely the use of years of schooling as a proxy for human capital. Trust can be seen as an input or output indicator or even as a direct measure of social capital, depending on one's conceptu- al approach. Collective action is an output indicator. Because of their dif- ferent perspectives, we suggest that these three types of indicators be tab- ulated and analyzed together, in order to provide a fuller picture of social capital and its impacts. flU TANTITATIVE A NT AlIC flUOF cOCIAL CAPITAL DATA A5 Structural Social Capital As explained in chapter 1, social capital helps to disseminate information, rPdu1cPQ nnnnrhinicfirc hPhvuinr ncd fnrilifttpe rnllertive' dprisinnmakinga The effectiveness with which structural social capital, in the form of asso- ciations and networks, filfills fhis rnle depends upon many aspects of these groups, reflecting their structure, their membership, and the way thou funrioinn The SOC'AT nuiocfinnnniro (nqe.ctinnc 4A 1 tn 4A 1 3) .nakces it possible to describe organizations along three key dimensions: the den- sity, nf nrnhambrcMnp t-he diversity. nf rnctrnlorc.hin and t-he ex-tent nf nar- ticipation in the decisionmaking process. DENSITY OF MEMBERSHIP. At the level of the community, density of mem- berschip c rismeasureA qn-n1-, ha, -ho 4 vrh-ohr nf -;vcdnn- nrv,n; ,',-rnc f(nr comparability, best normalized by population size). At the level of the household,Ait is m..easured 1-y the average numi.her of n,oemhbershi-c of each household in existing organizations (this can be normalized by house- 1olA size). This Ia n- ica-r;- n , bel --- sstabulate by loco1-4n (reg;o4n,- province, urban/rural) or socioeconomic characteristics of the house- 1_1A 1-_ _ --- --A _ __/A_._ - 1 144, - -4 1,.._s1- _ {_1;A ILUIlU3 klIL%LJlltC 5lUtFyDI a5C aILLI 6CILICI U LI eI IwtcaI. V1 ILU. LOLLIlt 1CiI6ILIIL ethnic groups) to capture the distribution of memberships. An example for InluueiLaCa 10 6gVCIL Ll LeIIIC 3.2_, wichII-L L 1aIIICO IILCIItLICIOIUyO L) typY of group and province. The types of groups distinguished are of course LULUl)' Oy-spefIIL, bLut U* eC taJICe IIIUsLtrIaa tsitat 1(1 C IC6LUIat ValtiatLUIoLn can exist in the pattern of organizational membership. For example, in the provinces of Jambi aLILd iNUsCt Tleng6aLa IlItUA, CaLbUuL 1I2 FpCILnILL oL meILrr- berships were in finance and credit services groups, but in Jawa Tengah L.hese groups accounte'U for almotoe-hr fl.efn.b-CDILIps. An alternative way to display the same information is to ask this ques- tionv: whLat percentage of hiouseholds participates in W-tIL Lype1 oL ULEIl organizations? This shifts the focus of analysis from memberships to the 1 … ii 'i'1l lilt 1 2 .1 … 2v..~~~~~~~-- _ - 1- 1-nn _ 3 nuusenuhu. idale 3.3, WIULI is uasedu on the samte Uda as ,ale 3.2, ilndi- cates that in Jambi 32.3 percent of households belong to finance and cred- it services groups, compared with 56.5 percent of uouselolds in 1Nusa Tenggara Timur (even though, as table 3.2 shows, the share of all mem- berships accounted for by finance groups is tne same in both provinces). This reveals that financial group membership is more concentrated in jambi and that financial services delivered through local organizations reach a much smaller share of households in Jambi than in Nusa Tenggara Timur. The classification of associations in tables 3.2 and 3.3 is functional- focusing on the prime objective of the association. Otner cdassifications 46 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Table 3.2 Active Memberships in Local Organizations in TdAne-i, by Prvin.e XILL AXL%LAV1KVaX9, Uy A LW.VAILLL~ lawa Nusa Tenggara Jambi Tengah Timur All Type of organization Number Percent Number Percent Number Percent lNumber Percent Social service 346 27.2 616 26.1 733 28.4 1,695 27.3 Production 45 3.5 61 2.6 129 5.0 235 3.8 Occupational 113 8.9 41 1.7 206 8.0 360 5.8 Finance/credit 148 11.6 740 31.4 307 11.9 1,195 19.2 Religious 348 27.3 272 11.5 527 20.4 1,147 18.5 NdTatural resources 19 1.5 08 3.8 U5 2.5) 1/73 2.Q Environrental 1 0.1 9 0.4 0 0.0 10 0.2 Government affairs 189 14.9 418 17.7 572 22.2 1,179 19.0 Recreation 63 5.0 97 4.1 37 1.4 197 3.2 Others 1 0.1 16 0.7 2 0.1 19 0.3 Total 1,273 100 2,359 100 2,578 100 6,210 100 Source: Grootaert 1999a. Table 3.3 Percentage of Households Participating in Local Organizations in indonesia, by Province Type of Jawa Nusa Tenggara organization Jambi Tengah Timur All cl:aiil serviice 57 5 72.8 79-0 69.8 Production 10.5 14.3 29.0 17.9 Occupational 25.0 9.8 39.0 24.6 Finance/credit 32.3 84.8 52.5 56.5 Religious 57.8 58.0 89.3 68.3 Natural resources 4.5 18.5 15.8 12.9 Environmental 0.3 2.3 0.0 0.8 Government affairs 43.8 67.5 87.5 66.3 Recreation 14.8 21.3 7.5 14.5 Others 0.3 3.0 0.5 1.3 Source: Grootaert 1999a. QUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA 47 can also be useful, such as whether the groups operate only in the village, are affiliated with other groups (inside or outside the village), or are part of a federated structure. Groups with linkages often have better access to resources, especially from outside the village, such as from government or nongovernmental organizations (NGOs). Such linkages have been described as "bridging" or "linking" social capital, in contrast to the "bonding" social capital that exists within the group (Woolcock and Narayan 2000). Federated structures of associations have often been espe- cially effective at opening up a wide range of resources for their members (Bebbington and Carroll 2000). The SOCAT asks households to identify the three most important grouns to which they belong (question 4A.3). Tabulating the "votes" received by each organization is often useful. The example from Bolivia in table 3.4 highlights the dominant nosition of the Agrarian Syndicates- considered the most important local organization by 41 percent of repnondents Created bv the Bolivian state in 1952 diiring the land redii- tribution reform, Agrarian Syndicates have become the central commu- nih-v rrani77Ationnq of farm rs tn manaoes lantl foreqft5, and water ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~________-1 of -- resources, and to resolve conflicts. The idetnfifiration of the three mont imnortant organizahfons for earh household is followed in the SOCAT questionnaire by a series of supple- mentary mies.tionn about these orcani7ationn rela-ino'g to tihe internal diversity of membership (questions 4A.4 to 4A.11) and the decisionmak- inmg roc nrd lardprshin (niioctioxnn 4A 19 and 4A 11A Table 3.4 The 10 Most Important Local Organizations in Bolivia Times cited Agrarian syndicate 584 (41.3) Nondenominational NGO 194 (13.7) Mothers club 65 (4.6) Agricultural producers association 60 (4.2) Denominational NGO 48 (3.4) ',%7--men's cer.ter 46 .... (.) Captaincy 43 (3.0) Associahtonnfederaifon of niral peonle 39 (2-8) Heads of family association 39 (2.8) Informal education group 23 (1.6) Source: Grootaert and Narayan 2000. 48 UNDERSTANDING AND MEASURING SOCIAL C APITr AL* DivERsrry OF MEMBERSHIP. It is not immediately obvious whether a high degree of intern-al diveDrcity is a nneiiju nr nPonatkro factnr frnm the point of view of social capital. One could argue, on the one hand, that an inter- n-a11-y ho.mogreu asciafion wATould mnake it- easier for mem.berS t-o trust each other, to share information, and to reach decisions.3 On the -th- I-, A fh-,, m-, le- l,- c cimiTlnr inf rmntin- Cn flnt 10Tcc IATA111 hos gained from exchanging information. Furthermore, the coexistence of a series of associations that are each irtemafrlly homrrogeneo-uss but a1l,-ongl Adif- ferent criteria could render the decisionmaking process at the village level more diffiCult. The multivariate an.alysis discussed later is the b,est vehicle to assess empirically the role of diversity of membership. IJablacJnL, AJVV -ever, are useful to show any reJ i oAMI r sociocnm patterns in diversity. T.he CSC'Ar questior.n.e rr it possble_ to ratea ;e n-trnal1 diver- j.IC J.te ;IjA~~JJUXUIQLC - -taa ,A J3-aUi.- L 1J LAU -AL ALI.ALI ASV sity of an organization according to seven criteria: kinship, religion, gen- ler, age, po4iL.-;ai alliliatlkJL, uLpyaL.C;LL, andL AL.caL.Ao. tIf aprF--ate inl a given country context, other dimensions such as race or ethnicity could be adUUeU. LjiVersiLy infCorrr,ation car be used separaly or colAILLJied in an index. Specific dimensions such as age or gender can be used to refine thle pruoile f oli Ulrg,niLatiosL, foJl t:AdlCltF, Ly irU.diaLItr, ILrVV h IowranLy hLaVC male-only or female-only membership. In Burkina Faso, for instance, the in1ciuenLce of gelUtel-btelegaCIMU asoCILVnLU is relatted oL LItC proVir.Ce adLU its ethnic composition (table 3.5). In Yatenga and Sanmatenga, home of the Mossi tribe, imnore tihan 90 percent of mtemLtberships are inL gender- mixed organizations, but in Houet and Sissili, the cultural preferences of otner tribes iead to male-orny groups in the miajority of cases. A diversity score can be calculated for each organization, ranging from zero to seven (a value of one on eacn criterion indicates tnat memcoers or the organization are "mostly from different" kin groups, religious affilia- tions, and so on). These scores can be averaged over tne three most Table 3.5 Distribution of Associational Memberships, by Gender, in Burkina Faso (percent) Gender Houet Sissdli Sanmatenga Yatenga All Female only 17.2 18.1 2.2 0.8 10.6 M.ale onl1v 50.4 62.3 6.6 5.3 34.6 Mixed 32.4 19.6 91.2 93.9 54.8 All 100.0 100.0 100.0 100.0 100.0 Source: Swamy, Grootaert, and Oh 1999. QUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA 49 important organizations per household and, for ease of comparison, rescaled from. O to 10 0 (in thinn rnrrocnnndina fn fthe bhiget nnocible value of the index). Return ing tote TI,rT,nonei n oavrnnln i- aa slhow cbniAr i-btha nrcgani7- tions are much less diverse in Jambi (average diversity index of 39 out of a -aximum. of 1iff 4-bs,, T -us T-n-a-r, Ti-r (averageo Advrcsitr index of 62). Another interesting finding is that female-headed house- 1,1oAds belong to less-dive1rs 4-b. -.1bA-A b- 1 1 obsehilAc (table 3.6). This; index- procedure -sulsta aLcI, e nhste aewih in measuring the overall diversity of membership. Alternative weighting LILCICO Far Uossible. Fo1;13 CAa.L1JIC., larg5 VeCrI CII I.c .e I,vent) thL eco= nomic criteria (occupation and education) on the assumption that an 0 r g cA-u _ .&=2 _ A_ &A. .&_ Ai s & 1 A. cA r_- uib,dLuZaLiu1L VI FCUVF1C VVlUtL ItUICICItL UL%IUFaLL1UILD VI CLUaLVLIUIL ICVCID FIC- sents greater opportunities for information sharing than, for instance, a group ol people withdifferl e UIiICItLL Cg. VVCI5IgtLL LarL alsiou Le dLIeVCIv iiuiLi a Table 3.6 Dimensions of Structural Social Capital, by Province and Household Characteristics, Indonesia Category index participation in decisionmaking Province Jambi 38.9 63.5 jawa Tengah 57.6 55.6 Nusa Tenggara Timur 61.6 71.4 Head of household Male 53.6 64.1 Anl 1 C7 I Vemlale 49.2 57.1 Religion Muslim 49.2 59.5 Catholic 58.7 71.6 63.7 70.7 Education of head of household None 52.5 53.5 Primary school-incomplete 51.5 60.0 Prmtnar;y sc-hool-complete 53.0 65.7 Secondary school-incomplete 54.1 68.3 Secondarv chnnl-cnmnpletP 64-0 72.9 Vocational 59.2 83.3 University / other 51.9 77.5 Source: Grootaert 1999b. 50 UNDERSTANDING AND MEASURING SOCIAL CAPITAL principal component analysis of the diversity criteria. Sensitivity analysis is recommendce to check iI results relating to ul relevaince Uo .he ulvel- sity of membership are affected by changing the weights. PARTICIPATION IN DECISIONMAKING. Organizations that follow a democratic pattern of decisionmaking are generally believea to be more eiiective than others. Question 4A.12 asks organization members to evaluate the relative roles of the leader and tne members m reacnmg decisions. A com- plementary question, 4A.13, asks for an overall evaluation of the effec- tiveness of the organization leader. Answers to these questions can be tabulated separately by type of organization (to assess whether certain categories of organizations are more democratic than others) or against spatial or socioeconomic vari- ables (to assess whether organizations in certain parts of the country tend to function more democratically, or whether organizations of the poor function differently from those of the rich). The two questions can also be combined in a "democratic functioning score" using a method similar to that used to determine the diversity score: 1. Scale responses to each question. Question 4A.12: 0 = The leader decides and informs the group. 1 = The leader asks group members for their views and then decides. 2 = The group members hold a discussion and decide together. Question 4A.13: 0 = The leader is not effective. 1 = The leader is somewhat effective. 2 = The leader is very effective. 2. Add up the scores and take the average over the three most important organizations. 3. Rescale the total score from 0 to 100. Table 3.6 shows a striking correlation between gender and education of the head of household and participation in decisionmaking in Indonesia. The index of participation in decisionmaking is 7 points lower for female-headed households than for male-headed households, and 24 points lower for households where the head has no education than for households where the head has postsecondary education. OTHER ASPECTS OF STRUCTURAL SOCIAL CAPITAL. The three key dimen- sions of structural social capital discussed so far-density, diversity, and functioning-describe the associations and networks in which QUTANTITATTVE A\AT.YTS OF SO(CIAT. CAPITAL DATA 51 social capital is embodied. They are input indicators. It is also useful to consider a few output or effectiveness indicators. The SOCAT contains questions on two such indicators: the extent to which the village or neighborhood is a source of mutual support in times of crisis (Questions 4B.l to 4B.4) and the degree of inclusiveness in access to services (ques- tions 4C.1 to 4C.8). Two hypothetical crisis situations (chosen to be locally relevant) are pronosed to the respondent, who is then asked to indicate the extent to which the village or neighborhood would act collectively to deal with the crisis and who would take the initiative to set un the resnonse. The five possible answers range from no collective action to full collective action. As with previons tonic's these npeshfions can be tabulated sepnaratelv against spatial and socioeconomic variables, or they can be combined in a single "muhual supnort score " As before. the required assumntion for the latter approach is that the qualitative answers to each question can be given 2 numeric Ialue on 2 common scale, so that addition or averaging becomes possible. Each analyst needs to decide whether this assupption ic acceptan-le in i-he cnntnext nf i-he npecific analys7is indieY-2tApn F'i-tor analysis or principal component analysis can also be used.to see if the ifour uuascfnnc cshara a cnmmnn in.Aonl,ino- fnci-nr Questions 4C.1 to 4C.8 probe for the existknce of exclusion at the level -of ivillage nonr neighhorhood adA for the characteristics that may be the grounds for the exclusion (religion, social status, and the like). The house- hold4 v-lcn asked -rB-fl vhel-ther 4i b-c o-er been 4i- -i-.-n of ectlu- sion. The most policy-relevant information will come from the detailed cross-tabulation of theL pretsencef of extclusion by type of serviJ e -agai-nt he characteristics deemed to be the grounds for exclusion. This tabulation will1 reveal whether exclusion exists across the board, due to specific char- acteristics such as gender or ethnicity, or if the reasons for exclusion vary biy tpe of senVice. For example, access to- C-iA-cai-iL A1) -i-lI-l accoing- to gender, but access to credit may be differentiated depending upon uiMtiLai CIiliICiLlUIL. VVLLi tyF)C LV1 tCtliiC IIavY aU5it L-itatLtiL-- V O i identifying sources of social stress in the community. To compare the inci- iLc of clusion across CLLLuiLiLlUtisC, CUt ClAionUiiLUI score" can Ue con- structed by adding up the answers from several questions; for instance, ULt 11 sUUdlbnWtse LU 4 CUt!LIUn '±L.1JU Cdan easily e scLUItrU in a coUmmLon scale, as can the 11 subanswers to questions 4C.5 and 4C.6. Cognitive Social Capital Measurement of cognitive social capital in the SOCAT is organized around three themes: solidarity, trust and cooperation, and conflict reso- lution. 52 UNDERSTANDING AND MEASURING JOCIAL CAPITAL SOLIDARITY. Solidarity, as an aspect of cognitive social capital, parallels the issuei of mutua!i csupport eiiscussedg earlier iindepr strucr-.lra! socia! capita!. The mutual support questions (4B.1 to 4B.4) asked the respondent 1A7,hOlh-6r f-he uillage or neiaghhnrhoondA -lAmCdelCl g'Pe o0-hor ft A1op de!A.i a hypothetical crisis affecting everyone. The solidarity questions (5A.1 and 5A.2) presont f- i-he m-hoi-.,rI cidi,nf-.n of an un r.-u-ate event happen ing to a neighbor or an individual in the village or neighborhood and throughout the community to assist that individual. Again, the crisis sit- uations nee-d to be selected carefully so as to bhe relavrwnt in te loc10al Con- text. The tabulation of this information can follow a pattern similar to that of3 rr.utual support, that is, Lhe answVers fro0m. each question car. be tabu= lated separately against spatial or socioeconomic criteria (table 3.7), or the an.sVVers tothe tnvwo quest.ons c-an be scored on a com,1.LLon scale ar.d aver= aged. The latter approach is most useful if one wishes to create a "soli- Aar-it score" 113r eacL village or neighbor1hod. TRUa IAND CDOOPERA0I1N. LIu L is au abLitracL concet_FL LLLaL is "L&LCUILt Lto measure in the context of a household questionnaire, in part because it mnay mean different thL-irgs to different people.4 The SOCAT approach therefore focuses both on generalized trust (the extent to which one trusts people overall) cadLL oIL LItM eALtnLL VI LiUsL *LtdL exL I-LI Lth LUILLcAL VI ope- cific transactions, such as lending and borrowing, or taking care of chil- Udren UUrUin LIeil iutitLb dUasItLtn.e lth geneitlized trusL questions dsk respondents to express their agreement or disagreement with general statements such as "most people in this village or neighborhood are basi- cally honest and can be trusted" or "people are always interested only in tneir own welfare" (question 5C.4). Tnese generai questions are balanced by questions asking the respondent to choose between two concrete alter- Table 3.7 Solidarity in Times of Crisis, in Burkina Faso (Percent of households who think they can obtain assistance beyond immediate household and relatives) Response Houet Sissili Sanmatenga Yatenga All Definitely 49.0 32.1 27.2 19.8 32.0 Prnhbbl-, 23.0 295 35.0 35.4 29.0 Probably not 7.9 23.3 5.8 16.5 13.3 Definitely not 10.9 18.8 23.0 24.9 19.4 Difficult to answer 9.2 3.3 9.0 3.4 6.3 Total 100.0 100.0 100.0 100.0 100.0 Source: Swamy, Grootaert, and Oh 1999. UITANTITATTVP ANA1YS!TS OF) S(OCIAL CAPITAT TDATA natives, such as owning and farming 10 hectares of land by themselves versus owning and farming 25 hectares of land jointly with one other per- son (question 5B.5). The joint option is clearly more beneficial to each par- ticipant, but it requires mutual trust between them. Thus. the assumption is that in an environrment characterized by high levels of trust, more peo- ple will choose the ioint option. Because trust is difficult to measure, the questions in this section have a certain degree of rednndancy to them Tn npart, this rpnefitiveness srves the purpose of cross-validating the responses to different questions. It is nposibhlp to, tahiil1tp the answerq to .;ch nuestion against the usual spnaftial or socioeconomic characteristics, but because of the complexity of the concept of fnust; it is recommmended that the analyst usps factor analysis or principal component analysis to identify any underlying common fac- tors across fth diffprpnt niipqeionc. Tncliuiing fihp siibnswers o certamin questions, the trust module contains 19 items altogether, and, in its iinprorgzcsse1 fnrm thisQarmouni-i- of i-nfk.rmit.nn isc imAipiql, Facto-nr analy- sis or principal component analysis provides a convenient way to aggre- This approach was used in a study of social capital in Ghana and TUJa-.d byJNaynar.- d Cassid r'q 2A 001. See qp-p" -pcon_ on-t.ver included in the questionnaire for Ghana and nine in the questionnaire for Uganrda (table 3.8). F.ractor analysis revealedA #A- all h11 varil for Ghana loaded onto a single factor. This discovery would justify con- cftri,ctfn g one index for trust -rid treat-ing it -as one var-iablle in th1be .y sis. In contrast, in the Uganda study, three factors emerged, identifying dAiff.er,,ent d-..enso n s of 4-1. Us t. T hei f.r st fct- C,4.o.r focus.ed, on 4t rust in agencies (police, government), the second on trust in members of one's immediate envirorn.rLent (lfmaily, vilflage., tribe). T1he hiLr fLA IC-LL(-I trust LinL t business community and ward officials. CONFLICT AND CONFLICT RESOLUTION. The presence of conflict in a village A~~~~~~~~-- - _ A I 1A A __ _____A________ -sA_ A 1_- I _ _ 1_ -I t_ -._ U1 nILCg1LLboLUU1II U1 or i aICl16C:1 CICCI In 0L:l alL IltUiLaLUI 0l L1tt IC l,Ll. V1 LiU: L or the lack of appropriate structural social capital to resolve conflicts or 1_6 A - -... C l. … a boLlt.6 Asi -uL onit.,se seen as art ou-p.idct,(r..ta perspective, it leans closer to the collective action indicators discussed in L.ih Itnex. se'iI).Te 4AT 1 qUetiUoliUldie cUIdultai a brief IIIUUUle (questions 5C.1 to 5C.7) to determine the extent to which a given village or neiglblorltouo li InI clofIUllC:t ndU 11 so, tie Ineciaruisms available to nelp resolve disputes. As a measure of conflict avoidance, two questions are devoted to determining the extent to which people are willing to con- tribute time or money to common development goals. Tnree of the questions (5C2, 5C.4, and 5C.6) ask respondents to com- pare their village or neighborhood with others in terms of amount of con- 54 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Table 3.8 Factor Analysis of Trust Variables for Ghana and Udganuda Ghana Uganda Variable Factor 1 Factor I Factor 2 Factor 3 Trust in people in your tribe/ caste 0.786 - - - Trust in people in other tribes 0.757 - 0.533 - Trust in people in your village 0.718 - 0.736 - Trust in people in same clubs 0.697 - - - Trust in business owners 0.645 - - 0.427 Trust in politicians 0.585 - - - Trust in family members 0.534 - 0.369 - Trust in government service providers - 0.719 Trust in local/municipal government - 0.593 'r-. , 4- ...4 - /_-~~_ f) AA '7 "L uS-LI J-fl6-/, - `/U F`-JSJU - Uf. Trust in community/ward officials - - - 0.612 - Not applicable. Source: Narayan and Cassidy 2001. flict, amount of contribution to common development goals, and extent of harmonious relations. Clearly, it makes sense to combine these three questions into a single indicator per village or neighborhood to reflect how the community views itself relative to its neighbors. Three parallel questions (5C.1, 5C.3, and 5C.5) ask respondents to assess whether their village or neighborhood is peaceful or in conflict, whether people make any contributions to common goals, and whether relations are harmo- nious or disagreeable. The answers to these three questions could be aggregated into a total score as follows: Question 5C.1: Conflictive = 0 Peaceful = 1 Question 5C.3: No time or money Time or money contributions = 0 contributions = 1 Question 5C.5: Disagreeable Harmonious relations = 0 relations = 1 Maximum score per village or neighborhood = 3 This type of aggregation obviously involves strong assumptions about underlying common scales. A priori, there is no reason to assume that the "distance" between disagreeable and harmonious relations is the same as between conflictive and peaceful. Yet, in practice, this aggregation method is quite commonly used, and resulting indicators have proven useful, especially in the context of multivariate analysis. However, as suggested several times previously, factor analysis and principal compo- 56 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Table 3.9 Participation in Collective Action (Gotong Royong) ILL 'dLLLneLsia1, bLy P rLoV1iLnc (pFe;LnLLI Frequency of participation Jawa Nusa Tenggara (per year) Jambi Tengah Timur All At neighborhood level 0 30.5 9.8 17.5 19.3 1-6 times 51.3 19.8 31.3 34.1 7-12 times 9.0 15.8 13.8 12.8 13-24 times 3.5 15.5 12.5 10.5 25-48 times 4.8 22.3 15.8 14.3 49+ firies 1.0 17.0 9.3 9.1 Total 100.0 100.0 100.0 100.0 At village level 0 27.8 55.5 17.3 33.5 1-6 times 51.8 29.5 40.3 40.5 7-12 times 12.3 11.0 19.3 14.2 13-24 times 2.0 1.8 7.8 3.8 25-48 times 5.5 1.3 10.8 5.8 49+ times 0.8 1.0 4.8 2.2 Total 100.0 100.0 100.0 100.0 Source: Grootaert 1999a. Combining the Different Indicators of Social Capital The proposed indicators of structural social capital, cognitive social cap- ital, and collective action can be tabulated separately or combined into a single index of social capital. We profess a preference for separate tabula- tion-as discussed in the previous three sections-because the indicators capture different dimensions of social capital that are each relevant in their own right for understanding social capital. Still, a number of stud- ies have constructed aggregate social capital indicators across various aspects of social capital, usually by means of factor analysis. A study of watershed management in Rajasthan, India, selected three indicators of structural social capital and three indicators of cognitive social capital to define a single index of social capital (Krishna and Uphoff 1999). The structural indicators tried to capture the extent to which informal networks and established roles helped the community deal with crisis situations and disputes and whether the village had a tra- dition of looking after common goods. The cognitive indicators captured certain norms and attitudes that represent a sense of solidarity and mutu- QfUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA al trust. As table 3.10 shows, factor analysis indicated that the six vari- ables loaded onto a sincle rommon factor- which accounted for about .55 percent of the combined variance. Hence, the six separate variables can validly be rombined intoa single 'orial capital index which can be used as an explanatory variable as well as a dependent variable in multivari- ate analyvss (see table 319) The study of social capital in Ghana and Uganda by Narayan and Cassidy (2001) cornmined a much larger nunm.br ocnia! rcapital variables and used factor analysis to identify common factors (table 3.11). In Cb22nn ton fnri-rc ipror irlonFifiAo rnthiirion f-iro Aimoncinnc rJf crntinl capital: group characteristics (structural social capital), trust (cognitive social capital), everyday socibiliyJ, volunteerisr, and together-ess. The ten factors explained 48 percent of the variance. In Uganda, the structure nf scnral rapift! proved c;trnnlor, vATidh if--r f -a-ctrsanhrin 6A percent of the variance. In both countries, structural social capital as measured by group characteristics was found to be the prncipl factor AAv1i+!mn+.an Analycis o nf Scial CapitaltD+at The A~ VabIAAfl;tn discusse in 0 she previousl se.a r.Lap the difeen J1L1 IIL'LlaLniLo tfOiA.tOfltAill LIC ±VTI4U ttIil .'nay L.LLC L4IIICICILL dimensions of social capital across spatial and socioeconomic character- i.s C. secton focuses onI LL1L ILfC LI1CC L{U CuL1U1LD tLhatL LIC be autdicsct ILy by multivariate analysis: * What is the contribution of social capital to household well-being, that is, are households wit.h alhigher level of social capital, as tLaZIuLId by the various indicators proposed in the previous section, better off? YVILaL is tLimpltoirLpUiLLtC VI oULicil Li,FiLcal for FupVeILY ltreULLIUIon * What are the determinants of social capital? Table 3.10 Structural and Cognitive Social Capital in Rajasthan, India: Factor Analysis Item Factor 1 Dealing with crop disease 0.73052 Dealing with commnon pastures 0.64826 Settling disputes 0.73272 D'%ea!:rl - L -_ rant chidre 0172029 LiCdhiii WVIUI tCIICTILL CLiiliUlLrI U/ ILYI Value placed on unity 0.78680 Trust -laced in others 0 66859 Source: Krishna and Uphoff 1999. 58 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Table 3.11 Factor Analysis of Multiple Social Capital Variables, Ghana and Uganda Percent of Cumulative Factor Eigenvalue variance percent Ghana GC 3.03 9.19 9.19 Trust 1.95 5.91 15.11 ES 1.84 5.58 20.69 ES 1.65 5.01 25.71 Volunteerism 1.63 4.94 30.66 ES 1.60 4.85 35.51 GC 1.25 3.80 39.32 Togetherness 1.18 2.59 42.90 ES~~~~~~ ~ ~~ 0.85 2.S 4S.49A GC 0.772 2.34 47.83 Uganda GC 7.86 41.36 41.36 GN and togetherness 1.59 8.40 49.77 NC 1.49 7.88 57.65 Trust 1.17 6.19 63.85 GC = Group characteristics; ES = Everyday sociability; GN = Generalized norms; NC = Neighborhood connections. Source: Narayan and Cassidy 2001. These questions address the role of social capital in the poverty reduc- tion strategy set forth by the World Development Report 2000/2001 (World Bank 2001), as discussed in chapter 1. The first question focuses on the role of social capital in creating opportunities for enhancing income and improving other dimensions of well-being such as health and education. We also look at the extent to which social capital improves access to cred- it and thus contributes to reducing vulnerability. The second question looks at the relative importance of social capital in the asset portfolio of poor households. The third question addresses the critical issue of build- ing social capital, a core element of the empowerment pillar of the World Development Report's poverty reduction strategy. The Contribution of Social Capital to Household Welfare To analyze the contribution of social capital to household welfare, we propose a simple conceptual framework whereby social capital is seen as one class of assets available to households for generating income and QUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA 59 making consumption possible.7 The household disposes of an asset endowment consisting of physical assets (land, equipment, cattle), human capital (years of schooling and work experience), and social cap- ital. The household combines these assets to engage in productive activi- ties, either in enterprises within the household or in the external labor market. This process involves making decisions about the labor supply of each household member and acquiring a number of productive inputs (agricultural inputs, credit) and services (education, health), which may need to be combined with labor supply in order to generate income. In this conceptual framework, each decision in the income-generating process is determined by the household's asset endowment, including social capital, in conjunction with the social and demographic character- istics of the household. This model can be formalized in a set of structural equations making up a conventional model of household economic behavior under con- strained utility maximization. By recognizing that the households' con- sumption behavior is a function of the level and composition of its income, the set of structural equations can be summarized by a reduced- form equation that expresses household consumption directly as a func- tion of the asset endowments and other exogenous characteristics of the household, and of the economic environment in which it makes deci- sions. This leads to the following estimation eouation:8 (3.1) a,,r; f= u r _ LC +^Hji +fN _yC + i+-z +H where E = household expendihire ner capita of household i, SC1 = household endowment of social capital, HC. = household endowment of human crnital oCi = household endowment of other assets, Y. = a vector of household characteristics, Zi = a vector of village/ region characteristics, and ui = error term. The key featumre of Lhis mnodel is the assumpti:on that social capital is truly capital and hence has a measurable return to the household. Social capital has ,m,,ii features of capital: it requires resources (especially .h-me) to be produced, and it is subject to accumulation and decumulation.9 Cnri 31 -rti-l fl~u - o . -i i-roA in -.n1 -,. r -r-rn -1 cof4 n iit I-n human capital (for example, schools versus learning-by-doing). Much social capital is - I duin ineacin that occu -fo; soia,reigos,o ou.stnan.0 LiLIA - -hIIL ..kLtCJLtIjLCO LILaL tittLti tItI DUtIa, 1CLL61UDI, VIi cultural reasons. This is reflected clearly, for example, in the pattern of organlizati;onall m- elnlbershi nIdnsa hr Talmost on -hal ofal membershL.cI ips Li a LIinLUFt o izati 1LULhat p VVI Lsu ali rLU s,L orLill ut aii memberships are in organizations that pursue sociaL, religious, or recre- 60 UNDERSTANDING AND MEASURiNG SOCiAL CAPITAL ational purposes (Grootaert 1999a). Other interactions occur in settings opedficalli arr.aed to yierlA orc r baenafisc TIn BRlivia for avnrIa #. k fhi is the case with the Agrarian Syndicates and other production-oriented g (Gnrootaerl a Najr-r-.n 2000) The key assumption is that the networks built through these interac- 4.ions have measurablIe beri.efits. tno the iti V inividas ar.d lad, directly or indirectly, to a higher level of well-being (Putnam 1993). This is the propositI-on tat can be tested empirically b o 3.1.10 The dependent variable of equation 3.1 is the natural logarithm of house- L1 _A AAA ; 1 A A1- 4-_15 11 ThA 1 ' s 5QlOons Os A; J-h _ -OA I 4iWV 'a F'L '' F' ' V.ay-_ i ti- *' LaL- Jtv V 1L V.A) WL 1 Li LC aoot endowment of the household, demographic control variables, and location- al ALU-ul -ariables. Household--- aset consis of p-L-4-1 ... aset, hul at L&.XA L tiY V OIIaVflO.. IAUU3 LU.JI aOD~CL3 ~ULSJIDJL VJI jJIL 9' OtLOt CXOO t, ILULL LLW capital, and social capital. Relevant physical capital variables include the arrourLt of larnd ownred or opera'.ed I- -u. houehld o-vv_Lerhi of a os JLUII .t IWL UVLU Ltuyta-tu Ly U LC ALU UDCtLUi.[A, uVVILtCLOIUF U]. a 11jISco (and its characteristics), farm equipment and tools, equipment and invento- ULCIL IO dL UI ILCUIIcLIL UUMtULU VIL cla dtLU Ct1j ry 'Latipato ar-norull oslll n.rrs,a. a.e As the earlier discussion has made clear, numerous social capital vari- ables are avaiUlable four use in Ul re1gressio. UU U1kVV1Lg car a tfew xaWUltplt of multivariate analyses that have used different combinations of variables capturing structural a.nd cogLnitive social capital. Onte bcL'l q-Ues.IulUL each analysis needs to address is whether different social capital dimen- sionsi should be introduced separately or as an index. Tne literature con- tains examples of both approaches. A study of Tar^-zanian villages by Narayan and Pritche-t (1997) used a multiplicative index of social capital, capturing the density of associations, their internal diversity, and a mea- sure of effective functioning. T1ne justm.tncation for using a multiplicative index is that the effect of the number of organizations to which one belongs may not be independent of the internal degree of diversity or the type of functioning of the organization. The effects are assumed to interact. The analysis in Tanzania found strong empirical support for the use of this type of index. A similar study for Indonesia also found a multiplica- tive social capital index to be significantly and positively related to household expenditure per capita, but in a study for Bolivia only an addi- tive index was found to be significant (Grootaert 1999b, Grootaert and Narayan 2000). Unfortunately, the conceptual and theoretical literature on social capital has not yet provided a sufficiently refined model to jus- tify one approach or the other. In part it remains therefore an empirical question to be tested in each case. Existing empirical analyses of social capital data have found support both for the use of an aggregate index (either multiplicative or additive) and for the inclusion of separate social capital dimensions. Table 3.12 shows a basic application of equation 3.1 to data from Indonesia. It is useful to compare the model estimated with and without nilT A NTTIrATT7rV A XT A T -COTC fl Crlf' A ICA ITAT I 6AT '.4LCI) I, i nil XV 1. 1 flif 1U .310 '.31 .3T0t1lL. '.t11 1 trilL VL'C.tI the social capital variable in order to highlight the relative contribution of social capital to household welfare Including social capital increases the R-squared from 0.21 to 0.24. More important, it reduces the coefficient of hiiimnn canitnl by ahnit nne-third. Thic sugogest-qs that it least Qnme nf the human capital effects operate through the networks and associations cap- hirept in the cnrial rcnifnl inrlex One interesting observation following from table 3.12 is that the esti- mnl-eo rofirnc tn hiimnn and snoia! cnaital aro quiteo cn-milnr A 10 porrcnt increase in the household's human capital endowment would lead to an increase in ovnonAi-iraa ner capifta of 1.65I npercent-,-cnaprnnr r,i+,it 1 1.1 percent increase stemming from a 10 percent increase in social capital endowment. This result is not unusual - nd -4blTn- in si.-uilar empirical analyses for Bolivia and Burkina Faso (see Grootaert 2001 for a ¶-r+J--nec f id +oe rner-lc-e\ lAT- -rnl ;7n rn f1-+ A; inl;-rrrn-4n A; J1 1 010 _o _ I0 .LtO J. _t i ILL-I- It -L1 L -t11t1101 L- the regression results critically depends upon the assumed exogeneity of soc.al capital. ThLis iS a ke.y ass-1..p L ILA1tion t1sdiscus 'ILII fu1lly Iater in this chapter. Ihe lin drawback of usiin a singLeILLIC I inde for soia ital i th1at iL provides little guidance as to which of the included aspects of social cap- Table 3.12 Hnusphnld Welfare and Social Capital: The Index Model Variable Coefficient t-statistic Coefficient t-statistic Intercept 12.7948 69.65 12.6782 67.59 Social capital index 0.0069 6.52 - - Household size -0.0972 10.23 -0.0923 9.59 Years of education per adult 0.0343 4.49 0.0454 6.11 Female head of household -0.0463 0.67 -0.0551 0.81 Age of head of household 0.0309 3.75 0.0354 4.20 Age of head of household squared -0.0003 3.30 -0.0003 3.71 Farmer household -0.2311 5.73 -0.2417 5.89 JaTwa Teingah U-0.163 n3.90 -07 2).40 Nusa Tenggara Tiimur -0.3271 7.24 -0.2201 5.21 Number or observations 1,137 1,137 R-squared 0.24 0.21 P-statistic 33.6 31.3 Notes: 1. Dependent variable = ln(household expenditure per capita). 2. t-statistics are based on robust standard errors (Hubert-White estimator for no-iduenticaily d buted resiuuals). Source: Grootaert 1999b. 62 I UNTCD A ERTAN DIN I fIT G A VNAID f I MASUTINGT I Ci T A PTL.A IDITIfAL ital produce the beneficial effect on household welfare. For example, is it more imnnorfqnt to hteong tn mnnv ornni.ni,qfi-ionn or is it more imnortant to belong to democratic or internally diverse organizations? To address tbis question onp needs to assume that each social rapital dimpnsion acts independently and that the effects are additive. This proposition can be emnirica1!y tested by incliiudincg a spripe of varirjlesp thalit crnhilre hip diif- ferent dimensions separately in the regression model. Table 3.13 shows the results of this exercise using seven dLifferent dirmensions of social cap- ital and the same data for Indonesia that underlie table 3.12.13 NoTei-, that fthe vrriailes irn thic ianmple iniclud only chwirflr] scria! capital dimensions. The results suggest that the most important aspects Table 3.13 Household Welfare and Social Capital: Disaggregating the Social Capital Index Variable coefficient t-statistic Intercept 12.5318 64.66 Social capital dimensions Number of memberships 0.0146 2.43 Diversity index 0.0031 3.16 Meeting attendance -0.0020 0.81 Index of participation in decisionmaking 0.0025 4.29 Cash contribution score 0.0113 1.46 Work contribution score -0.0008 0.27 Community orientation 0.0000 0.01 Household characteristics Household size -0.0947 9.87 Years of education 0.0322 4.22 Female head of household -0.0303 0.44 Age of head of household 0.0298 3.62 Age of head of household squared -0.0003 3.15 _rarmer h 1---e Uhl. _^ l Rn ^ Province jawa Tengah -0.1686 3.56 Nusa Tenggara Timur -0.3446 6.17 Number of observations 1,137 R-squared 0.25 F-statistic 21.7 Notes: 1. Dependent variable = ln(household expenditure per capita). 2. t-statistics are based on robust standard errors (Hubert-White estimator for nonidentically distributed residuals). Source: Grootaert 1999b. QUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA 63 of structural social capital are the number of memberships, internal iversity, and extensive participation in decisionmaking. The coemicien of the membership variable indicates that in Indonesia an additional membership in local organizations is associated with a 1.5 percent high- er household expenditure level.14 The results also indicate that the benefits from participating in inter- nally diverse organizations are higher than from participating in organi- zations whose members are more alike than different. The reasons for this may have to do with the exchanges of knowledge and information that occur among members. Members from different backgrounds may learn more from each other because they have different knowledge to start with. Similarly, people with different backgrounds may be able to pool risk more effectively because they are more likely to have different sources of income. The role of the different dimensions of diversity can be examined further by including each dimension as a separate regressor in the model. The Indonesian data suggest that the economic dimensions of diversity (occupation, education, and economic status) matter the most: organizations whose members differ in economic attributes yielded more benefits to their members than organizations whose members differed primarily in demographic attributes. Of course, whether this result for Indonesia applies elsewhere remains to be investigated.15 Finally, the results in table 3.13 suggest that active participation in the decisionmaking process of an organization increases benefits to house- holds. In Indonesia, the coefficient of this variable is quite large: a 10 point increase in the active participation score (which is a 15 percent increase) corresponds with a 2.5 percent higher expenditure level-a larger effect than from adding a membership. One of the important ways in which social capital can contribute to household welfare is by making household enterprises more profitable. For farmers, greater profitability can occur through better access to agri- cultural technology, inputs, and credit (discussed further below-see table 3.16). In the case of trading activities, good networks of clients and suppliers constitute social capital that complements a trader's financial, physical, and human capital. In situations where contract enforcement is often difficult and costly, these networks lower transaction costs and increase profitability. A study of agricultural traders in Madagascar by Fafchamps and Minten (1999) measured the extent of traders' networks and estimated the contribution of these networks, over and above that of working capital, equipment, labor, and management, to value added and sales. The two most important dimensions of social capital were the num- ber of traders known and the number of people the trader could count on in times of trouble. A doubling of each of these networks would add 19 percent and 29 percent, respectively, to total sales (table 3.14). 64 UNDERSIANDING AND MEASURING SOCIAL CAPITAL Table 3.14 Determinants of Total Sales of Trading Firms in Variable Coefficient t-statistic Capital and equipment Working capital ln 0.3519 4.649 Dummy if subsidiary 1=yes 0.9538 4.290 Value of equipment In 0.0313 1.114 Storage capacity In 0.0438 0.767 Number of vehicles In -0.4861 -2.191 Utilization of telephone !=yes 0.5986 2.186 Labor and management Manpower (in months/ year) In 0.7236 3.234 Dummy if full-time trader 1=yes 0.1340 0.854 Dummy if trader all year round i=yes 0.3834 2.412 Years of schooling of owner/manager level 0.0218 1.299 YearsL of experience in agHI1tLal tr ln 0.0913 1.199 Speak another language 1=yes -0.2227 -1.876 UocalU clu pialU Number of relatives in agricultural trade In -0.2737 -2.794 Number of traders kno.w.n In 0.1924 2.837 Number of people who can help ln 0.2875 3.270 Number of suppliers known personally In 0.0721 1.149 Number of clients known personally In 0.1103 1.704 Shocks Aggregate sales shock ratio 0.1926 2.235 Theft in the last 12 months 1=ves -0.2560 -1.518 Location In canital rity 1=yes -0.5231 -1.482 In another city 1=yes 0.1792 1.386 In Vakinankaratra region 1=yes -0.4315 -1.292 In Fianar/Haut Plateaux region 1=yes -1.2057 -3.403 In Fianar/ C6te et Falaise region 1=yes -1.1150 -3.038 In Majunga /Plaines region 1=yes -0.4667 -1.250 In Majunga /Plateau region 1=yes -0.9785 -2.610 Intercept 4.5114 7.447 Number of observations 681 F-value 32.47 R-squared 0.568 Joint test of nonfamily social capital F-stat p-value 4.59 0.0012 Note: Dependent variable = In (total sales) Source: Fafchamps and Minten 1999. QUANTITATIVE ANALYSIS OF SOCIAI. CAPITAL DATA 65 The multivariate analysis discussed so far aimed to determine the imrnrf nf Qsrcinl carpiftal on household^TeLfare iin genernal as rnfiiroed by income or expenditure. It is also useful to investigate whether social capi- tal ;m,rrnU.oc the nnrfnnlonI-On A iflloflCi of -.1*01 f4a2 Oecially 1 a1lf n"d education. We give an example of each. In many poor countries, access to nta-bl. W,vatr i a cri4t-cl deter.nar.tnof h1e104- A 4-.A- of 4--tr splyi systems in Central Java, Indonesia, found that social capital had a positive eff.ect on the desigr4~, constru.l, and m -tennance of water supply,sys= tems in villages, which in turn improved household health (Isham and Kflar.A.flteIL 199) U~In stUrL6y -, -thes effLect V . oLy for piped water systems and not for public wells (table 3.15). It appears that piped systLemrs reqUIe M.ore --Ae.-v ---.or- ar.d --pr. ocnsr.tar. 3~3LCIL3 CLjLfIC IlJ L~JIICL-LI VC CLII.U L L'.C.t L .UUyCI a .IUI L L' ..J LO'I LLI.LLM C1k,L maintain, and thus the role of social capital is more critical for their success. Table 3.15 The Effect of Social Capital on Household Health in Central Java, Indonesia .Type of water supp!y Sorial capital 7aariable Public wl4l.s Pined cnnnprtinns Social capital index -0.008 0.021 *** (0-.011') (0. 00 5') Density of membership 0.033 0.061 Meeting attendance 0.078 0.106 (0.157) (0.117) Participation index -0.120 0.159 *** (0.119) (0.056) Community orientation -0.077 0.031 (0.127) (0.136) Number of activities 0.062 0.093 (0.067) (0.025) Social interaction -0.08i 0.157 (0.074) (0.057) Neighborhood .List 0.004 -0.094A IN~~1~LIUUIILUUU LLU~~~L U uu'*-- U1 (0.049) (0.071) Notes: 1. Dependent variable is improved household health. Sample sizes are 289 and 588 households, respectively. Other independent variables in the modei (not reported) are agricultural land, years of schooling, household size, and previous use of' river or pondU. 2. Reported are marginal changes in the probability of the independent variable, calc1ultIed from mrrobi,i octin-,2-n (T4-1-hor-rdiiustc-d cf-nAnirA Prrnrc nr2 in nnrpn- theses). Significance levels are *** (99%), ** (95%), * (90%). Source: Isham and Kahkonen 1999. 66 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Better access to education often holds the key to the next generation's ab1ili7y to escap crno, "in-rhy A hli;-br imvolvxementf th ocommuity and parents in the schools can improve the quality of schooling and reduce dropout rates. Coleman (19) fis m-ae +k the role of social capital in the acquisition of human capital in the context of U.S. Tchih hl a it has prve validA i many oher couris as well. For example, a study of Burkina Faso used the village average of the nu-b-r of .9-.e oseod attendedA -arent-Upe asction(TYA) meetings during the year as an education-specific indicator of social cap- ;+al (GrooJtaert, nl, aId Swamy 1999). A poit A nnel 04f -t-non atten- dance, which controlled for many household and village characteristics, fIoun LtaL onlt exLia P Ia CtLCitendar.C ye; Itousildt svvo associatcu vILIL an increase of 3.5 percentage points in the probability that the child atL~~ten A sLi IU. IIn II6IL ViI LItf t UVV lo LAWLUi aLLtteLUCIdac rates LI inBurIinLa Faso, this is a substantial effect. Is Social Capital Exogenous? All the methods and results discussed so far depend critically on the ass-umptlon that social capital is part of th!e household's exogenous asset endowment, that is, those assets that determine income and consump- tion. Th1is assumption needs to be carefrully examined. The formanon of networks and associations of different kinds can be costly in terms of time and otner resources. Conceivably, therefore, nousenolds witn nigner income can devote more resources to network formation and thus acquire more social capital more easily. This is not unlike the situation of human capital, the demand for which also increases with income. The possibili- ty exists, then, that social capital, like human capital, can De at least part- ly a consumption good. The extent to which this is the case depends in part on the type of network or association. For example, demand for par- ticipation in social groups pursuing leisure activities is quite likely to rise with income because leisure is usually a luxury good. If social capital is in part a consumption good, then reverse causality, from welfare level to social capital, is possible. In econometric terms, social capital becomes endogenous in equation 3.1, and its estimated coefficient will be upward biased if the equation is estimated by Ordinary Least Squares (OLS). The standard solution to endogeneity problems is the use of instru- mental variable estimation, which provides an empirical test of the extent of two-way causality. The real challenge in applying this method is to find a suitable instrument set for social capital: instruments must deter- mine social capital, but not household welfare (nor be determined by household welfare). It is not an easy task to identify such instruments, and only a limited number of empirical studies have had any measure of QUANTITATIVE ANALYSIS OF SOClAL CAPiTAL DATA 67 success with this approach. Narayan and Pritchett's (1997) study of social capitap ;r.1 T-r - used aS ...easure of generalz rs as a40 for their social capital index that captured dimensions of structural social capital. T1he -ont -.hat ge- nLeralizt-A 4is buit over time, fis 10 c tion of village cohesion and norms, and is independent of the income Ie of vI a Df":CL`1t hItusehol. YVYILIICe LIs atCIL.CIL has a - a rtain vali,y, the difficulty is that essentially one type of social capital (cognitive social capiLCal) is UeCi as aCU iLLLIUIILMILL for1 CoLIUIhe LyFpe \ ULLU.L1ua1 so`ctl capF- tal). A similar approach was followed in a study on Burkina Faso by r _, _ - _ , n A _ve. .I._ . .cons n r -I X.. A AA_L .ALA __ i-A-, -C --. A "rootadetC, OtI, allU JSWVlmly k1777), VVLIU LULIULlM a lUL ILUCA 101 ULUbI derived from questions about whether people perceived others to be rlltaklir-l Lfil LUIctLiLULIoInb LU lcl:e.'CILLIVILICes caInU wheLULer FUeIC thought they could get emergency help from villagers. The limitations of ine instruient iIUtwIihstillitlg, bo.h sLUUld ILsLfIIU LILhaL IL pasbeU bLstan- dard statistical tests used to check the validity of instrumental variables and supported the interpretation that social capital is a LdUSca i factor of household welfare and not the other way around. The Burkina Faso study also proposed two otner instruments: * Length of residency in the village. Building a social network takes Time, and thus the longer the household resides in the village, the greater the potential for building social capital through interaction witn otner vil- lagers. * Trend of membership in associations. Households were asked whether they were currently members of more, the same, or fewer organiza- tions than five years earlier. The retroactive nature of this information makes it an ideal instrument since it clearly cannot be influenced by current income. it is possible to search for instruments at the ievei of the viiiage or neighborhood. Each of the following variables can arguably be seen as affecting the household's endowment of social capital but not its income: * Ethnic and religious diversity of the community. This variable affects directly the potential diversity of associations, but there is no reason to assume that the village's ethnic or religious composition would direct- ly affect a given household's income. * Density and effectiveness of institutions in the community. Clearly, the pos- sibility for a given household to join an association increases as more associations exist in the community. The likelihood of joining and being active can also be expected to increase as organizations are per- ceived to be effective. 68 UNDERSTANDING AND MEASURiNG SOliAL CAPITAL * The community's involvement in the procurement of social services and infra- structure. Such involvement is likely to encourage people to ,oir.ora nizations dealing with education, health, roads, or other infrastruc- tre T.- A de no , however, have a dir-ct effect on houseold inco-r., UI-J. It LJ-L -L) , ILL) eVL .,. kVL( _h L .tL. ) L. LL LIL) ld tLL)mIL which stems from the use of such services or infrastructure. * Taio. -i Tnce of a tradi-ional. council of village elders or similar institution mandated to deal with conflict or contract CLLUI%CIILUItL IILOay bC a ai6L of1 respctL foL L1QdLILLI arI. aQ rcect to norms. J 1st LULIfn UiLniy LInvUvenCIILCenL L IG IC;I_ LLL VU LJJ U;L/JL4-iLL/1LZ. XIL FCIVF.LIUIL that the community is actively involved in the creation of local organi- L'a 0 -L S mtay itncrease zh_ level of -rut ir, these ------za,=ons ar,.d m;ake- LdLUlL ilclyULL~d~ULC I%=VC1 01 L1UnL 11t ULSE: u1iLLirLL dJt KiC1t it more likely that households would join. Since this is a historical vari- able, u.tIe li 11 psbiU1ULYlit0 fo V%fevrs LciLuIlVoL. Ela-ch of1 ese canI`3dites ior lltbUlUlltIi vaiales LUtt Lcdl be ULJbjLLtdU LU standard econometric tests that will indicate whether the variable in - 2 2~_111 , - ---- - -- - - 1 question is a vaidu instrumnent.1 It is recuLinLi-ienueU uat analysts test eccl potential instrument separately as well as in different combinations to determine fow sensitive the resuirs are to rhe seiecrion or specific instru- ments. Most studies using "me instrumental variabies meMnod nave rouna rnat it led to higher coefficients for the social capital variable compared with the OLS model. Thnis finding suggests that equation 3.i is correctly spec- ified and that social capital is an exogenous determinant of household welfare. if reverse causality were significant, the coefficient of the social capital index in the instrumental variables regression would be lower than the OLS coefficient. The substantive interpretation of thnis result strengthens the case for viewing social capital as an input in the house- hold production function and as a causal factor of household welfare. However, the same econometric result would be obtained if social capital is measured with a high degree of error.17 Unfortunately, it is generally not possible to provide independent verification of the latter. One hopes that the use of multiple indicators for social capital as well as multiple instruments reduces the possibility that measurement error drives the result. Instrumental variable results are, of course, only as good as the instru- ments are believable, and it is desirable therefore to address the issue of causality in different ways. The theoretical ideal would be to have histor- ical data available, since no reverse causality is possible from current income to past indicators of social capital. Both the community profile and the organizational profile of the SOCAT contain several questions about the historical evolution of organizations and the community's role QUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA in creating them. This information captures elements of the creation of social capital in the nast and is ideally suited to test whether a tradition of collective activities and social capital building has an impact on current welfare. Many of the questions in the community and organizational pro- files lend themselves fairly easily to quantification, and variables could be constrLcted that can be added into equation 31. The Indonesian study from which several of the earlier examples in this chapter were drawn illustrates the potential value of surh data= The Indonesian village data files include information on the major develop- ment projects undertakenin the sample villages during the past 10 years and on the degree of community involvement in the design, funding, and im.plem.entat.ion of the project. These data made it possible to construct an index of past community involvement for each village as a variable mea- surming historica! social capital. This variable was added to equation 3.1, along with a number of village-level control variables, including one for the nrumber of development projects that the village undertook. The results indicated that past social capital had positive effects on current incoml,es over and above those stemM.rr ng fror.. cu-rrent socia! capit-a1 ad over and above the effect of the development projects themselves \3rooitaertL 100019b). Thias finLding could reflect the fact Lhat projects wL% id, high beneficiary participation are more effective (Isham, Narayan, and PrkitGheL 1995);, iL is also _in_ _line vvt Pun 1l' (01993) psiti4on tat itis thbe I LLLICLL177)). I I~aia IL AILC; VVILIL I LALILCIIL i0 k ..'j F.IO Ll.Jflfl 1 1 history of civic engagement that explains differences in the economic per- formarCE u.c of0 J n-L.riies. After the use of instrumental variables and historical data (where available), a thi.-d way to tackle the potential endogeneity of social capi- tal is to move away from the reduced-form model of equation 3.1 and attempt Lu estlate str.ctLural equafiions; utey captuCre dIirectly o fLt pathways through which social capital affects welfare. Structurally, social capital acI-eves inpr-velt in vv-elfare LI1UUUIL varioUs ZlII.IClUs: bet- ter access to information, better ability to cope with risk, better collective decisionmaking. it is possible to constrauct variables for eaclh of those channels. Narayan and Pritchett's 1997 study on social capital in Tanzania estimates several equations that explain farmers' access to new agricultural technology. The results indicate that households living in vil- lages with hign levels of social capital have a higher probability of using agrochemicals or fertilizer, although social capital was not found to be related to the adoption of improved seeds (table 3.16). Access to credit is an important way in which households improve their ability to manage income risk (World Bank 2001). The Tanzania study found that high village-level social capital led to better access to credit, although the effect was not very large (table 3.16). A study for Indonesia, however, found strong household-level effects of membership 70 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Table 3.16 Household Probability of Adopting Improved tWICILULCUL I 14aLL1ts in Tanlaia (calculated from probit estimates) Used credit for agri- Used cultural Used Used improved improve- Variable agrochemicals fertilizer seeds ments Village social capital 0.057 0.075 0.015 0.027 (2.35) (2.45) (0.737) (1.66) Household size 0.012 -0.006 0.004 -0.0019 (3.25) (1.43) (1.03) (0.742) L~~t.1i A A~nnA AA'7 Ano nMri nrn Average hlousehiolu 01.u 0.00/O 0.UI0 0.004 adult education (5.00) (1.56) (2.30) (1.21) Female head -(102 -0.112 -0.114 0.0035 (2.89) (3.46) (3.51) (0.143) Assets 0.049 0.110 0.058 0.0069 (2.45) (6.28) (2.63) (0.606) Self-employed in 0.046 -0.035 -0.037 0.027 agriculture (1.49) (0.958) (1.06) (1.03) Median distance -0.013 0.005 -0.005 -0.0052 to mlarkr-et ()2.4 A0.855) (1.16)i (151 Observed probability 0.217 0.197 0.169 0.093 Pred. probability at means 0.155 0.129 0.125 0.078 Number of observations 772 734 765 842 PseudIo R-squared0.204 0.2 54ArA 0147 0.71 I CWu INDtuacL UL.r jJv-ru I, '3, Notes: 1. In parentheses are the Huber-corrected t-statistics of the probit regression coefficients (not the t-statistics of the reported marginal effects). 2. included in tne regression but not reported were d'ummy variables for agro- climatic zones, and for missing values of the assets, schooling, and distance to mr,^lel^. alal Source: Narayan and Pritchett 1997. in organizations on access to credit (Grootaert 1999b). The study separat- ed memberships in financial associations from those in nonfinancial asso- ciations and found that memberships in both types of organizations con- tributed to access to credit. This is an interesting example of why social capital is truly "social": the building of networks and trust among mem- bers in the context of a nonfinancial social setting spills over into finan- cial benefits, such as easier access to credit.18 In summary, this section has proposed a multivariate analysis to assess the contribution of social capital to household welfare. Two approaches QUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA 71 are possible. The first one relies on a social capital index combining dif- ferent dimensions, usually in a multiplicative way. An alternative method is to include different dimensions of social capital as independent regres- sors in the model. This method brings out more clearly those aspects of social capital that contribute to household welfare. Either method assumes that social capital is part of the exogenous asset endowment of households. This assumption needs to be tested empirically. We dis- cussed three methods for testing: instrumental variables estimation, the use of historical data, and the estimation of structural equations. The empirical evidence from studies that have undertaken these methods generally suggests that the overwhelming direction of causality is from social capital to household welfare. However, the number of case studies for which this type of analysis is currently available is fairly limited, and we hope that the increased availability of the Social Capital Assessment Tool will lead to future analyses that will further validate these findings. Social Capital and Poverty The analysis suggested in the previous section can furnish evidence of the extent to which social capital has positive effects on household welfare and which dimensions contribute the most to that effect. One question remaining is whether social capital helps the poor to the same degree as it does the rich and whether investments in social capital can help poor groups escape from poverty. A useful starting point for answering this question is the distribution of the ownership of social capital relative to other types of assets. An example from Bolivia is given in table 3.17. Table 3.17 Ownership of Assets, by Quintile of Household Expenditure per Capita, in Bolivia 1 2 3 4 5 Variable (Poorest) (Richest) All Social capital index Multiplicative index 19.7 20.1 20.2 18.2 16.2 18.9 Additive index 21.4 22.7 24.1 24.7 25.0 23.6 Years of education 3.3 3.8 3.9 4.1 4.5 3.9 Land ownershin (hectares) 2.7 39 2-9 2-7 4 S .34 Animal ownership (number) 19.1 16.6 19.9 19.4 22.4 19.5 Farm equipment ownership (number) 0.5 0.4 0.5 0.5 0.5 0.5 Household durables 1.7 1.9 2.2 2.4 3.0 2.2 Sourre: Grontaert and Naravan 2000 72 UNDERSTANDING AND MEASURING SOCIAL CAPITAL The poorest quintile of households has only 3.3 years of education, compared with 4.5 years for the richest quintile. Land and household durables are distributed more unequally. In contrast, the additive social capital index is only slightly below average for the poorest quintile, and the multiolicative index is even above average-indicating that poor households in Bolivia have relatively more social capital than other assets. The question is whether this accumulation of social assets bv the poor is rational, in the sense that it indeed helps them escape poverty or at least provides them with relatively higher returns than other assets.19 This question can be addressed by three complementary methodolog- ical approaches: a probit model of the likelihood of being Door. auantile regressions, and split-sample estimation of equation 3.1. The nrobit model collanses the distribution of exnenditure into a bina- ry variable that takes a value of one if the household falls below the poverty line This annroach is notentiallv useful when the underlving expenditure data may contain nonrandom measurement error that is eliminated when the distrihition is collapsed (except for some possible misclassification around the poverty line).20 The explanatory variables of thep mn--depl r thp I-nams. as thosn for Pniuation n.1. Stuidipe that have usepd this method have typically found that social capital does significantly ruiiicep the nrnhohi1iht of hbinog noor Mpm hiprcs,in in intprnallv divprsP associations were especially helpful, particularly if members came from different educational arnd occupatonal backgrounds. This suggests that the mechanisms at work are primarily those of exchanging information ,nd knownrl-Age -nd perhaps also the pooling no riclc over households with different sources of income (Grootaert 2001). the poor and the rich in the role of social capital. Quantile regressions est:ftmatC the rergIression liun 1hrou1gh1 5given points on +&I. Aistr.1-utio oC the dependent variable (while an OLS regression line goes through the r.eanL) ar.d car assess wvletlhr -ce tai&. CAyiaCLaLly fctors are we-'-er or stronger in different parts of the distribution.21 Results for Bolivia, Burk11a IPdaU, anltU ILUt11i alI sUg6gesL LILaL LILt re-LUrIns tUo socUial tali, as measured by an aggregate index, are highest at the bottom of the dis- tribuilon (table 3.10). The third method available to investigate differential returns to social capital is the split-sample approach. The samlpie should be split accord- ing to an exogenous variable, such as education or landholdings. Splitting the sample into poor and nonpoor subgroups is not advisable, because it is likely to introduce selection bias into the results. Split-sam- pie estimation of equation 3.1 for tne same three countries indicated thnat the returns to social capital were in each case larger for smallholders than for households with higher amounts of iand (Grootaert 2001). QUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA 73 Table 3.18 Poverty and Social Capital: Quantile Regression Re Isuits Coefficient of social capital index at 10th 25th 75th 90th Location percentile percentile vIediani percent,ile percentile Bolivia 0.0059* 0.0052* 0.0055* 0.0055* 0.0044* Burkina Faso 0.0062* 0.0047* 0.0027* 0.0014 -0.0003 Indonesia 0.0096* 0.0090* 0.0078* 0.0048* 0.0049* Note: * indicates significance at the 90% level. Source: Grootaert 1999b; Grootaert, Oh, and Swamy 1999; Grootaert and Narayan 2000. Taken together, findings from the available studies suggest that returns to social capital are generally higher for households in the iower half of the distribution, whether by expenditure per capita or land own- ership. This is perhaps the sense in which sociai capitai is the capital of the poor; they do not necessarily have more of it, but it provides them with greater returns and hence occupies a more prominent place in their portfolio of assets. These results, like all the other examples in this chap- ter, are country-specific. However, in the settings investigated, they make the case that promoting the participation of poor households in local organizations is a potentially valuable ingredient of poverty alleviation policy. The Determinants of Social Capital All the analyses proposed so far have focused on the potential contribu- tion of social capital to household welfare and other outcomes, that is, they have treated social capital as an explanatory variable. This treatment is consistent with the conceptual framework in which social capital is seen as part of the asset endowment of the household. Although social capital shares many attributes with other forms of capital, it is funda- mentally different in one respect, namely, its creation can never be the result of one individual's action. It requires interaction between at least two people and usually among a larger group of people.22 If social capital is not subject to the same person-to-person market exchanges through which, for example, physical capital can be acquired or sold, then how does it come about? The literature has emphasized that the creation of social capital is a complex process heavily influenced by 74 UNDERSTANDING AND MEASURING SOCIAL CAPITAL social, political, and cultural factors as well as by the dominant types of economic activities. The construction of empirical models with social cap- ital as a dependent variable will therefore have to be much more complex than models that merely seek to assess the relative contribution of social capital together with other determinants of well-being. Hence, great cau- tion is needed if data from the SOCAT are used for multivariate analysis with social capital as a dependent variable. The number of determinants of the creation of social capital that could actually be captured in such a model based on the SOCAT data (or, for that matter, based on data col- lected by almost any practical instrument) is likely to be a small subset of the total set of relevant variables, so any such model would be subject to significant specification bias. Nevertheless, efforts have been undertaken in the empirical litera- ture to see to what degree certain observable factors can explain various indicators of social capital. One example is a study of the role of collec- tive action for conserving and developing watersheds in Rajasthan, India (Krishna and Uphoff 1999). Although the study focused on exam- ining how much social capital contributes to the development and the maintenance of watersheds, it also looked at the factors (at the level of the household and the community) that were associated with higher levels of social capital. Specifically, the study examined the relevance of eight possible determinants: (1) prior experience with collective action, (2) existence of rules of behavior in the community, (3) extent of partic- ipatory decisionmaking, (4) number of sources of information, (5) edu- cation, (6) economic status, (7) demographic characteristics, and (8) dis- trict history. As table 3.19 shows, the first four of these factors turned out to be highly significant predictors of the amount of social capital, as did district history. The coefficients of these four significant variables added up to more than one standard deviation in the household distri- bution of social capital. In contrast, attributes such as education, eco- nomic status, and demographic characteristics did not prove to be sig- nificant predictors. Another shtdy used a multi-enuation system to analyze the deitermi- nants of three types of social capital and their role in explaining why some commuinities in Dhaka, Bangladesh, werp ahbl to organize them- selves to arrange for the private collection of solid waste (Pargal, Huq, and Gilligan 1QQQ9'. nlid waste collection is a public good that involves positive externalities and thus the role of the community is vital since in-rPnfivPe for ir.dividual acti-on are limifi-e The JiilAr hypnhothesized that three aspects of social capital would be relevant, namely, trust, reciproci- ty, and willif-gness to sh-are goods with pople in n-ee, T1he st-A- -plic- itly recognized two-way causality whereby each of these three factors determines the probab;ity that a trash collection system.ou berga- QUANITTATIVE ANIATYSIC OFP SO)CTAT CAPITAL fDATA 7 Table 3.19 OLS Regression of the Determinants of Household- Level Social Capital in Raiasthan, India Variable Coefficient Standard error Intercept 27.38*** 2.49 Prior experience Experience of collective action within the last 12 months 5.27*** 0.27 Prior collective management of common lands 0.74* 0.30 Rules Clear and fair rules relating to common land development 4.85*** 0.66 Participation Participative decisionmaking vs. decisions by chiefs alone 1.09* 0.58 Decisions made by all vs. decisions by technical specialists 2.09*** 0.57 Information Number of sources of information 0.80*** 0.15 Education Number of years -0.59 0.45 Status Landholding 0.04 0.02 Caste status -0.16 0.27 Demographic variables Gender 0.41 0.56 Family size 0.08 0.12 Length of residence in the village -0.25 0.20 History (district dummy variables) Bhilwara -1.69* 0.81 Rajsamand 8.60*** 0.83 Udaipur 8.03*** 0.84 Number of observations 1,451 R-squared 0.453 F-ratio 79.24 F-probability 0.0001 Note: Significance levels are *** (99%), ** (95%), * (90%). Source: Krishna and Uphoff 1999. nized ana mat once sucn a system existed, it wouid contribute to ennanc- ing trust, reciprocity, and willingness to share. The authors therefore esti- mated a simultaneous equation system consisting of four equations, as summarized below. 76 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Trust = f (existence of trash collection; nousehold chl-laracteristics; community characteristics) Reciprocity = f (existence of trash collection; household characteristics; commnunity characteristics) Sharing = f (existence of trash collection; household characteristics; community characteristics) Existence of trash collection = f (trust, reciprocity, sharing; household characteristics; community characteristics). Tne study found tnat reciprocity and norms of sharimg nave a signii- cant positive effect on the probability of organizing for trash collection. In terms of the determinants of social capital, the share of households with a business and the share of residents that own their home were positive- ly and significantly associated with all three measures of social capital (table 3.20). This may reflect that the business community in Dhaka is fairly close-knit, but it also implies that the members ot the business com- munity foster trust and norms of reciprocity among other community members including those not involved in business. Likewise, homeown- ers appear to have a stronger effect on community social ties than tenants, who may be more temporary residents. While such studies illuminate the relative importance of personal attributes, the household's experience with past collective action, and institutional factors such as the presence of rules and participatory mech- anisms, still the studies cannot explain why some communities have clearer and fairer rules than others or why they have a stronger past expe- rience with participation and collective action. It is doubtful that a quan- titative analytic approach is likely to shed much further light on this question beyond the kind of results provided by the previous examples. Progress will have to be made by an integration of quantitative and qual- itative analytic methods. Much of the existing social capital literature can be classified as either quantitative or qualitative, but there are very few examples where the complementaries between the two methods have been successfully exploited. Caveats This chapter on the analysis of social capital data closes with a few caveats, specifically about how not to use the data from the Social Capital Assessment Tool. The prime objectives of the tool are to measure struc- QUI1ANTITATIVF ANALTYSTS OF rC)CIAT CAPITAI nATA 77 Table 3.20 OLS Regression of the Determinants of Social Capital in Dhaka, Bangladesh Trmst Reciprocity Sharing lAariable ~~~~~ ~~score score' score Intercept 0.8427** 1.8705** 1.4387** (0.1966) (0.1398) (0.1523) Presence of voluntary solid 0.1348 0.0916 0.1022 wastL e MI ri -a gCenCent systern. In089 (0.0596 in0.06S0)nlr Median tenure 0.0041 0.0040* 0.0024 (n nnOQ (n nn91 i (n nwiT Origin = Chittagong 0.1388 0.1641* 0.2580** (0.0889) (0.0632) (0.0688) Homeowners (%) 0.0104* 0.0074* 0.0167* (0.0061) (0.0043) (0.0047) Business jobs (%) 0.0076** 0.0070** 0.0111* (0.0037) (0.0027) (0.0029) Number of meeting places -0.1318** 0.0702** 0.0138 (0.0378) (0.0269) (0.0293) Number of private organizations 0.0736 0.0274 0.0915 (0.0841) (0.0598) (0.0652) NTurriber of public -raiain 0.0286Z "0499 ".0376 iN ti L'I VtJUIL 010 LILa Li1J1ItD 0.40 ¶J.JLt7 vUVol (0.0492) (0.0350) (0.0381) Adjusted R-squared 0.2626 0.2611 0.3819 Number of observations 65 65 65 t-statf;s.; foc1r exogene i.y test 0.60S0M 1.434S 1.791 LDLOLILIL 10 CAIJ6iLCY LCDL u.tJ.J 1.~to.lo 1.0/ 71 Note: Standard errors in parentheses. Significance levels are ** (95%), * (90%). Source: Pargal, Huq, and Gilligan 1999. tural and cognitive social capital at the levels of the household and the A 1;o ET nAr' ; 4 A, 'rl -._;A4 -44- s; als co 1 ..; b,, UlOll 11 LULl ty. 1In t-.UIJII 11. LU-l L1ta FLUL lJ 0 l LId. LI SIV I to understanding the links between different organizations and their rel- evance for 4uhe provision ofl various, -ervice :_ the orr,r,uitj Both 4.he ClC1`01 ute ~1VDUL 01 VCIIAoLuo DC;LI% V 0lAD UL ILL L0I1LIILuLILLLY. LJ0LItLtILL quantitative and the qualitative data can be used to assess the impact, p o1si L 0ve or negaL6ive, ol soc0ial capitLal on a range o01 outcom.es at the house- hold and community levels. The analysis discussed in this chapter has large'ly -tCayCu WVILiLU. LIllb I1CL11CVVkUJ1.. PZ VVtW IlCtVt F0ULtCU UUL, ClIOILb CIt explaining how the observed levels and types of social capital come abo-ut have been successful ou- to a very nyimited degree. WvtIe caution therefore against expectations that the data from the SOCAT can be used to explain now social capital is created and why it is weak or strong in certain communities. The dynamics of the creation of social capital are 78 UNDERSTANDING AND MEASURING SOCIAL CAPITAL complex and involve many political, social, and cultural factors. Efforts to analyze these dynamics need to rely on a wider range of tools than the SOCAT. We emphasize that the tool was primarily designed to assess the contribution of social capital and not the process of its creation. One of the important findings of the empirical social capital literature is that the social capital characteristics of a community can make a great difference in project outcomes. Communities that have strong social cap- ital are more successful in managing irrigation projects, water supply and sanitation projects, and many other types of infrastructure projects. Education projects benefit from the presence of cohesive and well-func- tioning parent-teacher associations. There are many such examples. It is therefore tempting to target interventions to communities that have strong social capital and to use tools like the SOCAT as a screening or tar- geting device. In principle, the SOCAT can be used for this purpose, but the danger of this application lies in the lack of available benchmarks. Consider the parallel with poverty targeting. Data from instruments such as the Living Standards Measurement Surveys are routinely used to calculate household expenditure per capita and,. on the basis of this, the incidence of poverty in different communities. Interventions are then fre- quently targeted based on the extent of poverty. This targeting can be done because poverty benchmarks are well understood and well estab- lished. The literathre as well as the develonment nractice of using pover- ty thresholds has a tradition of several decades. This is not the case with social capital Fven if an inftnirment like SOCAT were able to rneasure social capital perfectly, using the information as a targeting device would still be dangerouiis hecauise there is in'uifficient knowledoe ahoiit which dimensions of social capital are important for project success and what the crificrl levels are fnr earh nf them Ac we haxe dncuiimented-f in thbi chapter, the different dimensions of social capital have different effects on ouitcrnme anc these efcts are not cd- acrnc+2nn 2rrCcc crn,-nriac For earr,-_ ple, the internal diversity of associations was found to be a critical deter- rr inantof outorn. i 4-r ,nA Tndon ar.d -Burkina Faso but notin Bolivia Hence, at the current stage of knowledge, it would be extremely risky to set benchmarks for the density of ne51 VVorks ar.d orgar.-zaItions or theL1ir internal diversity and then attempt to target project interventions to com- IL.uItiesO IL.ttiLnL -sargeCO. Of couIrs, LIUs doCe noILt coJILLILULCe ant argument for not using the SOCAT and similar tools; quite the opposite. O.JIlUy thL,rio fICAJUCILL aFFpliaLiILD Uo LLLe LtoJI VVwi a Cw tucas Uo b UIIL uF from which generalizations will become possible and that might at some poirt Hi LIt ILUUit lteCadU o su'Lf11iCILL kno-Jwt-lUe Lu iLLUILtLY Lithoe as~JptcL. and levels of social capital that are responsible for success in project out- com-nes. Thiat M-0whledge -would provide vadluabUle li`UUI`0 atiO for iuture efforts to strengthen social capital. QUANTITATIVE ANALYSNIS OF SOCIAL CAPITAL DATA 79 Notes 1. For further information on these activities, see Grootaert (2001), Grootaert and van Bastelaer (2001), and World Bank (1998). A-, +he w.rld .A:de WlAie, see www.worldbank.org/socialdevelopment and www.iris.umd.edu / adass/proj / soccappubs.asp. 2. See also the reviews of the social capital literature in Grootaert (1997), Portes (199), ---IJU1j1.ik (1998), and Woolck.. --Nryan (200J0. -any cs)' are also discussed and cited in Krishna, Uphoff, and Esman (1997); Narayan (1995);. --Ao TT1993); an pof smlan, ar,d Yis.ra 198) 27. 3~), 1.. F-'13- k. 3 I.p a, its 1 3 -F.rt - .31 Lar - -L. X 1110~. k- -/I. 3. Evidence indicates that homogeneity facilitates the adoption of new technol- °by (R3 ogers lnn5, TfhLm l7n). 4. For a further discussion on trust, see, for example, Dasgupta (1988), 12 1 ------_ _ /1 nA=\ -3 fln,A70% FUIUydlilCl k177.y), aILU %.IMilUVULLtL k17/.0). 5. The SOCAT questionnaire does not include questions about trust in specific ndiviUUdali or 0y1pes of individudals (1reLaVti, 1ieLtIU, Itea Ltrs) or about -U,nst in institutions (police, courts, the government). 6. The genocides in Cambodia and Rwanda are extelme examples of the link- between violence and the lack of inclusive social capital (Colletta and Cullen 2UUU). 7. As explained in chapter 1, the SOCAT is an instrument for collecting social cap- ital information, and as sucn, it does not contain modules to capture-inuormation on land and physical assets, nor does it contain modules to capture income or expen- diture variabies. T-ne analysis discussed in tnis section rus requiires that mue CAI be combined with modules from other surveys to collect the necessary information. Chapter i ists several surveys availabie ror tnus purpose. 8. This reduced-form model has been used by a number of studies on social capital, such as Grootaert (1999b); Grootaert and Narayan (2000); Maluccio, Haddad, and May (2000); and Narayan and Pritchett (1997). 9. Events in transition economies such as Russia and former 'Yugoslavia are powerful evidence of the effects of the decumulation of social capital (Rose 1995). 10. As said, equation 3.1 is a reduced-form model. In structural terms, the returns to social capital could be measured in earnings functions (for instance, if one's network helps in getting better-paying jobs or promotions) or in the various functions that determine access to credit, agricultural inputs, or other factors that enhance the productivity of a household enterprise. 11. An altemative specification would be to use the natural logarithm of house- hold expenditure per equivalent adult if data are available to support the estima- tion of an adult equivalency scale. In countries where regional price variation is significant, the expenditure variable should be deflated by a regional price index. If such an index is not available, regional dummy variables can be included in the regression to capture price differences and other differences not observed across regions. so UNDERSTANDINTG ANID `AST-TRI SOCTIAT CAPITAL 12. The inclusion of physical asset variables in a regression with consumption as dependent variables could be problematic due to posible -nd-o-neity Some households may indeed sell assets to pay for current consumption. Ideally, this problem car. be addressed by incluId.. ina vriablels tbht crptlirp fip storklc of assts at the beginning of the consumption reference period. In practice, this type of information is rarely availaabble, and +he al,alnySt -i11 naed tn mzlcae ima, ment as to whether possible endogeneity bias would be outweighed by the specification bias resulting frorr. dropping -uch, Triables In rrra it i ri. -rm-mndeA d-hat equations be estimated with and without problematic asset variables to see if key 13. Some of these dimensions can be derived from the information collected by the SOCAT, nAot b -ly thenum.ber of ...e.,.berships, the dV-Sy index, ar.d the index of participation in decisionmaking. Other variables, such as meeting atten- dance an Acommuni,i orien-tion w not -c-edi the SCAT in prt because previous analyses did not systematically support their relevance. 14l.eei coeprllli h i.epta.6o of th, col,cet fh n and t. IILci IS a %_IVZC yaiaiiCI Ut LAIC XLLC.IjJIC LAVAIJL -i ULAC VI.II.IIL An -LILI L. social capital variables. The former represents the returns to years of investment ir.ULaeLIdLuca. thoILIugLh LjUI atLLCiLuanL_C. "L_ tLAC caseC V. sial capital, the II.OIL input is also time, and the coefficient measures the returns to the time spent in _1evelopin-g --woks a.enir. asoi.o .et1g,adtelk.1- "lLll .ieca UL~V~1OJUI~1Lt:LtWU1fb, IILLt:ILUU1Lr, aOULIaiutILPL LUrLUL6, aiLu UtC ILIAC. IILU LLIALC MaL be spread over many years. 15 A sh-lar cxf Inding -w-s ob.tai nej f or Bni r.a.. Fak- .so (Groo.aert OhL _.d C_._ 1999), but an analysis for Bolivia did not support the importance of internal diver- sity for outcomes (Grootaert and Narayan 2000). 16. Commonly used is the test for overidentifying restrictions proposed by Davidson and MacKinnon (1993). For applications of this test see, for example, Grootaert (1999b) and Narayan and Pritchett (1997). 17. in that case, the instrumental variable method merely corrects for tne atten- uation bias in the OLS results. 18. This interpretation of social capitai has been proposed by severai authors, including Dasgupta (1988), Fukuyama (1995), and Putnam (1993). 19. The ideal data set for answering these questions is a panel data set, which fol- lows the same households over time. However, since panel data sets with social capital data are very rare, we focus on the analysis of cross-sectional data. 20. There is some evidence to suggest that measurement error is strongest at the two extremes of the distribution. The probit model is especially useful in such a circumstance. For further discussion, see Grootaert and Braithwaite (1998). 21. However, the estimation is conditional upon the values of the independent variables and hence coefficients from quantile regressions are not comparable with those of OLS regressions. Specifically, the coefficients show the effect of a marginal change in an explanatory variable on the xth conditional quantile of the dependent variable (Buchinsky 1998). QUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA 81 22. This issue is related to, but nevertheless distinct from, the question of collec- tive versus individual ownAership of social capital. Both position have been advanced in the literature, with Putnam (1993) being perhaps the most noted pro- ponent of the view., l-,at social capital is a collective asset. Others, such as Portes (1998), suggest that social capital may well be individually owned even though Thus, the process of asset creation should be distinguished from its ultimate own- ership. 'M.e m.odel underty-ig the analysis proposed in this zA-te- -1-a- t4-e- the position that social capital can be individually owned. References The word processed describes informally reproduced works that may not e commioluy av a1 1_1u L-A-L LU LI zli1Illly avl:llA _ *UtIOU IL lUAlO eLbULbI6LU1t, P-1tLILU1Ly J., andLA IIOILa %_ IIarro. 2.000. IndUUL.U oULial Capital and Federations of the Rural Poor." Social Capital Initiative lvA3rklll PaperFt 17. Vv^3Orld Barl-, SoLcia "DeveulopmerLL Dea.UFtL1MLLen, Washington, D.C. Buchinsky, Moshe. 1998. "Recent Advances in Quantile Regression Models." Journal of Human Resources 33 (1): 88-126. Colemanl, James. 1988. "Social Capital in the Creation of Humanl Capital." American Journal of Sociology 94 (Supplement): S95-S120. Colletta, Nat J., and Michelle L. Cullen. 2000. "The Nexus between Violent Conflict, Social Capital, and Social Cohesion: Case Studies from Cambodia and Rwanda." Social Capital Initiative Working Paper 23. 'vvorld Bank, Social Development Department, VVashington, D.C. Dasgupta, Partha. 1988. - irust as a Commodity." in Gambetta, D., ed., Trust: Making and Breaking Cooperative Relations. Oxford: Blackwell. Davidson, Russell, and James MacKinnon. 1993. Estimation and Inference in Econometrics. New York: Oxford University Press. Fafchamps, Marcel and Bart Minten. 1999. "Social Capital and the Firm: Evidence from Agricultural Trade." Social Capital Initiative Working Paper 17. World Bank, Social Development Department, Washington, D.C. Fukuyama, Francis. 1995. Trust: The Social Values and the Creation of Prosperity. New York: Free Press. Granovetter, Marc. 1973. "The Strength of Weak Ties." American Journal of Sociolog, 78: 1360-80. Gr0aetr Chrilcfi-<< 1Q07 `Cn,4nI- Cn-ni+n1. Tho M;cc;T1 T inl,7" fh Ii- 6 in Expanding the Measure of Wealth-Indicators of Environmentally Sustain.able n-vlome.t 1A[-k;-gon, D.C.: WATorld B Me - e-1 ........ . . aflLi LtI ). .199910a. "ocal .^;-.n evc eie noei. Local Level Institutions Working Paper 5. World Bank, Social D. evelop...ent DL-.-t.e, LIIL6L'iL, D.. - . 19199. "0ocial C.apital, H e 10Lt3 U'.A v LYaI, ld iova ui Indonesia." Local Level Institutions Working Paper 6. World Bank, Socia1 Dleveioprnent TJp. L.ItILL, VvaDLL. LLt1 D.C. -.2001. "Doe - LIdal Cap-ital H1elp tLh Poor1 A Synthesis 01 Fir.ir.g from the Local Level Institutions Studies in Bolivia, Burkina Faso, and T__:T1 T _._1I T_I T_,:_-- IA_1: T--__'- n-- n YA7_12- )_1__ 1IU0L1tfbid. L.ULa L,eVe U Llb LiLULIULIL VL0lb VIrHL I lU 10. vVULIlU UdlL'u, Social Development Department, Washington, D.C. Grootaert, Christiaan, and Jeanine Braithwaite. 1998. "Poverty Correlates I T __ _1 t- - l __ - - - -f-- - -.l-- -_ - - _-- 1 _-_ _ _ . 1 - r __ and Indicator-Dased Targeting in Eastern Europe and tne rormer Soviet Union." Policy Research Working Paper 1942. World Bank, Washington, D.C. Grootaert, Christiaan, and Deepa Narayan. 2000. "Local institutions, Poverty, and Household Welfare in Bolivia." Local Level Institutions Working Paper 9. World Bank, Social Development Department, Washington, D.C. Grootaert, Christiaan, Gi-Taik Oh, and Anand Swamy. 1999. "Social Capital and Development Outcomes in Burkina Faso." Local Level Institutions Working Paper 7. World Bank, Social Development Department, Washington, D.C. Grootaert, Christiaan, and Thierry van Bastelaer. 2001. "Understanding and Measuring Social Capital: A Synthesis of Findings from the Social Capital Initiative." Social Capital Initiative Working Paper 24. World Bank, Social Development Department, Washington, D.C. Isham, Jonathan. 1998. "The Effect of Social Capital on Technology Adoption: Evidence from Rural Tanzania." University of Maryland, Department of Economics, College Park, Md. Processed. QUANTITATIVE ANALYSIS OF SOCIAL CAPITAL DATA 83 Isham, Jonathan, and Satu Kahkonen. 1999. "What Determines the L,11E:L111CILCDC UL _11U11LuL1LY-u~aScJ YVaVLC1I LJ..L,1AL ILUXII Central Java, Indonesia, on Demand Responsiveness, Service Rules, adL ASocial Capital." Social Capit l .- IaiVti vAUL1rkiL T .14___ . WorldU Bank, Social Development Department, Washington, D.C. Isham, Jonathan, Deepa Narayan, and Lant Pritchett. 1995. "Does Participation Ilmlprove Perflorlida-ICe rsaUbishii-g Ca-usality with Subjective Data." World Bank Economic Review 9 (2): 175-200. Krishna, Anirudh, and Norman Uphoff. 1999. "Mapping and Measuring Social Capital: A Conceptual and Empirical Study of Collective Action for Conserving and Developing Watersheds in Rajasthan, India." Social Capital Initiative Wvvorking Paper i3. vvorid Bank, Social Development Department, Washington, D.C. Krishna, Anirudh, Norman Uphoff, and Milton Esman, eds. 1997. Reasons for Hope-instructive Experiences in Rural Development. W'est Hartford, Conn.: Kumarian Press. Maluccio, John, Lawrence Haddad, and Julian May. 2000. "Social Capital and Household Welfare in South Africa, 1993-98." journal of Development Studies 36 (5): 56-81. Narayan, Deepa. 1995. "Designing Community-Based Development." Environment Department Paper 7. 'World Bank, washington, D.C. Narayan, Deepa, and Michael Cassidy. 2001. "A Dimensional Approach to Measuring Social Capital: Development and Validation of a Social Capital Inventory." Current Sociology 49 (2): 49-93. Narayan, Deepa, and Lant Pritchett. 1997. "Cents and Sociability- Household Income and Social Capital in Rural Tanzania." Policy Research Working Paper 1796. World Bank, Washington, D.C. Pargal, Sheoli, Mainul Huq, and Daniel Gilligan. 1999. "Social Capital in Solid Waste Management: Evidence from Dhaka, Bangladesh." Social Capital Initiative Working Paper 16. World Bank, Social Development Department, Washington, D.C. Portes, Alejandro. 1998. "Social Capital: Its Origins and Applications in Modem Sociology." Annual Review of Sociology 24: 1-24. 84 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Putnam, Robert D., with Robert Leonardi and Raffaela Nanetti. 1993. Making Democracu Work: Civic Tradition in Modern Italu. Princeton, N.J.: Princeton University Press. Rogers, Everett M. 1995. Diffusion of Innovations. New York: Free Press. Rose, Richard. 1995. "Russia as an Hour Glass Society: A Constitution withniit Citi7pnq " Fast Furnonean Constitutional Review 4 (3): 34-42. Sw nm.J, Anann Christian (Z rnotaprt ind GTi-Taik COh 1999 "Local Institutions and Service Delivery in Burkina Faso." Local Level Inst-itiions Working Paper 8 World Bank, Social Develonment Department, Washington, D.C. Uphoff, Norman. 1993. "Grassroots Organizations and NGOs in Rural Development: Opportunities with Ditminishing States and Expanding Market." World Development 21 (4): 607-22. Uphoff, Norman, Milton Esman, and Anirudh Krishna. 1998. Reasons for Success-LT earn...in ,- , Instructivei Experienoc in Rujra! De71lok7m1nf West Hartford, Conn.: Kumarian Press. Werner, Silvia. 1998. "Local Level Institutions and Collective Action." Processed. Woolcock, Michael. 1998. "Social Capital and Economic Development: Tol-wvar' a c_.-L__.;ca Syn.esi ad-I- Poic Frl.ev rk.'TLaer _r, IL) iu a tLICLLCI YL1d J 1t:RLi Ii LUL 1 .11i_Y 1-1ai11VVU1LS. I z UY"t Society 27 (2): 151-208. Woolcock, Michael, and Deepa Narayan. 2000. "Social Capital: inmplicationis for Level)pmIerIL 11ettUly, ReseahUI, aIIU Polc-y. Vvurtu Bank Research Observer 15 (2): 225-50. World Bank. 1998. "The Local Level Institutions Study: Program Description and Prototype Questionnaires." Local Level institutions Working Paper 2. World Bank, Social Development Department, washington, D.C. 2001. World Development Report 2000i2001: Attacking Poverty. New York: Oxford University Press. 4 Qualitative Analysis of Social Capital: The Case of Agricuiturai Extension in Maii Catherine Reid and Lawrence Salmen This chapter focuses on the cognitive dimensions of social capital, that is, interpersonal trust expressed through the relationships among a society's members, institutions, and organizations. The manner in which people relate to one another in and through institutions affects the oualitv of their lives and the degree to which they are able to improve them. Develonment becomes more effective when it is based on an under- standing of the trust placed in institutions and their intermediaries. This is particularly tnie of poverty reduction activities because of the lack of trust in formal institutions among the poor. The obiective of the research renorted in this chanter was to gain an operationally useful understanding of trust. We did so by studying the relationshins between agriculhtral extension agents and contact groups (working groups of 8 to 15 villagers) in the training and visit system of acrriciilhiral extension in Mali W.e sought to dpermine if the tnlst between farmers and agricultural extension agents contributes to increas- ine aoricliltral nrnoduucfion andi if the triict hewhArePn rcntacft orroinc and other members of the community determines the effectiveness of the groups as catalysts of comm.unity development. The very nature of the object of inquiry-trust-is one about which lit- tie useful information will ha r-e,aled linlcc a 1high degr-e nf -nist ic established between the interviewer and interviewee. To identify the peo- ni aA nc1-birhrnc one) Anna <"A Annern4 4-r, .e4- anA nrb,'4-n-irc bnaiA -.. - . ..v,v. _ A- -.. A- --f -..f. - .....J.i.A ... ly laden with social, economic, and political implications-one must first trWs shtte esn owoml one- is -ivulg_. -UGh -Arlaonwll not- use it in harmful ways. IT aLChive thIIs goval, vvC usacd LreiLciILIaiy asascaasct iLr.eLILIodology, which builds on the establishment of trust to derive insights. Beneficiary 86 UNDERSTANDING AND MEASURING SOLIAL CAPITAL assessment contributes to the creation of trust between interviewer and U ILtC V ICVVV by US 1Ag LIuaUlLa LIVe rIeseaIL. LI meLLLL J ant Ito Lr.LSIil -oL fidence and create rapport. These include using an interview guide for conversafi;onal intlelvie-wirg v whic1-aallows the inter:v-iew,Vee to bVecomle the LU1LV1~c1IUI ILC VIC vv UL6, VVIII It VV LLI IILI VCVCCIL subject of a discourse he or she leads; keeping paper and pencil out of sigh't as r..uchI as possible Iu ng A-: e ireve1_ re ntt ret h 51~~tL ~ IIULL 15 pSSI'ICUUIHI6, ULC IItLCI VICVV UI tjULUI ILUL LUJ LICCaLC: ULCL impression of an interrogation; and beginning interviews with questions itat estadbishi Ute ULLCLviCWCI concern f0r thle ai.teviewe. These and similar techniques show respect and put the interviewee at ease. vve iocuseu on three aspects of trust. ITI te fini 1 thLe LItUL - CLV eIe t.e agricultural extension agent and the farmer. Such trust may be necessary before farmeers are IwI-i Lun to foow exteltsion agen.s advice. iThe seI.UIcn is the trust that binds the members of the contact group into a cohesive whole. The thirr is the trust that iranscend. the group aind allouwLs it to acL for and serve the interests of the larger community to which it belongs. The Extension System in Mali Information about agricultural techniques in Mali is disseminated by traveling extension agents, who meet with small groups of villagers, wno then share what they learn with the rest of the village. Extension agents first meet with villagers during a general assembly to explain the training and visit approach. During the general assembly, the extension agent explains that contact groups act as conduits tor passing tecnnicai intor- mation on to the rest of the village and asks interested villagers to orga- nize themselves into small working groups of 8 to 15 members. The extension agent visits each of the three or four contact groups in a village every two weeks to discuss technical themes and demonstrate new techniques. Specific agricultural techniques are taught during hands-on demonstrations in the fields of a contact group member. In the- ory, demonstrations are open to anyone interested, whether or not the person belongs to a contact group. In practice, the process for inviting nonmembers is left to the discretion of contact group members and the extension agent. Technical themes address specific agricultural problems that are com- mon in the area or that the village has identified during an annual diag- nostic exercise that brings together the community, the extension agent, the agent's direct and regional supervisors, and various technical spe- cialists (for agriculture, environment, and animal husbandry). Information from the diagnostic exercise is then centralized and the upcoming year's themes are chosen at the regional level. Extension agents attend monthly training sessions that give them information about the technical subjects and themes chosen for the year. OUALITATIVE ANALYSIS OF SOCIAL CAPITAL: MALI 87 Research Design This study of the relation between social capital and the effectiveness of extension service is based on information collected through fieldwork in six villages in Mali. In each village, we collected data on four variables that may influence the effectiveness of agricultural extension: regional variation; the performance of the village (the ease with which villages created contact groups, attendance at biweekly meetings, and receptivity to and acceptance of new agricultural themes); social cohesion; and the qualitv of agricultural extension agents' work. SELECTION OF REGIONS. Two regions in Mali, Kayes and Segou, were cho- sen for study. Results from the two regions differed widely. Villagers in Kavec anneared to adont technical themes less often than villagers in - _ --F E _ -- --r - r Segou, and contact group members in Kayes were less willing to diffuse information to others Only a qniarter of the farmers in Kaves who were not contact group members reported receiving most of their agricultural information from rcntact group me-mbers living in their communities TIn contrast, farmers in Segou adopted new techniques more often, and con- tart arniun farrmers: servePd verv effPetivevly intermpdinries with 76 npr- cent of farmers reporting contact group farmers as the source of most of thoir naciri,lfiirnl infnrmra-nHr TDifferent levels of ocial cranital may explain the differences in information transfer, thus making Kayes and Segou i choic for this study. Kayes and Segou also differ in other respects. Kayes is nearly inacces- sible, .ATirh a rorcky mniin1Kqinniic terrain that makes agriculture and trav- el difficult. In contrast, Segou is the rice basket of Mali, and agriculture there. can' be c,cmnorriallx, rew.arding. Mianyr develnnryenl p~ro,ects are taking place in the region. Unlike Kayes, where a significant proportion of the rm.ale working popi--ation emigrat-es per.-M,an-nl-, em-i-ratin frnrm Segou is seasonal. SELECrION OF VILLAGES. Villages were selected on the basis of performance vari"ables related Ato agricaultural extension. Documrentation of vill-g v a IL'~O tl ULed t L6JL I to ,i LI ClUI LOtIt I.-, LIL. LA I,jLtJ V h L~ benchmarks by the regional directorates of the extension service provid- uI iLIUa.j;on Ul on tLedcuial aspecLt U. sucLt,io h LSUGh as hectares planted and number of demonstration plots. The regional director and supervisUory staaff -wtere thereforeJ1 askdV L Ito AUs V10vlagesC Ltheycniee to be high performers and 10 they considered low performers using three 'tLerla. tliC eae iit Wli-II V111drtf: CdLtfcLU %LUiLLcLL r'1uF0', LLLLUcIL criteia -.he ease -w-i, -w--hich vilgscetdcnatgop,attendance at biweekly meetings, and receptivity to and acceptance of new agricul- . I . 1 x 1 1 ,. 1 I 1 1 . J _ .1_ 12- / . L 1 tural tnemes. Villages were tnen rdnuomuily seleeteu fromii thlebe listb tbtUlt: 4.1). The research team decided to include one high-performance and two 88 UNDERSTANDING AND MEASURING SOCIAL CAPIfAL Table 4.1 Villages Included in the Study Type Kayes Segou High performer Tantoudji Kolodougou Coro, Soke Low performer Kassama, Sambaga Tingoni Bamanan low-performance villages in Kayes and one low-performance and two high-performance villages in Segou, in keeping with the results or an ear- lier beneficiary assessment. SAMPLE SELECrION. Within each village, in-depth interviews were con- ducted with male and female contact and noncontact group me mbers (table 4.2). In all, 90 individual interviews were conducted (see annex 4A for interview guide). Contact group members were selected randomly from the extension agent's list of contact group members. Other villagers were selected from census information, from the extension agent's recorded data, and from information provided by the local government. The team also completed a simple mapping exercise with key infor- mants to gain insight into each village's dynamics. Informants were asked to draw the different sectors of the village and then describe the population living there (by ethnicity, clan relationships, economic activi- ty and level) and how that population differs from that of the other sec- tors. The homes of contact group members were then indicated on the map. It quickly became apparent whether all contact group members were clustered within one sector or population-an issue that could then be explored by the interview team. Mapping proved essential to under- standing the inclusion and exclusion of different populations in the con- tact groups and ensured that each subset was interviewed. Two focus groups-one for men and one for women-were conducted in each village. Fifteen people attended each focus group-5 contact group members and 10 other villagers. Bringing together members and Table 4.2 Distribution of interviews Village Contact Contact Nonmember Nonmember group men group women men women T, antoudli 3 2 5 5 Kolodougou Coro 3 2 5 5 ____ 5 0 5 5 Kassama 3 2 7 3 Sambaga 3 2 5 5 Tingoni Bamanan 5 0 5 5 Total 22 8 32 28 QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: MAL! 89 nonmembers created a dynamic tension that fueled a productive dia- logue about the principles and reality of diffusion as seen from each side. Two of the three villages visited in the Segou region had no women con- tact grouns. In these villages. women's focus groups were held with each village's traditional women's organization (about 50 women participated in each rounp). Fxtension agents were also interviewed in each village about their experiences working in the area. In all, roughly 220 people were interviewed Lhrough focus groups. The Rnle of Snfinl Creihpznn The importance of another variable, not originally included in the study, became apparent while conducting interviews. Village cohesion, or unity, was:\ not scnorcific-allY njr1OroccoAl inM Fh lfriin1n _frnnr s hr appeared to be a very strong correlation between the two factors. All three high -performance villages shox,,ed numerous indications of social cohesion; the three villages rated low performers showed evidence of Interviewers recognized the relationship between performance and a preexistingI fLabric of rAIlati n s o hips thoAuL a ni finLduI vc. ProAcos Of UobOCi_ vation and open-ended interviews. Cohesive and fractured villages dif- fereA st.-ikgl,-- for- -xl.e in - thirailtyt meet or.llu. needs I I -AO I &rU5Y, iU Ml iJi, IlL LXLCLL aL'lJIlLy LV IICL %LU11UXL&1ULy IMCCLW (table 4.3). In every village with low cohesion there were stories of failed atter.pts to const..L.1ct or maintaiit; public goods. Fr-jU'h discussion revealed that each failure was rooted in the lack of unity within the com- munity. OpL en conflict or subtlCe dlssension arr.ong eCL-IL gru1ps, clanss , or family groupings often centered on the placement of public goods. In TiniLULL UamILrLtL, fUl eCamILlCe, a LoLal LULV V Cl1111 Lt:Ial 1rC[111i.dCLIU1L offered to provide the village with a water pump. The village elders coulLd. not agree on whether the pur. shoul be A_ce in _3_e ole -pr luIL IU 11CUL VIIMZII LILtC FUlltF bMLUUIU UR FICLLt:U IIL URLC Uluci FdlL of the village or in the new, more populated sector, which had recently b-een settled near a na,'onal road. La.k_ of agreemrent, resulted -r the can- LCCI CLLICU lici ILUILCIIi lULCI. LdIL'A VI drICCIIIIL CU1C lit Ltue LcIt- cellation of the project by the NGO (two neighboring villages received punrips). Similar incidents took place in each of the socially fractured villages. T, _J - --- ---- --- J - _ - - -1- . -1- . -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -* - -- .. ,- EachLI fail-ure seem to create more uisirust anu uissatsiacuon, wnicn in turn reduced the possibility of future success. In Tingoni Bamanan, the undercurrent oi recrirmination and hopelessness tnat foiiowed tne failure of the water pump project was exacerbated by seeing neighboring vil- lages with pumps. This viiiage's experience stood in contrast with the cohesive villages, in which infrastructure (water pumps, health clinics, community grain storage facilities) were visible proof of villagers' ability to work together. 90 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Table 4.3 Differences between Villages with High and Low LeveIls or aocial Conesion Tantoudji Kassama Attribute (high level of social cohesion) (low level of social cohesion) Village Village association Main naths covpred with weeds cleanliness weeds monthly. and trash. Diel-rihiil-inn nf Mnc infr,ctr,irr is . h,1^, HATC% +--mlr pumps, hni-4 infrastructure in village center, equi- located at an administrative according to distant from the building 3 kilometers from need different village sectors, village. Each half of village did other infrastructure not want to allow the other half spread throughout to have access to the water village. pump. Ability to Village built two Loss of outside funding for road organize for literacy centers and construction due to lack of maintenance or a communal grain organization. Bridge in village construction of storage facility. washed away because work public goods Youth group purchased was never completed. Three antenna and television school classrooms left for village. Village urifinished. credit association fun.cti^ons . Number of Village association, One informal women's group, organlLduuns youLll guup, and very Vne lilLUn tli failuiy ld-rnling dynamic women's group. Village meetings rarely group. Traditional -calld groups have been transformed to meet development purposes. LeTadership Villnap rchif deleate.dq T--unlcl fighting oer responsibilities to chiefdom. members of different clans and ethnic groups. Physical Several mosques in Mnonii ronf ravTing in condition of good condition. mosque Friday aftemoon Filled to capacity. Very few people present. nravpr attendance The presence of social cohesion cannot predict individual interest and ability to adopt technical themes, but it does indicate the abilitv to mobi- lize for group activities (such as contact groups) and to diffuse informa- tion within communities. It therefore becomes an extremely imnortant QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: MALI 91 element in the planning and supervision of development projects. As one woman put it, "Community cohesion is like a rope. If it breaks, water can no longer be drawn from the well." Quality of the Agricultural Extension Agent's Work The final variable to be considered is the competence of the extension agent and the quality of the agent's work. This variable consists of two components. The technical component considers whether the agent understands agricultural principles and the technical themes taught. The second component, which concerns rural development in a broader sense, considers whether the agent is able to understand a village's socio- cultural context and group dynamics and to motivate others. Assessment of the capacity of each of the six extension agents stud- ied was made in the field by the interview team, which included two agroeconomists and two rural development experts. The opinions of the team were then supplemented by each agent's supervisor. Villagers also evaluated the technical and pedagogical abilities of their extension agents. The training and backgrounds of extension agents differed widely. Of the six agents studied by the team, two were not formally trained in agri- culture (one had been a forester, the other was trained in animal hus- bandry). While agent background did appear to affect the technical themes on which the agent snent the most time, all agents appeared to have a firm grasp of the content and criteria of each theme. The interview team did notice considerable differences between effec- tive agents and ineffective agents in maintaining and updating their doc- uments (village history, contact groun member list, activities ner agricul- tural season) and their regularity in keeping biweekly appointments with contact group members. Tn two villages with ineffec:ive agents. neither contact group members nor the agent could firmly recall the appointment scheduile The more glaring differences between effective and ineffective agents were at thp nedagognirca! lovxl The NM-finnl EFxtension Servirp nrnvides agents with only a few general guidelines about development principles; even some supervisors appear inadoeiqutely trainpd in this s,iihiprt Supervisory visits appear to focus exclusively on technical matters. aonfc rornrd rhcarvr:nti nd rnmmonf-e m lAo hwu ci rsTicnre in, nthor official visitors in their notebooks. The interview team did not find a sin- gle r-mark an an of these notebooks regarding- eitbr Adif.cfuin nf irfonr mation or community concerns and dynamics. Comments were limited to technical matters, such as growth rates of different rice aand Millet vari- eties. Moreover, in two of the villages visited, the extension agent under- mined trust with key segments of the village by forming contact groups exclusively with rertain castes and ethnic groups, neglecting others. The ineffective agents also had a poor understanding of the yearly diagynoIstic meehing annd were iinmhl to itipntif-v nnnnrhinitipc for nnpro- priate activities. In short, they appeared to care little about the success of the peopl with- ---m --ey ---ed Effe-iv -gnt fet responsibili- ty toward their contact group members and to the village as a whole. Even villages w-ith little internal cohesion responded -ith trust to this conscientiousness. As one villager from a socially fractured village said, "...r aInt'C Anor ic -l1AInno -x,n ,fa rnA+niht- orr,,lA -A Q him b~~~ -- , ---.~-F-1 - -- - "-.5" -. --_ -.k A- to ask a question." Similarly, another contact group member said, "He never leaves w,n,fil he's sure hat ever-one understands the tech,ical message." Results and Analysis Two variables-the cohesiveness of the village (relative internal trust) aLr. the quality VI Lthe exCtLIsi agent-elA LUVC t LLe 6gCaLtsL iIILjULL on the success or failure of agricultural extension and other development LIVILZs (table 4.4). I 3L.LC tecalibe; VI LIhe extenionUI agentL is clearcly important to bringing about grassroots development, our evidence sug- gestLs sl.latcorra-iu-r.it-y cohesion -is LLe unda,-on on wh.ich ou.d, as,r. development must be built. Dievelopmrer.t agencites haCve trai t'ionltl"y plce ICLt il..acly atten.ionI VIL technical matters and the agents who transmit them. They have devoted little or nu fiurne and resoufces to gaminh an operadoiadlly relevaint un der- standing of the social and institutional fabric of the places where they work or to training agents to erunance this local context so tnat villagers are more receptive to technical themes. The need for this sensitization and capacity building on the part of development institutions ana tneir agents became very apparent during this study. This finding suggests that an understanding of 'te level of social cohesion in eacn locale should orient the steps to be taken by the agent. The results shown in table 4.4 constitute an interesting and potentially important addition to our understanding of social capital. Cohesion is largely embedded in a village's history and culture (box 4.1). it evolves over time, as it is affected by forces and changes from inside the village and from the world beyond. These changes facilitate movement from one category to another: a cohesive village could become less cohesive, just as a socially fractured village could become more unified. Through its poli- cies and the actions of its agents, agricultural extension is one of the fac- tors that can encourage, reinforce, or damage the level of cohesiveness. The synergistic effects of a cohesive village and a dynamic agent (or any QUALiTATiVE ANALiSIS OF SOCIAL CAPITAL: MAL 93 Table 4.4 Extension Agent-Village Interaction Outcomes Vilage nomnpetent diinamic Marginal; unmotivated Socially cohesive/ Significant successes Significant impact on contact high performer in agricultural produc- group members' agricultural tioii; synergy with production. Some damage to other projects; sociai cohesion through unin- informal diffusion. tentional discrimination and Villagers talk about lack of inciusion. Negiect of developing new social principles hinders deeper planning and manage- and more integral development ment capacities, more progress. confident in themselves Example: Soke and their abilities to find solutions. Examples: Tantoudji, Kolodougou Coro Socially divided/ Contact group Nonfunctional onntnct arniinp low performer becomes internal no adoption or diffusion. unifiier; nascent Techn,ical themes -ropoe are cohesiveness, but not adapted to village needs and ve;y5 little.. d.CifusionZ CthereforeJ d511as..g Utr- except through Example: Sambaga and Tingoni infor.a w. oten's Bamanan group. Extension agent, successfully balanced membership in contact groups between quarreling Example: Kassama of the other outcome possibilities) illustrate some of the intricacies of the development process. The experiences of the six villages in our study strongly suggest that social cohesion is significantly influenced by the design and impiementa- tion of development policies. Kassama, for example, a highly fractured village, had not held an assembly in two years because each of the fac- tions wanted the meeting to be on its side of the village. The extension agent for Kassama, a competent, low-key individual, held a general assembly in front of the administrative offices that was attended by more 94 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Box 4.1 Social Cohesion in the Village of Tantoudji While Tantoudji's history has much in common with other villages in Mali, the continuing observance of its founding principles and traditions stands out. Although dramatic changes have taken place, Tantoudji's peo- ple have adapted technologies and outside interventions to meet their own priorities and values, reinforced by appropriate training and some very talented deevelopment workers. UnderstanLdirg ho-w- Tandtoudji's social capital successfully channeled outside forces should help identify best practices for use in other villages The village of Tantoudji was founded in 1898 by two family groups, the Nomoko and the Sissoko. The two patriarchs discovered the site of the future village while hunting together. The two friends decided that one family would settle to the east and the other to the west and that the village chiefdom would be passed through the Sissoko family, where it remains to this day. Over time, the two families invited four other family groups to settle with them. Villagers trace their cohesion to actions taken by their ancestors: "even before unpacking, our forefathers made sacrifices to ensure that Tantoudji would be a village of understanding and agreement," one villager explainedU. BDasic priniL1ples WrIC also darettU on atL Lthis LtLL, sUL.h as hos- pitality and respect for elders. The vision of grand hospitality extended to everyone is expressed through village saag such as "the person w,ho comes to visit you is better than you are" and "you should act as a slave to vour guest." External organizations, including the National Extension Service and various NGOs, use traditional organizations to help them conduct their activities-often overwhelming these small local institutions with new rules and priorities-or create new groups, undermining traditional insti- tutions. The strength of Tantoudji's traditional organizations is that they participate in new activities without allowing outside forces to change their priorities. For example, the National Extension Service requires that a contact group have no more than 15 members, which effectively pre- cludes m ust traditional groups. Berikadi, Tantoudji's traditional wo-- en's group, allowed the extension agent to write down the names of 15 .vomen for reeporti;ng pturposes but required al! 60 wvom-.e to be pre-sent ait meetings and demonstrations under threat of fine. There are two traditional organizations, in Tantoudji. The first. Benkadi. is an association for all married women in the village. This dynamic group collaborates with the National Extension Service and several NGOs, from which its members learn skills. Currently, Benkadi's activi- ties include a collective field and weekly street cleaning. Training sessions in children's nutrition and cooking demonstrations, soap making, cloth dyeing, and literacy are provided at regular intervals with the help of NGOs. QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: MALI 95 The pf.de ~"-om h gopiste.-vo.............ebers who were ser.t to another ............ town to be trained as midwives. Sansene Ton is the other association in Tantoudji. Its membership includes the entire village, but the active members are in their teens. Girls enter at puberty and leave when they marry (often because they move to their husbands' village). The association hires out its services during the agricultural season, wix-l, young men doing field-.vork and yoinng w^.omen bringing them a noon meal, paid for by the women's families. Money earned is used to build infrastructure in the village and to match funds for NGO interventions. In addition, Sansene Ton pays all taxes owed by the village, something never seen before by any of the researchers. It also finances village festivities, organizing an enormous event on each girl's wedding day. Honorary titles are held by the village elders. The executive president is recommended by the age group or generation currently running activi- ties and is approved by the elders before being solemnly announced to the village during an assembly. The executive president has authority over all actlve f-el-uers andiU is chouein by his- ppere -r f llUld quali.fies, such as courage, seriousness, industriousness, and strength of character. Ac,tiveha mm-bers are organizoe inFr. , w rk hlc groius 0;1o-1 - rls In A by a member who functions as a supervisor. Members who are absent during work activities or who do not follow codes of conduct are sanc- tioned. Sanctions range from a fine of 10 kola nuts to a goat or corporal punishment of up to 20 strokes of a whip. Violations for which the lowest sanction is appropriate are judged and administered by the immediate age group, the next highest by the next oldest generation, going all the way up to the elders for the most serious decision (to use the whip). With the help of one of the National Extension Service's rare rural organization subject matter specialists, in 1986 the Sansene Ton was for- mally recognized by the government as a village association, allowing it to L'ecomie a legal cooperative. RretC,UL1UI0n as a legal coopera.ive hLa many benefits, some logistical, some financial. For example, a cooperative is eligible for certain government programs and funding and can also secure a loan from a formal bank. Recognition as a village association led to the creation of a village bank, with both savings and lending compo- nents, and a store, with grain storage facilities. The creation of each of these new institutions was preceded by a villagewide workshop to explain procedures and responsibilities. The village-level managers of each institution received extensive training in accounting, stock control, and other relevant areas. The entire village is extremely proud that mem- bers of their community have gained these new skills. A synergy has been created in which the viability of the village initiatives reinforces the tradi- tional structures and moral code without which these actions could not nave taKen place to begin witn. 96 UNDERSTANDING AND MEASURING SOCIAL CAPITAL than 300 people. While this might appear to be a simple decision, the vil- lagers talkted four Qdays CaboUUL L.ith 11"1acu1lUUs" LLULr CdtU VUWeU LU hold all meetings there in the future. Through his careful actions, the agent appeareu to be r.urturing *[te early stages uo a inew Legir,umUrLuf the village. This is an essential task-and one that is far from standard ueveipIIoenlt practice. Cohesiveness can be damaged by technical messages that do not respond to the communiries' perceived needs or lhat exclude women or other groups within the village. The extension agent in Soke, a cohesive viliage, nad worked with one contact group for more than 10 years (6 of the 10 current members had been present at the founding of the group). mnere had been significant increases in agriculrural production, but mem- bers had become so familiar with the themes that they conducted biweek- iy demonstrations witnout even waiting for tne extension agent to arrive. The rest of the village had become increasingly hostile to the lack of turnover, especially since two of the three sectors of the viiiage were not represented in any of the three contact groups. Clustering three contact groups in one sector effectively excluded two ethnic groups, since each sector of the village is populated with a different ethnic group. As one nonmember commented, "It's time that the extension agent starts looking around at all the villagers that haven't benefited one bit and replaces members that have been in the group for 10 years." By disregarding a foundation of development-community cohesion-the agent and his supervisors focused on existing, successful contact groups, failing to address the impact of exclusion on the village. Development projects create programs that deal with all communities uniformly. Yet villages are clearly not all the same. Nor do they have the same innate promise for advancement. Some villages are like dry, sea- soned wood that lights easily at the touch of a match. Others are like green wood, requiring kindling, attention, and patience to become a blaze. Both types of wood have the inherent potential to fuel a campfire, but each must be treated differently. In this analogy, the age and quality of the wood represent the cohesion of the village, while the lighting tech- niques represent the extension agent (as a proxy for the larger develop- ment program). Lighting a match to green wood and walking away would result in a smoldering mess that would soon be extinguished. Similarly, creating contact groups to distribute technical information without taking into consideration the cohesiveness of the village results, in many cases, in frustration and failure. Development projects need to adapt to the different levels of cohesion and what they mean for planned activities. This requires a level of under- standing and analysis about the communities in which they work. Rather than coming in with assumptions, extension agents need to have a ques- QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: MALI 97 tioning attitude. The village is not neutral territory, but an organic and LIIanri,H,U g se. Lo reladLlons"ips. It may be argued that an agent beginning work in a new village may find it difficul., lit. 0 ilLpossiblUe, to utinAeLstand- .hes relationsips. But it should be noted how much this research team discovered in less than a w-eekd ir e a C village by focusirng on ltevels VI LofcIhIon. Exftenillo CageItL. could spend the first few months gathering such information and work- ing on easy, successful group activities, sucn as poultry vaccinatioll, which does not require a learning group. This time could also be spent visiting groups, ton (peer-group orgarizations), anca associarions to see what possibilities for collaboration exist. Villages identified as cohesive (dry wood) could move on to more tec'nnically challenging actvities, using traditional groups wherever possible. Villages identified as lacking internal cohesion (green wood) could begin literacy projects and spend more time on high-impact, large-group activities. By acknowledging the importance of cohesion, the development worker wouid not be pressured to perform the same functions in all villages. Village Organizations That Manifest and Affect Social Capital Various organizations and groups, formal and informal, exist in each of the villages studied. This section describes these.organizations and exam- ines how they could become more efficient at diffusing agricultural infor- mation. VILLAGE ORGANIZATIONS AND TRADITIONAL GROUPS. Among traditional groups in Mali, the ton is one of the most important across villages and regions. A ton is organized by age groups and allows all boys (and girls in the case of Tantoudji) of about the same age to join together for work and recreation. Relationships based on age groups tend to be very strong and last a lifetime. In the three villages considered high performers, two were using ton for agricultural extension, transforming traditional groups into engines for development. These groups were perceived by members and non- members alike to be the most dynamic actors within the village setting. Traditional groups, including ton, are often larger than the contact groups recommended by the extension service. This could explain some of the noticeably greater diffusion of technical themes when such groups are used, but the implicit trust in such organizations seems to play a large role in their success. Ton were functioning even in two of the low- performance, socially fractured villages, although they had never been contacted by an extension agent. 98 UNDERSTAN!DIN!G AND M.EASURINC SOCIAL CAPTTAL CONTACT GROUPS AND THE DIFFUSION OF INFORMATION. While the impor- tance of diffusion is discussed and emphasized within government and the National Extension Service, the reality in the field is very different. Field agents often do very little to encourage information transfer. imnlic- itly or explicitly, from contact group members to others in the village. During an initial villaqe asselmhyv thp role of the contact gropin as in C0 -- -o__0 0o - - _ - intermediary is discussed. But in some of the study villages, the assem- hly had taken nplace asmuch as 10 years ParliPr, and no onp rpmpmhbered what had been said. Extension agents are not evaluated on diffusion, and sUnpreisors serm on Pmnphaizei7p uirely fprhnical asctnPs- of lhepir lA^orlc There was little apparent difference in the method of creating contact groups between high- and low-performanrce villages: 60 percent of con- tact group members and 23 percent of noncontact group members in all vTillcages reonrt tb-bat rthecontar oronupr Iasrc rreni-A 4ater a larog acce.^*ly where the system was explained to the village. The relatively small per- centage of people (especially noneM.bers) who rem.ember the flnrr-.ation of the contact groups appears significant because this assembly is often the o^nlyr time ,Ar,n th^ r.^nt^rt grr ^ ro^le a- an intermdi'ar' ic explained to the whole community. Significantly, it appears that in most villa.g1es cn--t-cm- g-roup Menbers wereA desi-ne b- I-h vI, l a11-e e-ers rather than self-selected. Durh.g i--t V-ItVV sn. contact group L.er.O said th4at they were hesitant to diffuse new information without permission from the exten- 5aIIL CilCIUL. IVILU3L 1ILCIILUCID ICFUI LCU LILaL ULCy tilt UAL AT1IVV LIlaL UILC IlIC of the contact group was to diffuse information. Diffusion among men vvwas taking place at a noticaleU level uiy in she I.VV cohesive village that had both ton and competent dynamic extension agents. In the third coittIve villagre, ULtere wdas builu--p frustration ctflLUItg i1tany 0IIILL0Ittact group members because they could see the impact of the new technical tietiies but uiu nutt feei welcome to participate in uie sessions at wiucl their use was explained. Their sense of exclusion discouraged diffusion and may have hurt village cohesion. The training and visit system of extension is based on the theory that contact group members change as different topics are introduced. Tnis concept does not appear to have translated well in Mali, where very little turnover was observed within the contact groups studied. Because there are so few opportunities to gain new knowledge in most villages, the con- tact group becomes a permanent, often stagnant, structure. Contact group members' perceived need for the extension agent's per- mission to diffuse information, villagers' need to feel directly solicited for participation, and the lack of turnover among contact group members are aspects of communication and information transfer that have not been considered by extension projects. These problems reflect issues concern- QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: MALI 99 ing information exchange that are endemic to Malian society. Such issues can or.ly be duiscoveredu Lby payuig d,LLniLU Lu i11o information transfer, as expressed in each local context. LITERACY AND TRUST. Literacy programs often seem to promote trust and uevelupm1efInt. MVadnty iiure villagers in Colhesive, Iigii-perIurIIidlle Vil- lages (76 percent) reported the presence of a literacy program than did residents of fracrured, low-performance villages (22 percent). Tne pres- ence of newly literate people within the village is testimony to the impact of these programs. Among tnose interviewed in high-performance vii- lages, 22 percent spoke of being aware of the presence of newly literate people in the village. in low-performance viilages, the figure was only 7 percent. Literacy relates to trust in three ways. First, participants in the pro- grams spoke of increased self-confidence, or "self-trust," and the role it played in encouraging them to try new ideas and techniques. Second, the role of the National Extension Service in promoting literary increased trust in agricultural extension activities. Third, and perhaps most impor- tant, literacy has synergistic effects on other development activities, including agricultural extension. A male contact group member in Tantoudji commented on the way literacy reinforced other programs: "I am a true product of the literacy program. I have never attended even one day of formal schooling, but today I manage the paperwork for our vil- lage bank."' WOMEN'S RELATIONSHIPS AND COMMUNICATION AS AN UNRECOGNIZED INSTITUTION. Women participated in activities of the agricultural exten- sion system in four of the six villages visited, but in only two of these vil- lages did agents make any real attempt to collaborate with women and respond to their needs. Yet women appear to be the only consistent dif- fusers of information and technology. They also play a key role in tapping into and generating social capital. Women's groups emphasized the importance of increased solidarity when asked to name the strong points of contact group functioning. In the most contentious village visited, Kassama, women appeared to be outside the traditional power struggle. They were able to maintain their working and social relationships with one another in a village physically and socially divided by a road. Even in the midst of quarreling factions, women still spent time together in common work areas, such as the water source and the marsh, which provided gardening sites for the community. Kassama's proactive extension agent built trust with women through his respectful behavior and willingness to address their concerns. Although the women are not 10n0 UNDERSTANDING AND MEASURANG SOCIAL CAPITAL formally organized, diffusion of new information nevertheless takes place; farilitated by the fact that their gardening pnlts are located side hv side. In several other villages, the men gave credit for the success of sev- eral development nrniects to the iinity of the wonmen's traditional ton If harnessed, this untapped resource could enhance the effectiveness of the extepnsion ePrlTicp and strencrthen the uillagpe in .W.icl it wnrks Some National Extension Service policies seem to discourage extension ag en4t frnrn. 1A.nrl(inC l.i-) 1ATC lon In Soke, fnr exvample, a w.Men'c 0-rou In with 200 members worked on a four-hectare gardening site with no techni- cal assistar.ce from. the exte-nsion agent. W.h.en asked to explain this neglect, the agent responded that he was told by his supervisor to work with only 1 of t1he9 'I 1AT24 l.an'- groupsl in his sct-o-+r XAT]IiA, he 1l.A7ns alrady r.g. Perhaps even more alarming, some of the agricultural extension tech- i;uque are incrrAeaingr the L I, ,r chinc, n( ATA,r,.,' ( popularAr T.n..1 saV1Xa xw ss[w-- 5 zW 1[J;sVA Vl tV[w[- WWV. ss VLLZ FVXA~~~~~~~~~~~~~~~~~~~~~~~~~~~ILfl*.ltJ.'O flt f t technical theme is composting. Millet stalks are gathered after the harvest and layered into large compost pits, thus remr.oving a prtant source of fire material for women and greatly increasing the time and distance required to geal1ir adequateu. firewo. NJ tJ ... f .L L'JflJ6itO, such as improved cookstoves that require less firewood, have been intro- 16. uce.4 Another problem is the extension service's lack of awareness of womlen's aclu V I.- ar.d how they diffe in;X_ -aG --i, lg- probl tas VVIJ1L~1 ~ CLUVILa a.Lt% 1LUVVV LIMY Li11t i-ti EaIAL Vi1a16v-a Fy1ul'ilL LJit makes it difficult to see or look for alternatives. Gardening is the only teh icl - the__ enorgdfr-l.r ven when_ the ar ivolved -- LCULIIJLCII LIMUM:L t:1tLVU1d6CU IUIJ VVV11[:1t, CV1 I VVL~1= t iitty adctiULVUVU IIL other activities or where gardening is not environmentally feasible. In iw-o oi 'L.Lt villages visi.ed, .heC ex.ensionUl dagen. had[ Ltr.dLU grder.i..g contact groups even though there was no water source. The women inter- viewed voiced frustration and disappoin-timent that utey had organuzed themselves only to have the extension agent suggest inappropriate tnemes. Even more striking to the research team was tne fact that women in these villages were heavily involved in intensive animal husbandry. Significant opportunities for improvement in tnis area existed, but tne extension program provided no encouragement. RELATIONSHIP BETWEEN EXTENSION AGENTS AND CONTACr GROUP MEMBERS. Extension agents were judged to be serious, available, and patient by 80 percent of villagers in high-performance villages and by only 47 percent of villagers in low-performance villages. As might be expected, the per- ception of the agent's capability correlated with technical performance within the village. Another striking difference between high and low performers was how the villages perceived the relevance of technical themes. In the three high- performance villages, 67 percent of contact group members cited the perti- QTUAITTATTI1V ANiATCTYS IS COFSCI AT ('A DTTAT IIX A T T '.2 tint.i jA s l M. Al L. I Jill VAI S'.S...IflA XI. . L.. IV. - -ut nence of the technical messages; in low-performance villages, the figure was iust 2n percent. It appears that if the technical themes are not pertinent-that is, if they do not respond to farmers' perceived needs-a loss of trust ensues, makino iti monr tdiffifillt fnr age nts ton ciiceedrpp wif Ihftr a.:ivrit:ies Unfortunately, the extension agent does not always control the perti- nenrce of teAnirq] themesc beinog taught. A!+.hough muchl tim-.e isa spn at the yearly diagnostic meeting eliciting the agricultural problems con- froning f-arrnersc anrd their prionries for +he cornrmg year, decisions for future technical themes are made by regional authorities, without input (mmn frrorc n-_c n rbn-p&, rn,n ,,-m xrnarrc b lca_x eb c,rra nrr-jblnrn e L1Vsk ~ ~ .1 L'LlJ'._Jlf *1,T IIIAIL) VhIIts,.o Ally'. tL. O1tIl'.?t J III C-V L I1 hV the availability of technical solutions from the research station. The exten- sion agent is +,hen mifornme.e of the techLnical package for the comdi -ar. The six extension agents interviewed all expressed frustration with th- sys -Arr. Lt, a.knowled bIL. thatJ- Lo..IL posdII LItIIILCO VVwer of little interest to the communities in which they work. Trust is betrayed at .4-0 evl drn this- proess Fr-st,th 4 o.r_t 4feels that the ef'rt- L VVU ICV5IO "L4LI1IL6 LILkCI yl5)LV0. Ii , LI LC %LUIIUIJtAIL1LY ICI LltaL ULLf VILUI L to understand its problems is somehow insincere. Second, extension agents feel let A--v Iby sue rarz-; 1~I lict %AjvL t u yV ULt vU,i6dJLI.aLIUI . Another important aspect of the relationship between extension agents and tIheier contact groupsLtI is L1Lt theALtIL LU VVIULIt CaentLLb keep CpFULLIILLlLes. In high-performance villages, 87 percent of contact group members reportedu that extLers-ion agents k'ep.t i.heir dpLpI11ttIILb III LInt Villdge; Hi low-performance villages, the figure was just 53 percent. Keeping one's w-ord is Ux IIttfly IrnpofrLdltt VwIUiUi IVianaIL sociely. Thle serious andU con- scientious behavior of an agent can be an important factor determining vllage perfOrianLce. I'n divisive viliages, IL Lcant Ut pIVLLdI. CUonciLusiIons andL& LecomL.enUaLlons Our results snow that the single most important factor determining the success of an external intervention such as agricultural extension is the degree of social conesion alreaay existing in a community. Tne predispo- sition of a community's residents to attend association meetings, gather together in places of worship, buiid and maintain puDilc infrastructure- these are the things that create the fertile ground in which external inputs, sucn as agricultural extension agents and contact groups, take root. Tne study's four major findings reveal the importance of this social fabric. Social Cohesion Is the Primary Precondition jor Development External agents, or project managers, should understand the fabric of the village before they intervene. If the village has both social cohesion and a 102 UNDERSTANDING AND MEASURING SOCIAL CAPITAL qualified, dynamic external agent, there can be significant success in agri- culture and other development endeavors. in such a situation, the pres- ence of several development projects can create a synergistic effect, resulting in new planning and management capacities. Although social cohesion is difficult to engineer, it can be enhanced by community organization and local institution building, as well as by lit- eracy programs, public health programs, and other basic development assistance that can increase the capacity for mutually reinforcing social interaction. Agents must be observers as well as actors within their com- munities. Sociocultural training, including tools such as mapping, partic- ipant observation, and conversational interviewing, should be included at all levels of intervention. Women and Women's Associations Are an Important Source of Social Cohesion Women and their associations are a vital and often overlooked source of social cohesion, representing latent social capital with great potential for mobilization and development in most villages. Where this potential is recognized and capitalized on, development is more likely to become a reality. Women's participation depends greatly on the external agent's com- portment and willingness to work with women. Development institu- tions need to encourage and train agents in both these areas. The amount of time an agent spends working with women can be affected positively or negatively by the supervisor's attitude and the quality of the diagnos- tics conducted. Trust Must Be Established between External Agents and Their Agencies Trust between an external agent, such as an agricultural extension work- er, and the community is important, as is the internal trust that underlies community cohesion. However, it is also important that trust be estab- lished and reinforced between agents and the central government agency for which they work. In the absence of such trust, the agent becomes demoralized and less effective in communicating with villagers. Low trust between agents and their agencies can thus reduce trust between agents and the people they serve. Direct suDervision of both the technical and sociocultural aspects of agents' work is crucial and should be empha- sized QUALITATIVE ANALYSTS OF SOCIAL CAPITAL: MAL! 103 Beneficiary Assessment Is Useful in Investigating Social Capital The qualitative research tool known as beneficiary assessment proved to be particularly well suited to the investigation of social capital, as its reliance on qualitative research techniques enables it to probe, and reveal, the nature of neonle's relationshins with one another and with important institutions. This inductive approach allowed the team to identify impor- tant factors as they presented themselves and to adant the design of the study accordingly. 104 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Appendix 4A: Village Interview Guide Interviewer: Date: Village: Region: Interviewee: Sex: (M) (F) Working with the extension agent: Yes No If yes, method of extension (check one that applies): (a) Contact group , membership (in numbers): (M) - (F) (D) Association - membership (in numbers): (M) -(F) : (c) Individual family . membership (in numbers): (M) (F) 1. Assessment of extension agent (a) Professionalism (keeping appointments, knowledge of technical themes, pertinence of technical themes, availability, technical support, etc.). (b) Involvement in village social life (behavior, attitude, reputation, integration into village life). 2. Assessment of village organization specialist (answerfor associations only) (a) Professionalism (availability, technical support, management training, etc.). (b) Involvement in village social life. QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: MALI 105 3. Relationship between women and extension agents (extension possibilities, avaiiabiiity, acceptance by husbands, etc.) 4. Relationship among contact group members (or association members) (a) Formation of the contact group (or association). How formed? By whom? Inclusiveness (gender/ income/ ethnicity). (b) Personal ties among members (friendship, neighbors, type of farming, family). 5. Evaluation of contact group (strong points and weak points) 6. Relationship between the contact group (or association) and the rest of the village population (a) Diffusion of technical themes between villagers working with extension agents and villagers who are not. (b) Motivation for diffusion of technical themes. (c) Reasons for lack of diffusion of technical themes. (d) Suggestions for improving diffusion. 7. Impact of the contact group (or association) on the rest of the population (a) Impact of extension on the village (agricultural production, revenue, health, literacy). (b) Impact on village attitudes (receptivity, social cohesion). 8. Technical themes that inspire or reinforce trust (a) Agricultural themes (insecticide treatment for seeds, composting, transplanting in line, etc.). (b) Animal husbandry themes (improved feed, improved pastures, vaccinations, etc.). 106 UNDERSTANDING AND MEASURING SOCIAL CAPITAL (c) Environmental themes. (d) Rural organization (management, marketing). 9. Suggestions for improving the level of trust (a) Between group members and the extension agent. (b) Between group members and the village organization specialist. 10. Observations of the interviewer Appendix 4B: Description of Villages Kassama, Kayes region! population 1,300 Low performer. Extremely troubled village with almost no village-level social cohesion. Very talented, patient extension agent. Three contact groups, one male (nonfunctional). one female (using traditional women's working group), and one mixed. Tantoudji, Kayes region, population 360 High performer. Very unified village, despite many different ethnic oriinh1 )vn:mir Pytpncinn nipnt Cn1v vi11aap hZith q villnawp nranni7- tion specialist (also very dynamic). Two newly literate men and two newAlyAT litpra xAmmen oive wppely eci-c' in rpedingn and writfing for the community. Four contact groups (two male, two female). Female contact groupsr~ use tr2aditiorna l.^^mr' asso.ations.W1 Sambhana Knaye region poulti n48 Low performer. So.e fintern.al colwsion, althoughf it is hIan.pered by severe water shortages and an ineffective, negligent extension agent. St-ong trad.4;o4nal group have not. been used fogiclualetr.in , L iLX.a.tUi Lal ,. 5lJiFJi LLV IM t) -W IJCCI L UD13CL Ul 5 L LI1.11C C 0 1L Huge potential in women's aviculture is not being exploited. Three con- tact groups, or-Le male, one 1feLmaIe, one ixed. --ly i- IuILLUUILaI. QUTALTTATIVP AMAIYS!S OF SOCIAL CAP!TAL: MATT 107 Soke, Segou region, population 1,700 High performer; socially cohesive village. Extension agent of average competence but no initiative. Three contact groups, all male. All contact group members live in one sector of village, that of the village chief. Tingoni Bamanan, Segou region, population 385 Low nerformer. Some tension between sectors of village. Below-average extension agent who was previously forest inspector. Two contact grouns. both male. hoth harelv fitnmtional Dvnamic vyouth groun not being used for extension. Huge potential in women's small animal hus- bandrv not being etxloited= Knlndtnonii Corn Ro6unu reorinn nnnulatinn 661 T-Tighrl, performe-r IHigh,ly l rnb0chivye vililage A^TithlM rroivated-A agoen w-Tn is somewhat constrained by lack of training in agriculture (his background is in anirr.a Ihusbandry). Pixve contact groups, fo1i mn.e o ne ferr.lale Women's group frustrated by lack of extension themes. Fielld researclJ for LItis study was conducted by MaLIradou Ca.araa, .Chick Kamate, Catherine Reid, and Maimouna Sow Sangare. 1. This finding was echoed in Anirudh Krishna and Norman Uphoff's 1999 311-4Y 1I,?, IT - , a--. -- - -bU11L OL--I -I t ....yLa. -l 'v-U F LC LUa.OL .LLIIO stuy -Mapping -an Mesuin ScalCaitl ACocpula,Er.icl Study of Collective Action for Conserving and Developing Watersheds in IJXaJtLIL IRI, Ir.dia, Socal C-aF ital UL1I.V;a VVUrItr .I Paper 13.. VVWolIU BanI, Social Development Department, Washington, D.C. 5 Qualitative Analysis of Social Capital: The Case of Community Development in Coal Mining Areas in Orissa, India Enrique Pantoja The rediscovery of social capital has seemingly provided a missing and powerfui expianation of how development processes work and how they may be used to strengthen related policy, programs, and projects.1 The potential contribution of social capitai to development appears to be immense, as corroborated by rapidly growing empirical knowledge (Putnam 1998). This potential contribution acquires even greater sig- nificance under current efforts to make development more equitable and democratic (Stiglitz 1998). international organizations, national governments, and nongovernmental organizations (NGOs) have embraced social capital as an important form of capital, indispensable to making development possible and sustainable. Yet the successful utilization of social capital may not be possible without understanding its potential negative as well as positive effects, and without promot- ing enabling environments that facilitate participatory, community-dri- ven efforts. The study described in this chapter explores the concept of social cap- ital through the analysis of two coal mining areas in the state of Orissa, India. The qualitative analysis is of community development processes structured within the broader political economy of poverty. One of the main goals in poverty alleviation is the provision of access to sustainable income sources and reliable and safe basic services. To help achieve this goal, it is assumed that poverjty alleviation strategies should support existing forms of social capital land promote the formation of new ones. The stuidy described here approaches this assumption critically by look- ing at the various forms, dimensions, and effects of social capital. It 108 QUAMITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 109 accepts the notion that social capital is a common resource (even if priva- tizable in certain instances) that can nrovide access to other resources. But the study also considers that under scarcity conditions, such as those nrevalent in developing economies and narficularlv among the poor; social capital can become an integral part of the structures of constraint created by gender. cr1a.s ethnicitvy and in the rase of India; religion and caste. For these reasons, the study pays close attention to issues of access to social capital resources and to control of resourres once such access has been facilitated. Furthermore, the study analyzes the role social capital nlavs in thp tnri-n lrpcr nf farilitation nr conntraint that chrrte1-ri7e a nar- ticular society at the local level. The unique "enclave developmnent" characteristics of the coal mining areas makes them an ideal research setting. Coal production activities in Tn,dia tak-e plre in remAfo areas, ^ATIorh recnl r]mining hasc beornrea the ron- ter of economic life while changing the local labor market structure, the -cina! fabric, -nA 4ho nat-ua ein irnor n After Ind1 nationalized rnost coal mines in the 1970s, Coal India Limited was established, and along v^With its subDsidiar.e'sl -loon became th.e rnost -run^a nsiuinlan developmental actor in these remote areas. The mining areas have b1ecomre de factdo coMrpany enclaves as dire-ct 6governMlent intenrvention haas diminished and the presence of the state has been rendered almost invis- ile. C-oal LTdia, ;L II- _t II cTOU l5y acquired .it; develop- ment responsibilities and become the repository of peoples' expectations regardn A mlo..l genraio --- -eric pro---n L "Utl II C:1 tFItV YIUy tIfJLL 6CiIMIaLlUt tfll CLI VIL }1U DI %_ FkV 1311)1. Interest in the coal mining areas increased when the World Bank fimarlced -LVCO-t--lCtf L tLo ILF.pIrv Llte plritblilty of 25 of '-oal LndILa' opencast mining operations (the Coal Sector Rehabilitation Project, or CSRP)' anA the Ao..ayscaaiyt eal Wi_. social and env-ror.len ...l 7 a LLI ftfELSJ LJ-.JI LY a LCtFaIL y LUJ LI~ V I fLi ,u tI aI U C:I L VI UIkJ Ut MI t- tal issues (the Coal Sector Environmental and Social Mitigation Project, or LEIMVPA ). ItnV-.e1tsILin ItL LIL3C LILC, III illay ULIoteCr aLcU URh Varius coalfields, were likely to affect (positively or adversely) traditional sUU1rce Ut 1IlLULIML CllU t:ltIJUyIItt:[LL UpFUUlLUHIlfUt, sULdal nltvWUor.ls anllU structures, and the form and quality of social capital. The ESMP did not entail specific activities related to socUial capital, but it did induut socUidl mitigation measures that affected the relations between the coal compa- ny and the coLmmlunities (vertical articuLaLions), antd withintll te LU111nU- nities themselves (horizontal linkages).2 The CSRP and ESMP are explained in more uetail later. …% - -1 IL1 A Database anu iviethuuuiugy This exploratory study identifies hypotheses and indicators that wii heip future researchers investigate more rigorously the concept of social capi- 110 UNDERSTANDING AND MEASURING SOCIAL CAPITAL tal within the framework of community-based development and better distinguish, as Portes (1998) advises, the concept itself from its alleged effects. Specifically, the methodology is responsive to field conditions, including (1) the need to avoid interfering with project implementation activi- ties, since the two Bank-financed projects were still being implemented, and many of their social mitigation measures were being tried by Coal India for the first time; (2) the need to obtain and maintain the collaboration of Coal India and the concerned subsidiary in doing fieldwork and deploying members of the local team; and given these requirements, (3) the need to find "safe" study areas with low levels of conflict between the mine and the villagers, where proj-ect implementation was advancing relatively unen- cumbered. The study team identified 13 mines that were receiving World Bank financing and 30 that were not as eligible for the study. The following cri- teria were used to determine eligibility: (1) the mine affecting the study area had to be an opencast operation, since opencast operations are more likely than underground mines to affect the social and natural environ- ment; (2) one of the mining areas had to come from the set included under the ESMP, while another one, of similar age, had to be from the rest of opencast mines; and (3) the mines had to have created the need for resettlement, rehabilitation, and community development.3 After considering particular factors such as geography, regulatory environment, and social, political, and economic context, the Talcher and lb Valley Coalfields located in the state of Orissa and managed by the subsidiary Mahanadi Coalfields Limited were identified as the most suit- able areas in which to undertake the study (see man). In consultation with Coal India and Mahanadi Coalfields representatives, one mining area was finally selected in each coalfield: Samaleswari in the Th Valley was the study area with a mine receiving bank financing, and Kalinga in Talcher was tihe area with a mine withouiit hank financing. The study areas' boundaries followed the definition given in the ESMP as the mine's 7Anfl nf infliipnrt' fnr rnmmiinif-, tJPxl7lnnmpnt i,hirh inrhlidA. ti1 CPf,H-1_ .- __ -- .._. -- ----r---- -- ---- ments within one kilometer of the boundary of the mine holding (World Ba.nk 1996). A major advantage of undertaking the study in Samaleswari was the Ovi cl-ni, ofl cmrioecrnr, noi,' baseline data*,scial,, and 7 evror.rr.ntal reports, and project progress reports prepared for the Bank-financed proj- ects. -0fst of th ormation for Samaleswari was collected fbhr.ug sur veys in 1994 and 1995 during project preparation, but an exercise to upd4atete4- rla^o n rJc-aetdpol was conduct4ed in 1998. LL ,ttCIZL ~i iIi Lei) CSi hli` i. C JFlmF WV -X LAin -Ut IA ,t -L~. C U.' Not nearly as much information was available for Kalinga. fol supplement the exisreng detae n: ft t u bs following activities were undertaken: QUALiTATIVE ANALYSIS OF SOCiAL CAPiTAL: ORiSSA, INDiA ill |R. TblSmap_XoptosL~byllw (+v¢t 5 _ _ >o IHAR ' "H % ) WEST MN,p --zL- BIHAR UNDARCARi_ BENGAL _BR Nve el f v I f I' .' .B~p.. f Il f,h olT.iaou i. MAYURBHANJ X X \v5(~Q Somb0lP9 (¸D*4EfGAR -/ .--KENDUJHARG:RAR,( J/ /' BARGARN N0SAMBALPUR W, r t MADHYA e C(N. G A U -r . _ I PRADESH / ~ y''' | ' > t AlAn g {___~~~~~~~~~~~~~~~E D.,E . SALANGIR I i'. ,NAYAGARH ,* 9 > ,0rov .; PHULABANI tF W JEI E LAG§TSINGHPUR t tveS,sSo ) KAIAHANDI 2 \9_ _ hI 0 9 M K RDN r-o-'E- ~Gk4JAM 7' < ~ ~ ~~)INDIA ~ ~ \p~ RA~AGADA¶~i )IS j;:." s{ttasT 1' | o 8 >.> 6rr IN A K 9 KORAPUT . V j \ , . SATE CAPITAL S , 015070' , ,1 , 0.tM BOU_NDARES Bengal SlAB! BEvNDA21ES / Y~~~~~~~~~~o. . ~ ~ ~ ~ plp1 . r.2 o.11 10ER _ } MpikonSg >v #_ t ~ ~ ~~~ g DT 20 "D w08010 120 1°I 0 uaoEes (2MAIKANOGIRf' J 150 rM_ AI'INDHA PRAflFSN 9 0 4 0 8 0 0 4 6 IOE0 * A household survey in each of the study areas to update the socioeconom- ic data and obtain a first insight into the social capital of the commu- nities; * Focus group sessions; * Unstructured interviews with relevant stakeholders, including officials of Coal india and Mahanadi Coalfields, government officials, and NGO and community representatives; * A stakeholders workshop (in Samaleswari only). Many of these activities were developed in collaboration with a team from the local consulting firm Operations Research Group, who periodi- cally visited the study areas between July 1998 and February 1999 to fol- low everyday events in the communities and maintain contact with the NGO operating in Samaleswari. As a member of.the World Bank's super- vision team for the ESMP, the author visited Orissa on multiple occasions in 1997 and 1998, where he met with Coal India and Mahanadi Coalfields officials, government officials, and community members. 112 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Context: Overview of Orissa Orissa, located along the Bay of Bengal, occupies 156,000 square kilome- ters of Tndia's territorv. Despite having huge mineral resources, Orissa is one of the poorest, most rural, and least developed states (Crook 1997, Repetto 1994)= Eighty nerrent of the state's nonulation denends on the agricultural sector, which has one of the lowest yields in the country. More than one-fourth of the 34 2 million people living in Orissa in 1996 were officially designated as scheduled tribes, and about 15 percent were consi-iprpel srhi-edii1lp carstep The majority of the scheduled tribes arp poor and illiterate (Fernandes, Menon, and Viegas 1988). Significantly, most of Orissa's coal belt as well as other sources of raw materials, cor- responds to its forest and tribal regions. TTnder the ronstitiltion of 19Q50 thoep gronup hblongingr to thp lowpr levels of the caste system and persons of tribal origin were classified into separate cerbduiie.c The0 cAhpiii1pJ rd-1ca andti t.bes Aere o.fficially iden- tified as the most underprivileged groups in the country (World Bank 1996). Sorr.e por groiupc that Aid not fall into either of these c-egor.ies were classified as "backward," and the state assumed responsibility for ,Adv,-icr4ng tlhe ornnonhir co-1i2l .nA oAirai7nr al ,*,olfnre o f ihe hnr- ward" classes, including the scheduled castes and tribes. In contrast, the general castes were considered "socioeconomically advar.ced" and no special provision was made for them. Profile of the Study Areas The coal mines in the lb Valley and Talcher fall under management of MahcuanadiL- CoalfielUds LAimited, wVILULVCIhC, opeate 21LUneS andu FrectLs across these two areas. Samaleswari, Ib Valley Coalfield Coal mining operations in the lb Valley area began more than 50 years ago and accelerated wihn thne natonalization or the rriiunes in ute 1970s ianu the implementation of opencast projects in the 1980s. In the Ib Valley, Maananadi Coalfeids runs five underground mines an rfive opencast mines. Three of the opencast mines, including Samaleswari, received World Bank financing (World Bank 1996, 1997). The Samaieswari open- cast mine is located in Jharsuguda revenue district, between Sambalpur and Sundargarh revenue districts. Actual coal mining in Samaieswari started in 1993, although preparation work began in 1987. The mine has an estimated project life of about 293 years. QUALiTATiVE ANALYSIS OF SOCiAL CAPITAL: ORiSSA, INDiA 113 The study area in Samaleswari includes four villages, which in turn col..t.fl1 seve hLabitAat.ons: tflucli, Ka,apal-11 TdA-all;, LaTjkI-- Mundapara, Orampara, and Sukhpara.4 Total population within the stLud.y area is ayoi. i7atey 2,71 (orI A 70 households). AboutAl #-A 6.7-perc of the total population belongs to scheduled tribes, while 12.7 percent b-elongs to scheduled ---es The rest - eon to otherA bacwar ca-e UCUL LV MLJLCULIUICU. %..aDLC3. I ILC IC3L LICIUVIL6 LVU LLIII ULaL~VVQ.lL ~"L..I30 (35.4 percent), who are the second largest group, and to general, upper castes 5.2 pe1rc1tL). PILLUIUM6 LU Lito 19947'± DVL1UJLkU1[UI11L baselinLe sUrVCy, an estimated 1,655 persons (398 families) would be affected by the min- UL, UUCldLlUltZ UUt LU ldlIU dLLjUlZlLlUll, U1 WVYIUtil ZL. We1r LnUal1,' (Vivekananda Palli Agragami Seva Pratishan [VPASP] 1994a, 1994b). The total numiiber o lentitiele proJect-affected persons in DamIales-W-ali -W-as eWsti- mated at 986.5 Land acquisition had been finalized along with payment or compensation, except in cases where the ownersliup of the land was under legal dispute or tenure was not clear. Ironically, many families who nad to resettle because or tne mine project had been resettiec into tnis area after their native places were submerged during construction of the Hirakud Dam. Kalinga, Talcher Coalfield The Taicher-Angui area is industrializing rapidly. Several large industries in the area, such as the National Aluminum Company and the National Thermal Power Corporation, and a good number of medium and small industries, depend on the coal company for their power. Three opencast mines, also under the management of Mahanadi Coalfields, are receiving Bank financing here. The selected study area, Kalinga, is adjacent to one of them, Bharatpur. The first coal mine in the area opened, although on a very small scale, in 1926, but it was not until 1960 that coal mining began to intensify. The Kalinga Opencast Project, located in Angul District, had been started about 8 years before the study. The expected life of the mine was 27 years. In Kalinga, eight habitations are included in the study: Bramhanbahal, Danra, Kalam Chhuin, Majhika, Nakeipasi, Natada, Nathgaon, and Solada. The study area has a population of 15,095 (or 2,351 households). The majority of the population (65.5 percent) belongs to the other back- ward castes. Scheduled tribes are second (15.2 percent), followed by gen- eral castes (9.7 percent) and scheduled castes (9.6 percent). Bramhanbahal was scheduled to be resettled completely, while Majhika and Solada were scheduled to be partially relocated. The other five habitations were not directly affected by the mine's land acquisitions but were receiving com- munity development assistance. 14 TTNDERSTANDING A NTrD MEASUR A TING SOCIAL CAPITA I Social Organization in the Study Areas In 1995, when the area was surveyed in preparation for developing the ES MP t-hp eCnnnmvy of Samaleswari_ although still predominantly agrar- ian, was changing rapidly (ORG 1995). Over time, the mining industry had created many new acirviies inthe area, trannfArrninr what had been a subsistence economy into a market-driven one. The coal company had also acquired a significant amount of land, and mtanxy inhabitants had lost their agricultural plots. More than half of the male population had found emrln"nl07mnt,in , n i-1h- rnets in 1 99'5 in- cn,Am h ome l h'ndeco A ni1r lhnnroirc in mine-related activities, but others had been left without access to income- generation opncrti OCRG P'cs initial cisirvey alsc indAirnrcat tha inequahty was increasing because those working at the mines had steady incorr,es-4 ard o scriy wie h livelihood of th-s left w^vithout harid and alternative income-generation opportunities was deteriorating. A ch-rge in th- le-adesi pa-4tte. m was also observed, ass the Mnah-anadi Coalfields employees began to take power away from the traditional leadershlip. The caste hierarchy has historically been strong in Orissa, and it is t1A.lJug thLis .--taly that v-ilage=level it.fl0tC 1eaL1.JL.n ar.d -sciabil ity must be viewed (Lerche 1991). With mining development and agri- cultural trans &on..a..A-n- -h -td aras --v.- social ordrsan LLILU IC L1aL tJ11ILLcIuUL, ULtt MIL UYCLICIZ ICV1J 11vaiL16 ZIU Cli Vi.L%Az10 CULLL rules for social exchange have certainly changed. But a villager's position inL Lith soULcia hLierarLchy LU1jLULs LU be UteLt:1LUitzU la1rgly by Lith satmLt factors that have long defined it: power, wealth, and ritual prestige. The sociuoellgiouub nte.w-or sLL1UU11t6 Lil.e Vlldaes' Lcdte-baseU UdivisUil of labor has proved to be highly impervious to change, and many of the changes in interc:aste relations nave occurred within this system's frame- work. risroricaliy, the degree of ritual purity or pollution or eacn caste has been related to specific occupations, and specific exchanges and services 1 . s . { ~~~.1 . I .. , . 7 - -. , .. , . I nave been derived from tnis relanonsnip.7 Tne punty-pollunon hierarcny is also expressed in a number of norms regarding what a given person is allowed to do and not to do (such as eating rules and whom one may associate with). Changes in occupations in the study areas, especially the decline in and even extinction of many traditional caste occupations, have undermined caste differentiation and accelerated social change. With the loss of land, traditional farm-labor arrangements have declined, albeit more in Samaleswari than in Kalinga, where agriculture is still more relevant to the local economy. Long-term relationships between employers and laborers have decreased in both study areas, while more casual, impersonalized contracting and subcontracting in various sectors has increased. c'I,I,,A I AT XT AIVC1CC7 rQ TTrDTA ~TC TNDIlA QUAfLITATI VE fLNAL1 OF SOCIAL CAPITAL: ORISSA, INDIA 115 Despite these changes, two main divisions established by the ritual hiirirrhv are still remqirkhlhv strong- thpe dividing lineh,btween Brahmin and everyone else, and the dividing line between clean castes and pol- luted castes ("untourhables," or Har ins) (Neale 1990). The hierarchy establishes the Brahmins at the top, followed by the clean general castes, then fhe unclean general castes, and at the bo'ttom. the peripheral castes, including the Harijans. Although the unclean general castes are consid- eredt a pzrt of the1 WdInOrPe thep peri"phra! rcates nar mnrojnna7qzed spa.al]X7 and socially, living in hamlets of their own located at some distance from Coal 1ndin T imitod Coal India ULimted, established in 1075, 1i, in Utermsof -4 Luts and employment, the world's largest coal company and India's largest public sectuoLr L nterise.8 It has abou*t. A90 LmiLns unAer its ---l-L--LaL, of VwViL approximately 140 are opencast operations. Structured as a holding com- panly, Covall sTus A a;- 1- 1wussscr vlss Fa v; ca autonomy. Like other large, complex, and hierarchical organizations, Coal India iD characterizeLd by i..Lper V1Vou1sns LU chItdtIL6 anIU UoVbUs challenges regarding information flows and delegation of authority. A subsidiary, f'or instar.cte, hasl Lthree layersb ofI IItt.a6UatgeL.e, Lcorpora,e ltevel, area level, and mine level. Mahanadi Coalfields, which became a sub- blUldly lliLy IL 1774 Ilidlidgt:b Illlteb 11 llVC ditdSb. Coal India is a coal company, so, not surprisingly, its management and oflicers see LIte proJJucLLon ofl UCoald athCID prirrliCly r1L.issio. IThe LUIIrpCa- ny cannot ignore, however, the social and environmental issues intrinsic to coal .l,irg In_ I_.e ---. rur. -3-.e corprybcre dsl dr,fe LULUI 1LUh1tg. lit UL-Lt WUI%, lUII, UtC LUiltpdlly L'tCLU1Il-b Liubtfly IUeLItlIeU with the particular regions where its operations take place, to the extent that the rmining areas thenmselves, including the surrounding villages, become implicitly company grounds, at least in the eyes of state and local governrnent officials and tmose living within their confines.- ine rutures of most of the individuals living in the mining areas are closely tied to the mining operations, tnrough direct and indirect employment or through the ways their lives are changed by the new social, economic, and envi- ronmental conditions created by such operations. Paradoxically, the coal company's capacity to address the demands of the surrounding communities, iet alone to fuifiii its promises regarding community development, resettlement, and rehabilitation, is limited. In turn, a relationship of dependency and a sense of distrust tend to devel- op, often simultaneously, between the affected communities and the coal company. This relationship is complicated by the fact that everyone knows the coal company will eventually leave the area, returning the 116 UNDERSTANDING AND MEASURING SOCIAL CAPITAL land to the state government once mining operations have been complet- ed and the area has been restored.'0 The establishment of de facto company domain over a region starts with the decision to acquire land in areas with commercially exploitable coal resources. Coal India and its subsidiaries acquire the larnd for mining under the Land Acquisitions Act of 1904 and the Coal Bearing Areas Act of 1957.11 Holders of the land must give it up to the coal company, and the Land Acouisition Act requires that the company compensate all landowners. The companv's interaction with the local norulation thus starts years before the actual mining operations begin. Local labor is usually recruit- ed early on bv the companv, mainly for semiskilled and unskilled jobs. Further along in the process, the state and local governments tend to scale down their level of intervention in the social and economic develonment in the mining areas. With time, the centrality of the mine economy increase- A more defined market economy emerges as avriculhllre and other traditional forms of livelihood decline while more and more local people start working at the mine. The comnanv also builds new farilities and improves the existing infrastructure to facilitate the movements of vooAsd in and roit oif the mining areas while rqrrvino onit rnrnrniinith development activities in villages near the mines. Expectations and a sense of entitlemrent sim.ultaneously increase among local communities. Since most of these areas, at least initially, are isolated, a clear case of "ncia s developnrn ent" take place. The World Rannk-Financed Prniorts I I~. AJ .Jt.t.J J I -Uff I C tflf1 - . -1 1J .L1 - S - -.JlI CII ItI CA 101 Ll1 .-JI The Coal Sector Rehabilitation Project, or CRUZP air.ed, atasit -gCa India to make coal production financially sound and more efficient. The project, wit-h ar. esg,r.ate costA -of TUS$1. bilo (of -whiAh the- 1A^llA Bar-1 was financing US$500 million), was approved in September 1997 and was expectEed toclsei J1_ - . T 0z Seiicly through Suhe in-esll.en COl cA ~ L U. L -I.) S CU IlL JWLL 1-UUI). LJ~IIa 1y,LUAULLIt ULIC II1VI.DLIILC:ILL %LUIILL ponent, the Bank loan was financing the cost of a large fleet of heavy earULh-rLUVi1n equiLpLetL LU m1LUC11IULtz VI 25 opencast lcLaLs . Tnl addition, the CSRP also supported coal sector reform and the restructur- Uin VIl %.Ud1 ftld UIUUroli LtL.ulL-al ass-istarLce artdU Lrainlg. The Coal Sector Environmental and Social Mitigation Project, or ESMP, was approveu Hi MyV19d96 and itS aUuIviIIes were tO irusn by June 2001.12 A complement to the CSRP, this project supported Coal India's efforts to make its production more environmentally and socially sus- tainable, while ensuring that any possible negative effects of coal mining expansion would be alleviated.3 lTotal project cost was estimated at US$84 million, of which the World Bank was financing US$63.3 million. QUALITATIVE ANALYSIS OF SOCIAL CAPiTAL: ORISSA, iNDiA 117 An important change induced by the project, and key for the study, was Cnnl TndAi'c nAndfion of a new - -4 m nt a lA.on policy and the revision of its community development guidelines, which all sub'sidi-aries, w^ould elventumllY have to foIl-ov. SclJCd -4--.tflf-L L.a- sures included individual assistance through specific rehabilitation and compensati;on packages. In gerieral, dependi on the Uc Catego.y a project- ed-affected person was in (landowner, landless, or member of a tribe us;_ --- co.o rsuce)oal T- .da would offer assistance for estab lishing nonfarm employment through the provision of infrastructure, sm.all contr acts, and4 establ-I.S.en of& copeaivs wol a_h ~1L11 Lr LLL~ aiu ~LaUi1.1L11MC1L (L LUUy1L1 ,vvLlUU FJtY tlt: replacements costs of homesteads, where applicable, and provide an ILCIiLaauVC: IRUDIIh6D E: aiLU wvuiu. UL1E1 E:litFl`-Y1I LUIL InIILtC lII1tlitf VIttfII feasible. Moreover, Coal India would offer village-level assistance AhroUghI coLrr.nuni-y uevelop.i-lent plans 'kLnown as iuigeiious peoples development plans) that included provision of physical and social infra- struc-ture, traiing for self-employi-nent, and support for community activities such as watershed management and reforestation. %.Jdol lind s niew resUeitltlemet dalU renaoiitrahon poiicy and commum- ty development guidelines were likely to have several effects on the social capital resources or the mining areas with vvorld Bank-financed mines, including Samaleswari. These effects included the establishment of new cadres of environmental officers and community development, resettlement, and rehabilitation officers; the hiring of an NGO to facilitate community participation in and implementation of social mitgation measures; enhanced consultation with people affected by the project; and creation of a community development council and viiiage working groups to identify and carry out development activities. Previously, com- munity development was ad hoc and the assets provided by the compa- ny to the beneficiaries were absolutely free. Under the new guidelines, the coal company was requiring payment from the communities in the form of mandatory contributions in kind or labor. The most important changes related to the resettlement and rehabilitation policy, however, counted loss of economic assets as well as loss of land in the compensa- tion process; expanded the definition of a project-affected person to cover not only landowners, but any adult whose livelihood was affected, including the landless, and entitled all eligible adults to compensation; and replaced jobs in the mines as the major form of rehabilitation with multiple rehabilitation options. The mines receiving World Bank financing were considered to be a testing ground for the new social and environmental measures that the coal company was planning to apply eventually to all mines. Problems with implementation of the new measures arose immediately. This was not the first time that Coal India had changed its practices for dealing 118 UNDERSTANDING AND MEASURING SOCIAL CAPITAL with the effects its mines had on the surrounding populations, and peo- pie were understandably confused over what to expect. Ivioreover, con- fusion and potential for conflict was destined to arise because adjacent communities were being treated differently even though they were expe- riencing similar conditions. Analysis The study of social movements and collective action has long required social scientists to pay attention to issues of cooperation, social cohesion, and conflict. In general, however, sociologists have tended to oversocial- ize individuals" actions. Conversely, economics has tended to underso- cialize these actions (Granovetter 1985). The works on social capital by Bourdieu (1986), Coleman (1988,1990) and Putnam (1993,1995), among others, have provided new conceptual and analytical elements to balance the two. Their work provides valuable new and relevant insights to sup- plement narrowly constructed economic models whose limitations in explaining social and political behavior of individuals and groups have long been evident. The work on social capital contributes significantly to the effort to overcome these limitations by integrating nonmarket factors into the analysis of political and economic life.14 The coal mining areas in Orissa provide an excellent opportunity to explore the concept of social capital. The study areas, Samaleswari and Kalinga, as part of de facto "company enclaves," represent a scaled-down model of the working of social capital in Indian society as a whole. Although it would be misleading to extrapolate from the particular find- ings in the study areas to the country level, the relative isolation of the study areas allows researchers to explore certain dynamics more clearly and to identify issues critical to a better understanding of the nature and effects of social capital in general and of social capital at the community level in particular. Specifically, the study focuses on the vertical articulations and nonar- ticulations between the coal company and the concerned communities, understood as a replication of state-civil society relations. The study also examines horizontal interactions among community members, that is, the local civil society. As mentioned above, the focus is on communi- ty-based development, but within the broader framework of the politi- cal economy of poverty. The analysis is developed as a virtual matrix that takes a particular form of social capital and looks at its width and depth, at the way the issues of power and politics weigh in to give it a relative use value depending on who is assessing this value, and at its relevance for community-based development. The virtual matrix is thus multilevel and multidimensional, targeting the structures of facilitation QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 119 and constraint that affect access to resources (including social capital resources), control of access to resources, and control of the resources once access is obtained or granted. Through the development of this ma_nt~rix, the Inmain -assumptions of thle study are itet-ively tsted. against the findings. Forms of Social Capital All forms of social capital can affect collective action, governance, and e IonoL..icl pcrfO.uhnnl.ce itn nMtany vvays, at- 11 oa tLhI tave a role LU play in the creation and maintenance of generalized trust as well. These forms vary fromL. one soi-oraia6 a --------g to-- the next Cua-;s 199).o They- vayIUl ILC S0%_lU-C7Tra1LlLaLlUltaI 0CLLUL6 LU LILt: lItAL kl.ZCtLIl~ 1770). 1lIM) do not exist in isolation, many do not have clear or real boundaries, and mn ar eM- , Mleul in_ L'iL e 1 _r __ £-1 _ _ hS1 _ iilaity airs t:lILU~IUtUt: 111 ULllI LUlIlt UVI ULIdL Ld.)ILCd UI d1C 1ltUfbbdiy inputs to or outputs of other forms of social capital. Following Harriss 2rv. f:.1 flnn7\ i 1Alno\ 1, /Ilnfl I nn.\ __ - - and 1e Ierzlo (1977/), P-oLteb (1770), J[ULIamII (1773, 1998), anid Rose, Mishler, and Haerpfer (1997), the forms of social capital considered in this study are: ramily und kinship connecrions, ncluiumg the single nousenola, tne extended family, and the clan, based on "strong" ties of blood and affinity. An important factor is tnat family and kin reiationships are created mainly by birth, not by choice. v'vider social networks, or "associational life," including networks of indi- viduals, groups, and organizations that link individuals from different families or groups in common activities for various purposes. This is the form of social capital closest to Putnam's 1993 definition of social capital in terms of "networks of civic engagement" or "local associa- tions." This form of social capital covers a full range of formal and informal horizontal arrangements. * Cross-sectional linkages, or "networks of networks," including the net- works that link organizations from various sectors of society (for example, NGOs, grassroots organizations, government agencies, pri- vate firms) and allow them to combine resources and different types of knowledge to find solutions to complex problems. Through these net- works, public-private cross-sectoral linkages and mutually supportive and complementary relations are established. This form of capital pro- vides the articulations between horizontal and vertical associations and organizations. * Political capital, including the norms and networks shaping relations between civil society and the state, thereby allowing a society to medi- ate conflict by effectively responding to multiple citizen demands. Political capital is related to informal institutional arrangements that 120 UNDERSTANDING AND MEASURING SOCIAL CAPITAL may result in clientelism, rent-seeking, and exclusion, or in effective representation, accountabiliiy, and participation. * Institutional and policy framework, including formal rules and norms (constitutions, laws, regulation, policies) tnat regulate public lire. Generally identified as macro-level social capital, this form has a dou- bie nature, since it may induce tne creation of otner forms of social cap- ital, while it constitutes in itself a resource that facilitates coordinated actions. Moreover, existence of a consistent institutional framework is necessary to generate and strengthen generalized social trust. * Social norms and values, including widely shared cultural beliefs and the effects such beliefs have on the functioning of society at large. Norms and values support other forms of sociai capitai as weii as rep- resenting the most general form of social capital. Institutional and Policy Framework: Tne Extent of Generalized Social Trust Hatti and Heimann (1992, p. 62) explain that "one of the most important concepts in the Indian value system is the one of trust." Accordingly, trust or the lack of trust is a critical component in evaluations ot relationships among Indian people. Descriptions of family members, friends, or strangers always include an assessment of the specific individual's trust- worthiness, which according to Hatti and Heimann is usually assumed to be lacking. Historical and contemporary forces both have exacerbated ethnic, class, gender, and familial tensions in India, which, combined with the scarcity of resources, have made access, distribution, and use of existing social capital highly contested and social capital resources them- selves fragmented (Morris 1998). In India, communalism has translated into conflict and violence across religion-based communities, such as Muslims, Hindus, and Sikhs. Generalized social trust has been weakened by the deeply ambivalent relationship between secularism and religion (Basu and Subrahmanyam 1996). A high level of "abstract trust" has resulted in strengthened identi- ties based on mechanisms of inclusion and exclusion. Diversity of voices and interests has resulted, in India, in a pluralist democracy that with all its limitations may be its greatest strength. At the same time, the modus operandi of this pluralism may be India's greatest weakness. Institutional and Policy Framework: Civil Society in the Study Areas Samaleswari and Kalinga are characterized by a weak civil society, a weak state, and a weak state-civil society synergy, all of which are strong- QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 121 ly mediated by the highly bureaucratic, justifiably profit-oriented, and complex organization that is Coai inuia. m ine srudy areas, the institu- tional and policy framework that could potentially foster generalized trust and promote cooperation among inaividuals and social groups lacks cohesiveness and consistency. The coal company's influence is so strong here that its policies and programs are perceived by iocai residents as integral to, if not a replacement for, the larger institutional and regula- tory context enabling social exchange in Indian society. For instance, the only NGO that operates regularly in either of the study areas is CART, the NGO brought in by Mahanadi Coalfields to Samaleswari as part of the World Bank-financed projects. The new vil- lage working groups created in the Bank-financed mines to implement community development activities officially came into being when they signed an agreement with the company. As a result, these groups tend to be perceived as a contracting agency for the company, while CART is often seen as an agent of the company. Moreover, most of the new lead- ers in the villages are employed by the coal company and tend to identi- fy themselves strongly with its interests. In general, then, the interests of the company have permeated civil society in the mining areas. These cir- cumstances provide the specific context for studying how social capital works in the study areas. The social environment that prevails in Samaleswari and Kalinga is not conducive to generalized social trust. Granted, given the larger con- ditions affecting generalized trust in India, promoting an enabling envi- ronment where generalized trust can grow and solidify is not an easy task or an effort that can render immediate results. At a minimum level, how- ever, and within its capabilities, "any institution with a developmental agenda must be at once engaged with the communities it seeks to serve and capable of maintaining its own credibility and effectiveness." (Woolcock 1998, p. 178). Yet Coal India has been unable to cement trust- ing relationships or maintain credibility with the communities affected by its operations. One of the main constraints is that the coal company is simultaneously the main agent affecting the social, economic, and natur- al environments of these communities and the agency in charge of help- ing those affected by these impacts. The relation of dependency and sense of mistrust that develops between the affected communities and the coal company shape the ways the company's policies and programs are applied. The communities' demands and attempts to bargain with the company test the limits of the company's commitments and cooperation, causing the "rules of the game" guiding intervention strategies to change constantly. Multiple sets of game rules were in fact found to coexist in Samaleswari and Kalinga, which undermined coherent social interaction 122 UNDERSTANDING AND MEASURING SOCIAL CAPITAL and increasingly disconcerted local civil society. Financing requirements, state policies, and the relative autonomy of the coal company subsidiary all prevented the development of a single, consistent, and clear policy for handling the community development responsibilities assumed by the company. Before 1980 there were no clear guidelines for community development. With the Bank-financed projects, Coal India prepared coherent guidelines for community development, for guiding the actions of company officers, and for informing communities about what to expect from the company. But these guidelines were applied only to 24 mines and to those villages located within 1 kilometer of the mines' lease- holds, and not to other mines or villages in the region, and that differen- tiation represented, at least during the study, an additional source of con- tradiction and inconsistency. As Hyden (1997) explains, individuals can- not be persuaded to cooperate genuinely or respect each other where the institutionalframework is neglected.15 These inconsistencies created differ- ent expectations that have not been completely met by the coal company or the government. Family-Kinship: The Weaknesses of Strong Ties According to Buckland (1998), in South Asia, cooperation beyond the extended family unit is relatively uncommon. In the study areas, as for other regions in India (see Hatti and Heimann 1992), the household is central to an individual's life, and a generalized sense of distrust toward nonhousehold members prevails. In this context, the social capital creat- ed within this family or kinship circle may not be as valuable as is often assumed (Putzel 1997). A key issue that hinders the evolution of informal institutions-of norms and values-is precisely that they are constantly reproduced within the family realm and enforced through strongly held ties based on dependency relationships that demand submission from certain members. Often, then, collective action for community develop- ment is not based completely on voluntary participation or is not as col- lective as it seems. Any assessment of the nature of social capital in a given community must thus probe into the strong ties created by family and kinship relations, and the extent to which these ties influence the access to and control of resources, including social capital itself. Family structure in both Samaleswari and Kalinga has changed signif- icantly with the development of coal mining. The number of nuclear fam- ilies has increased, in part because of the presence of the mines. Often, when one brother gets a job, for example, and another does not, tension and a diversion of interests develoD. and the extended familv disinte- grates. This is perhaps more pronounced in Orissa, where the resettle- ment and rehabilitation guidelines stinulate a "one family-one iob" QUALiTATiVE ANALYSiS OF SOCIAL CAPITAL: ORISSA, iNDIA 123 approach. The 1994 survey of project-affected persons in Samaleswari indLicaLItedL that Ithe nrLb oILfl. nucla ILLt..L iULlies increaIsed frorr,L 39.7 pecentI before mining operations to 47.7 percent at the time of the survey. Srlt LWLar.eousl;, thCeyIC p nae6 M. JoL IUtalIIiS MCaI dCeaScrs fromII 6u.. percent to 52.5 percent (VPASP 1994b). Incidence of female-headed I It IJlulO I L Oll ll V V CI) I) iVW. Power of decision-that is, "the room for agency" of particular social actors-is highly,1 concentrated ihr h ncerf .iisi h suy 0 i L1r,LUy %.U1L%LC:LL1Lu VILILU-L LIlt ILULItCI IC1dLL111b Ult LIItZ bLUUy areas, especially since high levels of dependency are common. As this sLUUy corroboraLte, Vinte VI Lift lluair. Va1riLJ1C:b ClcLLllI ,hee dtpendency levels is gender (Kapadia 1997, World Bank 1991). Like Singh (1995), this s,dUy Uu1iU Ldtaat r1i if-t operainoup s iihave fiitueu women s economic opportunities and correspondingly increased their levels of dependency anu vulinerfuimi'Ly wiutUi nthe houseolu andu Ine community. Beiore mne mines opened, women found jobs as agricultural workers and earned income from collecting forest produce. More recently, access to the forests has been restricted, and agricultural work has decreased significantly, while me mining economy has not managed to offer sufficient "suitable" occupations for women, primarily for cultural reasons. The rate of female employment is thus dramatically lower than maie empioyment, wnile self-employment is practically nonexistent among women. Tne structures of constraint affecting women at the family-kinship level tend to be stronger among caste Hindus than among scheduled tribes and castes. Tney also tend to be stronger among landowning culti- vators than among landless laborers or marginal farm families.16 In part, this may be because women have a higher status among the scheduled tribes and castes, where they are appreciated as an economic asset, than among the middle and higher castes (see also Femandes ands RaJ 199Z). Although tribal and lower-caste women are not considered equal to their husbands, they do seem to have greater freedom than the middle and upper-caste women, who might go through life subordinated first to their fathers, then to their husbands, and then, in widowhood, to their sons. These differences notwithstanding, as social capital resources undergo cycles of contraction due to economic stress, households in general tend to withdraw from the larger community, furthering the isolation of women (Moser 1996). In summary, at the intrafamily level, the use value of social capital varies for each member depending on his or her gender. Furthermore, a family and the community to which it belongs may have plenty of social capital resources but this may not provide access to resources that will help women and other vulnerable family members overcome poverty or benefit them directly. Access to forms of social capital outside the house- hold and to other resources depends, in this context, on the degree to 124 UNDERSTANDING AND MEASURING SOCIAL CAPITAL which one's links with the outside world are mediated by other family members. Availability of social capital resources may therefore be inde- pendent of opportunities for access and control of resources. This is par- ticularly important for women, since the power of mediation of males is likely to increase with contractions in social capital resources, particular- ly under conditions of stress, scarcity, and poverty like those found in Samaleswari and Kalinga. Horizontal Networks: Social Capital at the Community Level It is generally agreed that social capital, particularly among poor com- munities, is scarce in South Asia (Woolcock 1998). Upon a cursory look, India does not seem to be an exception. The study found, however, that mutual trust within the community is plentiful in the study areas, and that there might even be an oversupply of certain forms of social capi- tal.'7 Yet, while social capital resources in both Samaleswari and Kalinga may be abundant, they correspond to bonding social capital, which is not always conducive to the cohesiveness of the community at large.18 Moreover, extreme poverty has probably discouraged collective action and curtailed its effectiveness by limiting people's time horizons and social interaction while augmenting self-interest and distrust toward out- siders. The considerable lack of horizontal linkages in the two villages most likely stems from a highly fragmented social structure, character- ized by closed groups with high entry costs. The pattern of segregation that emerges from this is both social and spatial. COMMUNI LIFE. The village system, although certainly in flux due to the changes triggered by the mining industry, has maintained a set of struc- tural rules (that is, guidelines or external norms and values) that influ- ence what people do for and receive from other community members (Neale 1990). The distinct social stratification pattern characterizing the villages located in the study areas is a direct reflection of these rules. In Samaleswari, social and spatial segregation are high. In two habita- tions, Sukhpara and Orampara, the entire population belongs to a single scheduled tribe, while in Mundapara most of the population belongs to a single scheduled tribe. (table 5.1). Although they are a minority of the population, the Dixits (Brahmins) dominate local politics and village management. They are the main landholders and have traditionally been the major decisionmakers of the area. Lajkura has the highest concentration of scheduled castes, who are still restricted to an isolated hamlet and treated as untouchables by the upper caste group. Because of the rigid social norms and values imposed on them by the community at large, scheduled castes are not allowed to QUALITATIVE ANALYSIS OF SOCIAL CAPiTAL: ORISSA, INDIA 125 Table 5.1 Total Population of Samaleswari, by Category Scneduled Scheduuled Orher General Village castes tribes backward classes castes Total Kudapalli 41 345 260 77 723 Lajkura 219 190 254 61 724 Sukhpara 0 130 0 0 130 Mundapara 7 229 27 0 263 talltpalli -t 108 200U 0 4J1 Orampara 0 112 0 0 112 Karapalli 34 96 232 6 368 Total 349 1,285 973 144 2,751 enter trhe temple oi tne rmain village or to use the oatning platform tgnar). Nor are they allowed to participate in the main village functions or in any collective decisionmaking process. As a group, the scheduled caste pop- ulation displays high degrees of internal social connectedness and rela- tive cooperation. Despite these characteristics, tney are stii highly depen- dent on the higher-caste group because they lack economic resources and access to market opportunities, especially employment and credit. The level of dependency has diminished with the consolidation of the coal mining economy. A higher level of social cohesiveness exists in Kalinga, largely because the traditional power groups have managed to maintain more tradition- al cooperative arrangements at the village level, and agriculture is still a relevant economic sector. This higher level of cohesion seems to stem from the existence of groups with strong internal ties and the economic and institutional resources necessary to impose their will on the rest of the community (table 5.2). In Kalinga, the Chasa (farmers) category-the largest among the "other backward classes" population-is perceived as the main power group. This is not surprising, since the farmers have tra- ditionally been the main landholders in this area and have high social and economic status. Due to their power-wielding status, the Chasa appear to dominate the decisionmaking process and to be able to center the community consensus around their own priorities. According to Grootaert (1997, p. 80), "...the creation of trust and reci- procity is more likely in horizontal groups, especially those based on kin- ship or other dense networks (for example, based on gender, ethnicity, or caste)." The study found that in this respect, Samaleswari and Kalinga have plenty of social capital resources, but they were fragmented among castes and gender and not always conducive to the cohesiveness of the community at large. Portes (1998) identifies four potential negative con- sequences of social capital: exclusion of outsiders, excess claims on group 126 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Table 5.2 Total Population of Kalinga, by Category Scheduled Scheduled Other General Village castes tribes backward classes castes Total Bramhanbahal 140 240 190 160 730 Danra 610 280 3,000 110 4,000 Kalam Chhuin 420 60 2,825 55 3,360 Majhika 70 25 655 0 750 Natada 280 430 1,150 15 1,875 Nathgaon 0 0 63 7 70 Nakeipasi 125) r, r, r, QA 6Q9 1,25)0 Solada 650 350 1,925 135 3,060 Total 2,295 1,450 9,888 1,462 15,095 members, restrictions on individual freedom, and downward leveling norms. All of these consequences, except the iast one, seem to exist in the study areas.19 Under the conditions prevailing in Samaleswari and Kalinga, community feeling tends to be fostered by inclusion-exclusion mechanisms, and social capital can become an antidemocratic and poten- tially disruptive force. When alienation from the community is deep, as is the case with the scheduled castes in Lajkura village, for example, the sense of suspicion can turn into hostile relationships toward "foreign"" neighboring communities or individuals. As a group, both men and women belonging to the scheduled caste in Lajkura do not express much sympathy for the upper castes. Exclusionary mechanisms in the study areas discourage the formation of bridging social capital and simultane- ously promote the emergence of closed groups. Significantly, the closed nature of these groups tends to exercise excessive claims on some of their members, notably women and disenfranchised social groups. Ethnic and caste divisions constitute an important factor affecting social interaction in the study areas, which has hindered the emergence of generalized trust and increased the transaction costs of social and economic exchanges (see also Collier 1998). Gender, again, is an important variable. In both Samaleswari and Kalinga, many women have developed very informal, yet relatively strong, "support networks" with other women outside their households, although within the confines of their own social and economic, and often, spatial location. Much as males are isolated in their social interaction by caste, class, and related issues such as occupation and identification with work, women also tend to separate themselves by caste and class. Not only do women continue to occupy a completely separate social dimen- sion from men, even when facing the outside world, but they do not attempt to build bridges either across networks or to groups of women QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 127 different from "their own kind." Since most local women lack autonomy and econormic resources, their socia; capital represents instrumenrta potential, but this potential is of little real value. The resources made availabie to those women wnen they use hneir sociai capital are minimaI and insufficient to make a noticeable impact on their lives.20 Within exist- ing socioeconomic arrangements, their room for action and "me cnances o0 improving their well-being through collective action still remain highly constrained. A sense of "fictive kinship," to apply Kapadia's (1997) term, could be detected among company employees, as evidenced by the tendency to address each other as if they were part of a large family, the coal compa- ny "kin." With the entrenchment of the mining economy in the area, many males have acquired a strong sense of identity with the company, while the prevalent aspiration of males-and of many females-not yet working for the company is to become a Mahanadi Coalfields employee. This close identification with the coal company has provided some inter- village linkages of similarly socially and economically positioned men. This is not surprising, since occupation, workplace, and the organization of production are likely to play a key role in the development of social ties and horizontal linkages between individuals and groups of individuals. In the study areas, the centrality of mining activities may make these factors even more relevant.21 The networks of fictive kinship found in the study areas, however, are loose and without a strong sense of collective interests. These characteristics are the result of two interrelated factors: first, the ties of fictive kinship do not cross gender or caste divisions, because so few women are regular company employees and certain castes tend to concentrate in certain activities; and second, the interests of mine employees often differ from others in their communities and even from members of their own ethnic or caste groups or family members who are not part of the company. The latter factor directly affects the feasibility and nature of collective action, more so as company employees have increas- ingly taken power from the traditional leadership (ORG 1995). The coal company's scheme for compensating people depending on their designation as a formal employee, a daily wage worker, or an enti- tled project-affected person has created new power relations and fur- thered inequality. In Samaleswari, the situation is especially precarious. All the villages in Samaleswari are entitled to community development benefits, but many villages coterminous with them fall under the old Coal India policies and residents feel that they are treated differently under the same circumstances. Moreover, because the resettlement and rehabilitation packages benefit individuals rather than the community as a whole, the incentives for cooperation and collective action within the villages are low (Eyben and Ladbury 1995). Instead of working together 128 UNDERSTANDING AND MEASURING SOCIAL CAPITAL to try to obtain benefits for everyone in the community, individuals com- pete for designation as one of the categories of people who receive com- pensation. Tlrhe development of a market economy in the study areas has also been an incentive for individuals to act on their own rather than within a community structure-perhaps even a more important incentive than the divisive compensation scheme. In a traditional village economy, the sense of community is tied not only to the place where people live, but also to a system of traditional economic production and transactions. As the market increasingly provides individuals access to the same goods they once had to obtain through collective action, individuals feel less compelled to cooperate with those outside their close-knit groups. Increasing commercialization of agriculture (in Kalinga) or the virtual disappearance of agricultural activities (in Samaleswari), coupled with the monetization of the economy and the provision of infrastructure and certain services-at least for some people-by the coal company, has dramatically eroded the need for collective action to provide collec- tive goods. In turn, the traditional system of resource sharing and ser- vice obligations has given way, and the reciprocal caste-based relation- ships between landowners and laborers have been replaced by market relations. Lower-caste groups are not compelled to interact as much with the landowner castes. Social cleavages are simultaneously strengthened as the more powerful groups feel threatened and lower groups become more independent of them, albeit more dependent on the coal company. COMMUNITY-BASED ASSOCIATIONAL LIFE. Social capital in the study areas is high but simultaneously fractured by gender, caste, and class. Significantly, but not surprisingly, the study found that community- based associational life is very low (table 5.3 and table 5.4). Most social interaction takes place outside community-based groups, within the divi- sions established by the existing social hierarchy. Although most survey respondents in both study areas indicated that they are aware of the exis- tence of the community-based groups discussed below, they were unsure about the groups' functions and showed a low level of interest in getting involved in these types of associations. The onlv visible groups in the villages are the mahila mandals (women's clubs), most of which have died out due to lack of interest and resources.22 In Samaleswari, women in four villages organized mahila mandals, but only three are relatively active.23 In Lajkura, a club organized by scheduled caste women in 1987 to raise the social standing of the scheduled castes is no longer in operation, due to lack of guidance, inad- equate funding; and intermal power conflicts. The situation is similar, if QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISJA, INDIA 129 Table 5.3 Community-Based Groups in Samaleswari Mahi`la IYoulth V7illage Village mandal club working group Lajkura inactive Active Active Sukhpara Nonexistent Nonexistent Active Ku'apudapill Aicve [iacLive tcCive Orampara Nonexistent Nonexistent Active Ainapalih Acti;ve inacti;ve Acti:ve Karapalli Activea Inactive Active X\4nA-a--n A r+ -,e Nonexistent Actfi:ve a The mahila mandal of Mtmdapara also covers Karapalli. Table 5.4 Communiry-Based Groups in fKalinga Mahila Youth Village management Village/town mandal club committee Solada 0 4 0 Bramnhanbahal 0 0 0 Nakeispasi 0 2 0 Majhika 1 1 1 Kalam Chhuin 3 5 3 Natada 1 1 1 Danra 1 3 1 Nathgaona 0 0 0 a. Nathgaon is a new settlement where all households are still dependent on their respective "mother" villages. not worse, in Kalinga, where most of the women's clubs found in four of the villages (Danra, Kalam Chhuin, Majhika, and Natada) are inactive.24 Youth clubs, common throughout india, aiso have been established in the study areas to promote youth participation in a broad range of activ- ities including sports, cultural activities, literacy programs, heaith aware- ness, and community development activities. In Samaleswari, four vil- lages have youth clubs, but only one, in Lajkura, is registered and active. The others suffer from lack of interest and of any binding force. In Kalinga, youth clubs are found in each village or town except Bramhanbahal and Nathgaon and are much more active than in Samaleswari. In general, clubs in the study areas have between 40 to 50 members, especially in Kalinga, where population density is higher. Most of the members are educated. Membership in the club is restricted to youth of a particular village or hamlet. 130 UNDERSTANDING AND MEASURING SOCIAL CAPITAL An important qualitative difference between the two study sites is that Lie tLLadi,iol gd UraaI pariclUa L aitU i (village II LL ta6naeLUI t L VL LLtteLi.L) I a survived in Kalinga but not in Samaleswari. Villages across India were tradiltionally managedu bPy this LyFe VI LVilUL,ILL, bULutta hire:le:van%ce diminished, and many disappeared after the consolidation of the pan- chayar systerm- for ndalaguid g comLLLImuLLity developmLent.25 Contrary to this trend, every village but one in Kalinga has an active village management committee.26 The traditional village managem-lent systemi, has s-urvived it Kalinga because most of the villages are much older and more populated tnan the ones in Samaleswari, and they are compo!sed oI various caste groups. Agricultural work and land-based employment is still significant in Kalinga, where villages are much less dependent on the coal company than they are in Samaleswari. 'Village management committees in Kalinga are primarily responsible for resolving disputes and conflicts within the community, managing common property resources and temples, and organizing cultural events.27 The committees also make sure that the village's pond water is equally shared among villagers for irrigation during the drougnt season. The committees are informally organized and composed of village elders, following traditional arrangements. The leader is a senior male, who in Kalinga is called Sabhapati (President). Most of the presidents and other members of the committees are from the upper-caste "farmers group." Participation of scheduled tribes and castes representatives is, as it has historically been, very low. Members of the panchayati raj pay respect to and consult the members of these committees about important decisions such as the selection of beneficiaries for government-funded programs. No horizontal linkages exist between the management committees of the various villages, except for limited agreements among certain villages to guard the nearby forests on a rotating basis. Disputes between two vil- lages are often resolved at a joint meeting of the two committees, while major cases are taken to a court of law. In both study areas, the panchay- ats were the only grassroots institutions with an intervillage character, but they tended to reflect the villages' social structure and its related cleavages, and many people said that they had lost trust in them. Under Coal India's new community development guidelines, village working groups were promoted in Samaleswari to ensure beneficiaries' participation and enhance their sense of ownership. These working groups are supposed to have 10 to 15 members representative of all castes and tribal populations in the village; at least two of the members are expected to be women. Instead, the groups tend to reflect the existing social structure and power relations of the villages (table 5.5). In Lajkura, for instance, only 1 of the 17 members of the working group belongs to the scheduled castes, even though the scheduled castes represent about 'Table 5.5 Composition Profile of Village Working Groups in Samaleswari Percentage of scheduled castes and tribes By gender By group - - Percent of Percent of Other scheduled scheduled Total Scheduled Scheduled backuward castes in tribes in Village members Male Female castes tribes classes Others village village Ainapalli (431) 14 12 2 1 2 11 0 11 42 Karapalli (368) 10 10 0 0 0 110 0 9 26 Orampara (112) 11 11 0 0 11 10 0 0 100 Mundapara (263) 12 9 3 1 7 4 0 3 87 ILajkura (724) 17 17 0 1 6 1 1 30 26 Sukhpara 1(130) 10 10 0 0 10 10 0 0 100 ]Kudapalli 1(723) 18 13 5 0 5 6 7 6 48 Note: Total population of village is given in parentheses. 132 UNDERSTANDING AND MEASURING SOCIAL CAPITAL one-third of the total village population. In the entire study area, only three ~ ~ ~ I- 4vo-in grushd e.ae m.el-L.bers at thte t;met of the tudy theY LIUCC VVVUI'~LIL 5lkUt O 11QOA I di,a I iLCiU1- aLL 4--IAI UILC OL.1- "A- .Ay,_L other villages said there were no "suitable" female candidates. Despite concertMeA e.._t by1- tefC-1ita-9n NTGO'- hiredA 1- Coal TA.dia to eul t.kAC1C C11IUI Ly LItC IGU..IrLaLi1L6 VW%3'.J ILLICI.A \.Ua LILUIca LU ~JAl ize" the village working groups, they tend to be dominated by the allCeay pUower"2. Neighboring villages in Samaleswari do not seem to find any common ilit restLs. D'espitLe h'Ligh1- levUelsn of in[-group cateCb VI ethL.slidait- ; wUCiL VLIL.i the habitations, no strong networks of horizontal associations linking the same groups from'L one village to the next could be identified. Simlilarly, no horizontal linkages among the various village working groups have ueveloped 11d in S aleds-Wdli, UespiLte nll proximity of tthe V1i1daes adUU o1 the coal mines themselves. Considering this situation, it is likely that cornnunity-based deveelpmIienlt Lulu achieve oiuy very iimouest result here. Bargaining power is fundamental to establish respect for freedom of association, which in consequence demands a degree of "scaling up" or organization that must transcend the local level. Regional horizontal net- works and organizations are an option for scaling up, since tney provide opportunities for linking dispersed solidarities, increasing bargaining power, and facilitating access to information. None of this is present in Samaleswari, or the nearby mining areas that comprise lb Valley. Nor does the coal company seemed very interested in addressing this issue. iNTERPERSONAL TRUST AND THE NEED FOR COOPERATION. Table 5.6 provides a comparative measurement of degrees of generalized trust in the two study areas. As it indicates, individuals say they count on the entire vii- lage to help out in times of personal crises such as a death in the family or a conflict between two individuals or families. Yet cooperative ettorts in other situations have not been particularly successful. Indeed, cycles of conflict and cooperation have been a central feature of the interactions among the various social and institutional actors involved in the com- munity development and resettlement and rehabilitation activities in the study areas. Periodic cooperative efforts have failed to resolve issues of power, equity, and access to resources in any lasting way, reinforcing atti- tudes of suspicion and distrust and creating a vicious circle. The survey showed that more residents of Samaleswari (84.9 percent) than of Kalinga (74.0 percent) were aware that collective action could improve their communities' quality of life. The higher level of awareness in Samaleswari was to some degree related to the efforts of the facilitat- ing NGO working there. Close to 84 percent of survey respondents in Samaleswari also expressed their willingness to contribute either time or money toward activities that would improve the quality of life in their communities. Considering the level of social fragmentation, this stated QUALITATIVE ANALYSLIS OF SOCIAL CAPITAL: ORISSA, INDIA 133 Table 5.6 Degrees of Generalized Interpersonal Trust (nprcrent) Social unitl Samaleswari Kalinga Type of event institution (n = 141) (n = 235) Festivals Within family 15.0 5.5 Neighbor 6.8 6.8 Caste group 7.5. 3.8 Entire village 68.7 82.6 Neighboring village 2.0 1.3 Economic loss No one 9.3 13.2 Relatives 30.5 28.5 Neighbors 43.7 48.5 Village money lender 15.2 8.5 n ......- L..1. 1 A I I Ix I a:1 VCIlil I .'t 1.J Dispute between Among themselves 18.8 1.7 two individuals/ Neighbors 9.4 3.8 families Caste group 2.7 0.4 Entire village- 66.5 92.3 No response 2.7 1.8 n case of INO onte J.2 U any death Relatives 12.3 26.4 Neighbor 26.6 30.6 Caste group 7.8 7.2 Entire viflaLe 48.1 35.7 Note: Indicator: Whom would people count on in case of various events? wiiingness may appear to be some form of cognitive dissonance. A clos- er probe, though, showed that "community" meant something different to each respondent. in addition, aithough peopie expressed wiiiingness to contribute time or money, they expected that the coal company would provide many of the needed public goods and that their contributions would not in fact be required. Finally, despite the realization that cooper- ation was important, most interviewees rarely interacted with people from other social groups and village clusters, and even felt that other peo- ple from their own villages were self-centered and did not care about the welfare of others. Not surprisingly, confidence in the sustainability of group effort was low. An important area for cooperation is the management of common property resources (table 5.7). in Samaleswari, where these resources have dwindled along with any kind of villagewide management system, a majority of survey respondents still recognized the need for coopera- tion. Yet roughly only 3 out of 10 respondents were willing to help orga- 134 UNDERSTANDING AND MEASURING SOCIAL CAPITAL nize and run this management system. In Kalinga, 6 out of 10 respon- uents were willlrng ,o sitCar Lits IebpUsitZ1UL1iLy. InLLt:1u1=L.lY, Udesite LLth low percentage of people in Samaleswari willing to cooperate in manag- ing cornmlion property resources, most respondentls were wullrtt to con- tribute money for it. Apparently, many people would not mind paying ror provision oi collective ser-vices, i-i uley couiu aIuoru IL, biecause u Iti past collective action was not very democratic. Traditionally, higher-caste groups assumed the role of supervisors, while the acrual physical -work was done by individuals at the bottom of the social hierarchy. This forced labor nas tended to disappear with tne increased influence or the mine economy and the related demise of the agricultural sector, and many of tne iower-caste groups do not want it back. In Samaleswari, willingness to contribute money for future infrastruc- ture maintenance was even lower tnan tne wiiiingness to participate in the management of common property resources (table 5.7). Most people felt that operations and maintenance were the responsibility of Mahanadi Coalfields or the government. The coal company had been providing infrastructure for years, and many people in Samaieswari did not agree with the new rules that required them to take on the responsibility, par- ticularly since the coal company was still maintaining the infrastructure Table 5.7 Attitudes toward Management of Common Property Resources and Viiiage infrastructure Samaleswari Kalinga (n = 141) (n = 235) Attitude Percent Percent Felt that community-based management is necessary 58.2 88.1 Willing to participate in the management of common property resources 27.7 60.0 Willing to contribute money for management of common property resources 75.9 62.6 Community-based operation ind maintenance (O&mM) of village infrastructure perceived to be necessary 31.2 74.0 Willing to participate in O&M 25.5 28.5 vvilling to contribute money for O&M 25.5 43.0 QUALITATiVE ANALIYSiS OF SOCIAL CAPITAL: ORISSA, INDIA 135 in some nearby villages. In Kalinga, people felt stronger about participat- ing nroprationn and rranfon nro activities fof village~o inr2ctn rh ir Here, the Periphery Development IProgram-the ad hoc form of commu- nity development implemented by Mahanadi Coalfieldscontinued to operate, and the villages were enjoying its benefits without having to A .J{ A A Llt AL LoI Lt~LI UAt L' . voctZ r u gr andl . *v1. Vertical Articulations -The Company and "Its" Communities Largely because Coal India has assumed more and more local community UeVeU}lopmen1LL IUsnIbL;Ls, LIVII sIL1cit ULi t)al 0 svLa1iYa1 ard Kxaliur.a ILh forged few vertical articulations with the local and central government agerciLets ULCtL VVhUIU itave cod1UUcLt,U ULLCO aLUVLl;es in uLt coai c1tFally absence. Its mounting responsibilities have gradually overwhelmed the coal coriLpar-y, Uir. putrL because of1 its ilL,y tU 101 Cl a pa1L1tes1lIp vvilU ULhe communities it is meant to assist. The company's own lack of capacity and interest in dealing wiul cununIuiUUty developumelt issues has kept it fromn promoting the creation of the very community networks that could have raciltated community developmenet activiies. DUt aithough the com-lpaniy could be doing more to address this situation, it is highly unlikely that it could have been rully successful on its own. nven ir stronger honzonrai linkages were established across the study areas, not all the development issues could have been solved at the community level, let alone solely bly building social capital through the promotion of associations. On the one hand, local and community development and poverty alle- viation are tasks that cannot be assumed entirely by a single agency, even one with an increased capacity to build social capital. That is especially true when the agency in question is a coal company under great pressure to improve the bottom line of its mining operations. On the other hand, by becoming the de facto central agency in the area, the coal company has displaced an important form of social capitai: the mutually supportive and complementary relations that could be provided by other sectors of society such as national and international NGOs, government agencies, and private firms. Without these vertical articulations, which are neces- sary to break down the isolation of local community groups and provide them with better access to other (informational, economic) resources, the potential of the social capital prevalent at the community level to improve people's lives cannot materialize. To assess social capital in the context of community development, it is essential to look at the actual and potential winners and losers in trans- actions mediated by social capital. Social capital does not have an absolute value, but a relative one. Whether social capital has a beneficial 136 UNDERSTANDING ANLD MEASURING SOCIAL CAPITAL or detrimental nature depends on the social or institutional actors' stake it ,he poVLVess ul LlllUluLty ueve1011Ltp1Len-a _1LaIa%_LI Li tLiat imia es the valuation of social capital difficult. On the occasions when collective effOrts, s-uclh as -hLe formia,ion VI v1lage wurIltr groups, apealu to bt;: compatible with the goals of the company policy, they were well received by minle offic:ials. But on the few 0LcdsionIs Wit-h stome peopl Ut Like study areas decided to come together to pool their bargaining power and oppose certain decisions on te company, nune oficials perceiveu this form of social capital development as undesirable. Many mine represen- tatives saw tne norton or building bridges across viiiages as a inreat to the smooth operation of the mines and therefore counterproductive. Implications for Future Research The study's findings suggest that social capital has a multidimensional nature and a relative and contingent value tnat depends neavily on tne social, economic, cultural, and political context. From these characteris- tics it follows that individual and group access to a community's sociai capital reserves are not equally distributed, while, at the same time, each form of social capital can have different, sometimes contradictory, effects. It would be rather limiting, given these conclusions, to assess social cap- ital resources by focusing only on associational membership or norms of reciprocity and trust, and by assuming that social capital always pro- duces beneficial forms of civic engagement (Edwards and Foley 1998) or that more of it is always better (Woolcock 1998). To assess the value of social capital, one should therefore consider its multidimensional nature and its overall distribution and accessibility vis-a-vis particular commu- nity and household members. And one should take into account the effects of the nature of distribution and accessibility of these resources on equity-in terms of control of resources and control of access to resources-and democratic outcomes in a community. In this respect, two key sets of variables affecting social capital that need to be investigated more intensively are political engagement and power relations. Relative and Contingent Value of Social Capital Social capital may have a different meaning and use value for each of the concerned individuals and groups, depending on their specific social, economic, cultural, and political context (Burt 1997). In the particular case of community development, whether social capital is experienced as ben- eficial or detrimental varies according to the stakeholder. It would thus be relevant to identify both the range of possible effects of social capital and the particular costs and benefits of social capital for the concerned QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 137 stakeholders. Social capital at the community level should therefore not be assumed to be the sum of "individual" social capital (Edwards and Foley 1998, Portes and Landolt 1996). And social capital in general should not be assumed always to have a positive value. Under certain circum- stances, a form of social capital that might otherwise play a critical role in facilitating certain actions or resources may become useless or even harmful (Coleman 1988, 1990; Harriss and De Renzio 1997). The use value of social capital is also affected by its own social loca- tion, independent of who appropriates it (Edwards and Foley 1997). To put it differently, the value of a given form of social capital for enabling some action depends to a large extent on the social and economic location of the social capital in a community. In fact, the availability of social cap- ital cannot be assumed to be equivalent to the resources obtained through it, given that these resources can be simultaneously limited and con- straining (Portes 1998). While some social capital may be nested within dynamic sectors, for instance, other forms of social capital may be con- nected to declining or contracting sectors. A group that appropriates this latter form may obtain some benefits, but these may be short-lived. Distribution and Access to Social Capital Resources Access to and control of social capital resources are not equally distrib- uted within the family or throughout a community (Bourdieu 1986, Granovetter 1985), while social capital is often considered valuable pre- cisely because of its scarcity and exclusivity (Whittington 1998). In par- ticular, access to social capital depends on one's social location and is con- strained by various factors such as the level of geographical, cultural, and social isolation; lack of financial resources; and the specific institutional arrangements that structure everyday life (Edwards and Foley 1997). Thus, at the group level, it does not always hold that the resources a per- son obtains through his or her relationships within a group are available to all members of the group or that all members of a group have equal access to the group's resources (Astone and others 1998). Effects of Social Capital The study's findings further suggest that each form of social capital can have different, sometimes contradictory, effects. In the notion of social capital that has developed on the basis of Putnam's work, there has been a tendency to romanticize the image of community and to neglect some of the adverse effects of sociability (Portes and Landolt 1996). This study indicates, as does other recent research (for example, Portes 1998, Woolcock 1998), that sociability can have desirable and not-so-desirable 138 UNDERSTANDING AND MEASURING SOCIAL CAPITAL consequences. When evaluating the social capital resources of a given community, the analyst should identify and account for both the positive and the negative effects of social capital (Putzel 1997; Rose, Mishler, and Haerpfer 1997). Social capital can facilitate all kinds of individual and col- lective efforts, ranging from community development organizations to death squads (Edwards and Foley 1998). Often, though, it has been assumed, misleadingly, that social capital is not just a public good, but intrinsically for the public good (Putzel 1997). The normative underpin- nings of what is desirable or undesirable notwithstanding, it is safe to assume that social capital can be a public good and a public bad, alternative- ly or even simultaneously. In other words, as a common property (social) resource, social capital may have positive and negative externalities. Social capital can certainly enhance the efficiency of physical and human capital and can facilitate economic exchange and coordination. But, as observed in the study areas, some forms of social capital are illib- eral and socially exclusive. Certain civic society groups can use their social capital resources to exercise control over other community mem- bers and secure a considerable share of the community resources (Woolcock 1998). Importantly, as Granovetter (1985) puts it, although social relations are often a necessary condition for trust and trustworthy behavior, they are not sufficient to ensure that behavior. On the contrary, social relations may actually provide opportunities and means for malfeasance and conflict, perhaps to a greater extent than if they did not exist at all. As the case of Kalinga suggests, a group may actually have access to too much social capital. Certain networks of civic engagement may be a source of trust within, but they can also incite distrust from without. Political Engagqement The networks and relationships created by associations do not guarantee by themselves political outcomes (Putzel 1997). There are significant dif- ferences across various kinds of social clubs and organizations regarding levels of political mobilization and community action. In the long run, an individual's rich social life may not necessarily translate into political competence (Almond and Verba 1963). It is relevant for future research, therefore. to distinguish between the "mechanics of trust" (the oneration of networks, norms, and the like) and the "political content and ideas" disserninatpd through these networks and reflected in the shared norms (Putzel 1997). This diiqf-inri-on impnlie nnt only that the capacity of well-mnhilized groups to make effective demands on government can remain limited. It ailso imnlies, as .Wbifinatnn (1Q99) rcorprecly argues,z that civic assocria- QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 139 tions can cultivate resentment that leads to political turbulence and dis- LL-Us. 11ere is no IUuarantee tha. UeII.andU will -lways be dLALLULIdLI Iln nature or that they will not conflict with the social order or, as in this spe- cifii case, with the stated objectives of developmi-tent projects. Mobilization of the population can lead also to the emergence of very particularistc aemands that may undermine aemocracy more tnan enhance it. Power Relations As observed during fieldwork, social capital resources exist within the same social space as power relations. According to roucauit (i98u), power is coextensive with the social body, and power relations are intri- cately woven into other kinds of social relations, such as economic pro- duction, kinship, family, or sexuality. Therefore, a full understanding of social capital requires not only an understanding of these relations, but also an understanding of the exercise of power and strategies of resis- tance that take place at the moment of social exchange. This element of the social exchange involves-or to be more precise, produces and repro- duces systematically-relations of inequality intrinsic in power relations (Agger 1992). Simultaneously, people are positioned differently with respect to the resulting power structures, which are constituted by vari- ous axes of identity, such as gender, class, and ethnicity (Kapadia 1997). In other words, social relations play a conditioning role for power rela- tions while being simultaneously conditioned by them. This assumption, which implies, in Foucault's words, that "power is always already there" in social exchange, does not mean that it is useless to promote social cap- ital. Foucault understood the extent to which government institutions mattered, but he also realized that the dispersed mechanisms of power that existed outside the state, in civil society, were just as important. Altogether, these power relations constitute, as stated above, structures of facilitation and constraint to access resources and to control the access to and resources available in a society. Conclusions: Specific Implications for Community Development Lack of social cohesion represents a major challenge for successful imple- mentation of community-driven development in India. Community development strategies have increasingly become dependent on the social capital of the target beneficiaries. They require community consul- tation during project preparation and participation during implementa- tion to identify problems and priorities that arise under the constraints 140 UNDERSTANDING AND MEASURING SOCIAL CAPITAL imposed by low-cost investments, minimal subsidies, and cost recovery. This approach is also demand-based, requiring a high degree ot commit- ment by the communities to ensure sustainability. All these assumptions lead to one logical conclusion: the standard approach can work best through community-based groups that promote existing social capital or can build the needed social capital. In India, however, the standard approach has been based on unrealistic notions of the nature of community in the villages and of the possibilities and democratic content of collective action. Communities are assumed to be inherently democratic, and "inclusion" and "participation" guaranteed by promoting new groups that include representatives of all groups in a community. The fundamental constraint identified during fieldwork is that communities-village civic society in this case-are not always struc- tured in a way that facilitates the equal distribution of external resources or any kind of aid across the intended beneficiaries. In India, village civic society remains extremely hierarchical, and despite many positive changes brought about by development, persistent inequalities remain embedded in social capital resources. Under such conditions, villagewide groups, as indicated by the study and corroborated by other research (for example, Jayaraman and Lanjouw 1999; Mosse 1995, 1996), may not be effective mechanisms for democratic planning and collective decision- making. Most villagewide groups tend to ignore certain sections of the population and are dominated by the most powerful groups; women are given a very limited role, if any. If they are not closely monitored, new vil- lagewide groups are also likely to resemble this pattern. The study also highlights the difficulty of identifying appropriate social and spatial boundaries of the target community, given that these boundaries seldom correlate. Most community development programs in rural India assume that the village is the community and should therefore remain or be made self-sufficient. Under this premise, community is usu- ally associated with the official boundaries of the Revenue Village, as is the case under the new community development guidelines of Coal India. In reality, the village, composed of several hamlets, does not have a clear physical identity or a homogeneous social identity. The emergence of a sense of communitv at the level of the village would therefore be bet- ter approached as a dynamic process shaped by the various forms of dependence, competition, and factional conflict that characterize the hamlets' relations (Mosse 1996). Current approaches to community development see the association of beneficiaries as a necessary ingredient to achieving development goals, under the rationale that the social capital produced bv these associations makes the provision of collective goods more cost-effective (Stolle and Rochon 1998). Accordinglv. the strategy for nrenaring a community QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 141 development project has been to search for "developmentally valuable fUoirS of1 soCidl CdpiLdl, dItU lit Lit. dVabseLLtc of .Lhts, Lu 1toL .UiH:L1LU Lthe promotion of "desirable forms of social capital" (Beall 1997). Social capi- ta; in this context acquires uevelTprnenta1 value mainy as cuiyensatu- ry resource to address the deficiencies prevalent in society, including the unmet social needs of the intended beneficiaries. By extensioin, civic cooperation becomes narrowly defined as a tool to implement develop- ment programs and pians, while the use value of social capital is restrict- ed to cooperative and voluntary social relations demanding leisure time. Tlne problem is that in poor communities like those inciuded in this study, distinctions between voluntary and involuntary association and cooper- ation and between leisure time and working hours are generally moot. As much as sociability and collective action have benefits, they also have costs. investing in social capital, particularly in voluntary associations to implement community development activities or to manage certain com- munity affairs (such as operating and maintaining community infra- structure assets) puts additional clemands on certain individuals and groups in the community-generally the poorer and more disenfran- chised households and social groups, who are paradoxically already overworked and underemployed. It is in this context that the promotion of social capital for community development must be understood in order to avoid overburdening the intended beneficiaries with "volun- tary" membership in community groups. Finally, the instrumental value of social capital in project implementation has generally been overem- phasized to the detriment of the potential use value of translating these relations into political resources.29 It is reasonable to assume that NGOs could act as mediators between vertical and horizontal networks. NGOs, however, are often external agents operating in the middle of the delicate balance of interests and power existing in the comnmunities. Like Buckland (1998), this study indi- cates that the necessary emphasis on short-term objectives, such as income generation and provision of social services, causes NGOs and other agencies working in community development to neglect longer- term issues such as social organizing to maintain and build social capital. Promotion of community networks and extended normative behavior (generalized trust) are generally insufficient, while cultural patterns and social and economic trends hinder the emergence of new forms of civic engagement. Simultaneously, social capital promotion tends to introduce new norms and incentives for interaction that are heavily mediated by the NGO, which makes the sustainability of the enhanced social interac- tion heavily dependent on such mediations. In summary, the main implications of the study for social capital and community development are: 142 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 1. Social capital cannot be built, promoted, or transformed exclusively from within the community. 2. New types of community organizations should be promoted, but these cannot be designed in isolation from the social and economic context of the community and from the wider institutional framework facili- tating or constraining the generalization of trust. 3. Instead of one particular model of local organization (for example, vil- lage working groups), a wide variety of community organizations should be promoted, according to the various objectives and activities of the community development strategy. Reliance on group-based activities should be supplemented with other associational forms such as networks of individuals, and ideally these networks should reach outside a particular community. 4. Building social capital through community development requires trig- gering a process of social reorganization that takes advantage of infor- mal, often invisible, forms of association. New organizations are unlikely to change existing social relations immediately or in a short period of time. 5. When feasible, it may be better to start promoting the creation of small groups within existing social solidarities. In other words, it is critical to start working the existing social structure from the inside out. However, creation of small groups in isolation might exacerbate social cleavages in the long run. Horizontal linkages across these groups should be facilitated, and vertical articulations with state and private organizations deliberately sought. 6. Promotion of social capital must be complemented with concerted efforts to generalize social trust. Inconsistencies regarding the applica- tion of procedures, regulations, and requirements that affect the insti- tutional framework supporting the existence of generalized social trust should be avoided. The coexistence of many old and new rules of the game contributes to generalized mistrust. Simultaneously, credi- bility of the development agency is adversely affected when beneficia- ries in similar circumstances are treated differently. 7. Excessive reliance on community-based groups to achieve the objec- tives of community development might create unidentified dependen- cy relations among individual members, in particular the most vul- nerable ones, within the community. The greater the benefits expected from the community group, the more people are willing to tolerate excessive obligations. If the interpersonal bonds of dependency that may already prevail in a given community are transferred into the newlv created groups bv making them the main or onlv providers of certain services or public goods, chances are that these relationships of dependencv will he strengthened rather than weakened. QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 143 8. Sociability has costs as well as benefits, and the costs may be higher for the poor who do not have as much leisure or idle time to invest in building social capital purposely. 9. Tte external social actors, be they NGOs or other types of organiza- tions, should have as an objective to make themselves redundant in a reasonable time period. implications for Policy Social capital does matterfor community-based development, although it might be better understood as one of a variety of assets available in different degrees to individuals and groups in a community (Moser 1998). Social capital by itself cannot provide the solutions to major social and econom- ic problems and may actually be of limited value if not combined with other forms of capital, namely, natural, physical, and human capital (Serageldin 1996). The structures of facilitation and constraint that characterize a society demand that social capital be promoted simultaneously at the local level through networks of individuals and groups, and at the institutional and policy level. The state still must play a role in diminishing the close and often risky "personalized dependencies" of people on each other. Traditional associations that may have been necessary under certain cir- cumstances but that perpetuated subordinate roles for women and other vulnerable groups may be bound to disappear, and may be replaced by government institutions. The enthusiasm surrounding the concept of social capital stems from the acceptance that it is a resource or asset that provides an alternative to the heavy hand of the central government and an effective tool to improving governance. It is a concept that fits perfect- ly into the current promotion of decentralization. However important the features of social organization are to commu- nity development policy and projects, they are significantly affected by political institutions and their capacities, including the state. Attempting to create social capital without recognizing this fact is to attack "the symptoms, not the causes of the problem" (Tarrow 1996, p. 396). Societies and communities, such as those in the study areas, may possess abun- dant-if fragmented-social capital but lack other key resources and assets that would allow certain groups to escape poverty or make signif- icant progress in terms of democratic political participation (Warren 1998, Harriss and De Renzio 1997). Civic society and its social capital matterfor community development, but in the context of government institutions and the general institutional framework of society at large. Associations, albeit not all, may be built up with the sup- port of the state. As Putzel (1997, p. 947) concludes, "a strong state and 144 UNDERSTANDING AND MEASURING SOCIAL CAPITAL strong civil society must go together." In general, it appears that all com- munities are endowed with at least a minimal stock of social capital, from family and kinship ties to cooperative arrangements among friends and neighbors. The difference lies in how social capital is scaled up and whether it is scaled up through the interaction between the state and pri- vate and voluntary organizations in order to create solidarity ties and social action that reaches levels of political and economic efficacy. Finally, if the main objective of community-driven development is to empower individuals, to foster autonomy, to promote personal growth and self- realization through market processes, the notion of social capital is rele- vant in the sense tha. it underscores the importance of other necessary means of empowerment to support market processes and to facilitate the emergence of democratic, civic communities. Notes This chapter draws from a larger Social Capital Initiative study of the Coal Sector Environmental and Social Mitigation Project in India, of which Jelena Pantelic was the task manager. Field research was done in collaboration with Operations Research Group (ORG India) and Jonathan Glass, an intemational consultant. Christiaan Grootaert and Thierry van Bastelaer provided useful comments and suggestions during the research and preparation of this chapter. Beltrania Scarano contributed maps and graphics. Susan Assaf facilitated the management of the study's funds. David Fissel provided invaluable support during the preparation of this chapter. 1. According to Hyden (1997), the notion of social capital may be traced as far back as the 19th century. The notion of social capital entered development think- ing with full force in 1993, after Robert Putnam published Making Democracy Work. The notion had been around in contemporary thinking, implicitly or explic- itly, many years before, as made evident by the works of Jacobs (1961), Homans (1961), Granovetter (1973), Bourdieu (1986), and Coleman (1988), among others. 2. As in other World Bank-financed projects, these social mitigation measures were consonant with the bank's safeguard policies designed to avoid unnecessary social harm and to mitigate and compensate when needed to ensure that people's livelihoods and overall quality of life are not drastically affected by the new investments. 3. There were 24 onencast mines included under the ESMP scattered across 11 coalfields in 5 states of east and central India, managed by an equal number of suhsidiarips. 4. Every village has a given "revenue boundary" that comprises several habi- tations. In nothr worrds, vi!!age in tlis nApar refers to a "revenue vi!!7wag. Information on the study areas, unless otherwise specified, is derived from the household surveys and the database of project-affected people. QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 145 5. Entitled project-affected persons in Samaleswari include those individuals affected by the itVes-Cents finane by 4he IAIo1l4 P-l -ho were above 18Q y-- Lilt, lilY '.0 iIlit.Ali0 ililitlI ..Li L'~~~y -it V Vt'ii1 -.-ILIU -VI1,0 - - - -.~JJV UyC i of age by January 1, 1994 (assumed as the cut-off date when baseline surveys were undertaken in the area). 6. This information was provided directly by the Chief Mining Officer of the Kainga area. 7. The most important castes in Orissa have been the Brahmins, the Karans, the anLdLayats, aitU LIth Chasas. T iL I LiMULLi UL-UpplLI1iLo ltdVe UbeIL, IetpciL,vLV- ly, priests and scholars, writers, landholders and warriors, and cultivators (Lerche 19971). 8. Unless stated otherwise, this information comes mainly from the two staff appraisai reports prepareudu r the projects 'v^vorluDdIlak 199, 19/). 9. The task facing Coal India was enormous; according to Singh (1995), the mines are "spreadu over 30 uIstricts our uLierent states in inula, approximately cov- ering 2000 villages... 10. The usual active life of an opencast mine is between 25 to 30 years. 11. The Land Acquisition Act has been amended several times since its adoption, tne last time in 1984. 12. Implementation of social and environmental measures did not reach full speed until the middle or lYY/. The closing date of the project would later be extended until July 2002. 13. Early estimates indicated that about 16,000 persons wouid be affected by the project, about 10,000 of whom would have to be resettled. The total number of people entitled to rehabilitation assistance was estimated at about 9,200. 14. See Lenci (1997) for a valuable discussion on the "reconvergence" of eco- nomics and sociology. 15. Institutional framework in this context means, first, the vertical articulations between the concerned social and institutional actors (for example, Coal India, Mahanadi Coalfields, various levels of government, private sector agencies such as banks, and the horizontal associations, networks, and individuals themselves), and second, the policies, strategies, and instruments that provide a sense of gen- eralized social trust independent of (and that may acturally enable the expansion of) interpersonal trust. 16. Similar findings can be found in Fernandes and Raj (1992) and World Bank (1991). 17. In this respect, see also Madsen (1993). 18. This is based on an important distinction established by Putnam (1993) between bridging and bonding social capital. Bonding capital is limited to groups with similar characteristics and may enhance social divisions. 19. Downward leveling norms, according to Portes (1998), take place in cases when group solidarity is based on shared adversity and opposition to main- stream society. ln this case, individual success stories are seen as negative because they threaten group cohesion, especially since this cohesion is based 146 UNDERSTANDING AND MEASURING SOCIAL CAPITAL precisely on the alleged impossibility of the individual success of any group member. 20. Incidentally, this finding is congruent with Portes's (1998) assertion that social capital is not equal to the resources made available by accessing it. 21. The study's survey found that in Samaleswari about 50 percent of males between 18 and 55 years of age were employed in the mine. Kalinga has a more diversified economy, and the survey indicated that about 20 percent of 488 males over 18 years of age in the sample were regular mine employees. 22. These organizations have been promoted by the central government to increase women's participation in community development across villages in India. The official objective is to ensure social equity, economic empowerment, and self-reliance of women. In the study areas, few of these clubs were found to be formally registered under the Society Registration Act. 23. The active women's clubs met regularly once a month. The clubs were under- taking similar activities, such as meetings to raise awareness among women regarding health, family planning, and nutrition. 24. In Kalinga, none of the women's clubs has its own building, which made it difficult to conduct meetings and keep records. 25. The panchayati raj consists of a three-tiered system of local self-government to administer community development in India: a panchayat in each village, a coun- cil for each community development block, and a council at the district level (or zila parishad). 26. The village is Nathgaon, and the committee's lack of activity is due to its recent creation. 27. Sources of funds for the committees have traditionally included leasing out the village ponds for fish cultivation, charging fines during dispute settlement, and collections during village festivities. These funds are deposited in a common village fund managed exclusively by the members of the management committee. 28. Mosse (1995) similarly concludes that new organizations tend to reproduce existing power structures. For a contrasting view, see Fisher (1994). 29. Buckland (1998) offers a useful distinction in this respect. At the simplest level of analysis, social capital may be seen as central to facilitating cooperation and community participation as a means to reduce project implementation costs and achieve immediate results. However, when the "use value" of social capital is seen as "empowerment," the implication is that social capital in the communi- ty will facilitate social and political organization that, in the long term, will pro- vide wider access to other resources. References Agger, Ben. 1992. The Discourse of Domination. Evanston, Ill.: Northwestern University Press. QUALiTATiVE ANALYSIS OF SOCIAL CAPITAL- ORiSSA, INDiA 147 Almond, Gabriel A., and Sidney Verba. 1963. The Civic Culture. Princeton, N.T.: Princeton Ur-v-si Ir-. Astone - a -MT- a rie, Constanc A. N athansor -x br -c or -o n -J. C 1 -- - -.1.JL, li -vintic, - - --IOCCLL -n. iN LLUOtU-r,tL~lC~JitJiL CL J. Kim. 1998. "Family Demography, Social Theory, and Investment in Social '-a pital." 1AwTorkH- TPaper 9801. Al u. HoksPou.- jJ1I I ayI LQI. V vuIr..i6 i 'F; ~'U-Vu. JUiULD I 1uFyr,uL I OUyLIauui Center, Baltimore. Processed. Basu, Kaushik, and Sanjay Subrahmanyam. 1996. "Introduction." In T., 1-:1- 1DA4A-C+LD_ .._ -A _ 7 T __1:__l l-: 1xausIIllJrI LabU 1LU OdlLJay JUULJIUllIttyC1Il, CUb., Lruurcfnihls LLc IVUlLurt. Sectarian Conflict and India's Secular Identity. New Delhi: Penguin Beafl, Jo. 1997. "Social Capital in 1VVastea Solid InvesUnent?" Journal of International Development 9 (7): 951-61. Bourdieu, Pierre. 1986. "The Forms of Capital." In J. G. Richardson, ed., Handbook of Theory anid Research for the Sociology of Education. New York: Greenwood. Buckland, Jerry. 1998. "Social Capital and Sustainability of NGO- Intermediated Development Projects in Bangladesh." Community Development Journal 33 (3): 236-48. Burt, Ronald S. 1997. "The Contingent Value of Social Capital." Administrative Science Quarterly 42 (2): 339-65. Coleman, James S. 1988. "Social Capital in the Creation of Human Capital." American Journal of Sociology 94 (Supplement): S95-S120. 1990. Foundations of Social Theory. Cambridge, Mass.: Harvard University Press. Collier, Paul. 1998. "The Political Economy of Ethnicity." Paper presented at the Annual World Bank Conference on Development Economics, World Bank, Washington, D.C., April 20-21. Processed. Crook, Clive. 1997. "Survey of India's Economy." Economist 342 (8005): S1-S26. Eastis, Carla M. 1998. "Organizational Diversity and the Production of Social Capital." American Behavioral Scientist 42 (1): 66-77. 148 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Edwards, Bob, and Michael W. Foley. 1997. "Social Capital and the Political Economy or0 Our -isconten-i." American benuuiorau Scientist 0u 53): 66-7. 1998. "Civil Society and Social Capital beyond Putnam." American Behavioral Scientist 42 (1): 124-39. Eyben, Rosalind, and Sarah Ladbury. 1995. "Popular Participation in Aid- Assisted Projects: W'hy More in Theory tnan in Practice." in Nici Nelson and Susan Wright, eds., Power and Participatory Development: Theory and Practice. London: intermediate Technology Publications. Fernandes, Waiter, Geeta Menon, and Philip Viegas. 1988. Forests, Environment, and Tribal Economy: Deforestation, Impoverishment, and Marginalization in Orissa. New Delhi: Indian Social Institute. Fernandes, Walter, and S. Anthony Raj. 1992. Development, Displacement, and Rehabilitation in the Tribal Areas of Orissa. New Delhi: Indian Social institute. Fisher, Julie. 1994. "Is the Iron Law of Oligarchy Rusting Away in fhe Third World?" World Development 22 (2): 129-43. Foucault, Michel. 1980. PowerlKnowledge. New York: Pantheon Books. Granovetter, Mark. 1973. "The Strength of Weak Ties." American Journal of Sociology 78 (6): 1360-80. . 1985. "Economic Action and Social Structure: The Problem of Embeddedness." American Journal of Sociology 91 (3): 481-510. Grootaert, Christiaan. 1997. "Social Capital: The Missing Link." In Expanding the Measure of Wealth: Indicators of Environmentally Sustainable Development. Washington, D.C.: World Bank. Harriss, John, and Paolo de Renzio. 1997. "Policy Arena: 'Missing Link' or Analytically Missing? The Concept of Social Capital." Journal of International Development 9 (7): 919-37. Hatti, Neelambar, and James Heimann. 1992. "Limits to Cooperation." Asian Journal of Economics and Social Studies 11 (1): 55-71. Homans, George. 1961. Social Behavior: Its Elementary Forms. New York: Harcourt, Brace and World. QUTAL ITATIVE ANJATLYS!S OF SOCIAL CAPTTAL: OR!SSA, INDIA 149 Hyden, Goran. 1997. "Civil Society, Social Capital, and Development: Dissection of a Complex Discourse." Studies in Comparative International Development 32 (1): 3-30. Jacobs, Jane. 1961. The Life and Death of Great American Cities. New York: Random House. Jayaraman; Raishri. and Peter Lanjouw. 1999. "The E.volution of Povertv and Inequality in Indian Villages." World Bank Research Observer 14 (1): 1-30. Kapadiat Karin 1997 "Mediating the Meaning of Market Opportlnifies: Gender, Caste, and Class in Rural South India." Economic and Political W. eekly (November 27): 3329-35. Lenci, Sergio. 1997. "Social Capital? From Pizza Connectfiron to Collecve Action. An Inquiry into Power, Culture, and Civil Society." Working Pnnror ?dA Tn,c+if1it nf Crri:l qfilHiiPC TOP T-Tnmia L erche, TonQ 1QQ1. "-con-r. fl,opr.nt A Tr-ncCnrrn,4-nn of Traditional Social Relations: A Case Study of the Viswakarma 1B1,,ksmith and Carpntero, Class of O'risa India." CDRP Project Papenr 92.1. Center for Development Research, Copenhagen. Madsen, Aase Mygind. 1993. "Divisive Cohesiveness and Changing Caste Relations in ka-,lta-." CDR r AT.A- ,-- P-- 03 1 '-enter for Development Research, Copenhagen. Morris, Matthew. 1998. "Social Capital and Poverty in India." Working Paper 61. U.niversity of Susex T-s.",t_t of Deeli -ntStdis V Jr~ ...JUV 1OL U. .JUDDCA, LI LOLILUL S.C l LJV IJtIIL JLUI3 Brighton, U.K. Moser, Caroline 0. 1996. "Confronting Crisis: A Comparative Study of H-Tousehl Respnse to --vrt and Vleait;nFour TUrb-an iiuu ILI%iA l-xC7DUIt10C; LV I VVC:ILY CIILU VUIILCICIUL'fly UL LUI' Ulu)H Communities." Environmentally Sustainable Development Studies artd MVIonographts Sers 0. VYVL)Il L.kLt, VYC1IL1LrtLU1L, D.C. . 9O ".hLe A .......A Li... Rea.s.ssir.g Urban 1770. Litt: tA-eL VUuIICLdul1ULY rialllttewull. 1ttibxea5b11t6UfUc Poverty Reduction Strategies." World Development 26 (1): 1-19. Mosse, David. 1995. "Local Institutions and Power: The History and Practice of Community Management of Tank irrigation Systems in South India." In Nici Nelson and Susan Wright, eds., Power and 150 UNDERSTANDiNG AND MEASURING SOCIAL CAPITAL Participatory Development: Theory and Practice. London: Intermediate Tccholob, Publicati;ons. . 196 "TLocal T 44.-44-tin --A Far-.i- C-ste.-, D lp-1--nt- ±~U. L.JLII.ILOILLs- JILo. wLLI I iIuL,.y OLCIm L.~ -CMJF.XLCLLL. Thoughts from a Project in Tribal Western India." Network Paper 64. fl61 iLLtItusa, flC3 L Leat I C 13311 INC IExtesio Ne.LVVIJt r -A L NJ, LoUnol L. N'eale, VValtrI C.. 17799. Develtupli,l Rxurl Il;.d-PoULLiLte, PUliticu, ua,u Progress. Glen Dale, Md.: Riverdale Co. ORG (Operations Research Group). 1995. "Indigenous Peoples Development Plan for Sarnaleswari Open Cast Projet. (Mahanadi Coalfields Ltd.)." Coal India Ltd., Calcutta. Processed. Portes, Alejandro. 1998. "Social Capital: Its Origins and Applications in Modern Sociology." Annual Review of Sociology 24: 1-24. fl. j Al nfl ,rIf 2I I- x Portes, IAejandro, and Patricia Landolt. 1996. ine uownsiue of Sociai Capital." The American Prospect 7 (26): 18-22. Putnam, Robert D., with Robert Leonardi and Raffaella Nanetti. 1993. Iviaking Democracy 'vvork. Princeton, N.J.: Princeton University Press. Putnam, Robert D. 1995. "Tuning in, Tuning out: The Strange Disappearance of Social Capital in America." Political Science and Politics 28 (4): 664-83. . 1998. Foreword to "Social Capital: Its Importance to Housing and Community Development." Housing Policy Debate 9 (1): v-viii. Putzel, James. 1997. "Policy Arena: Accounting for the 'Dark Side' of Social Capital: Reading Robert Putnam on Democracy." Journal of International Development 9 (7): 939-49. Repetto, Robert. 1994. The "Second India" Revisited. Washington, D.C.: World Resources Institute. Rose, Richard, William Mishler, and Christian Haerpfer. 1997. "Social Capital in Civic and Stressful Societies." Studies in Comparative International Development 32 (3): 85-111. Serageldin, Ismail. 1996. "Sustainability and the Wealth of Nations: First Steps in an Ongoing Journey." Environmentally Sustainable Development Studies and Monographs Series 5. World Bank, Washington, D.C. QUALITATIVE ANALYSIS OF SOCIAL CAPITAL: ORISSA, INDIA 151 Singh, Mahip. 1995. "Study on Impact of Coal Mining on Tribal and Other Backward Communities." Coal India Ltd.,. Calcutta. Processed. S:iglitz; Joseph E. 1998 "More Tnstnumen-ts and Broader Goals: Moving toward the Post-Washington Consensus." The United Nations IUnivprsih-v World Inctihite for Dlevelonment Fconomics RPeParch Helsinki. Stolle, Dietlind, and Thomas R. Rochon. 1998. "Are All Associations AU,it2?" Amprir,nn Rohm7,inrnl Vrion+ictf A (1) A7SAr Tnrrn'-x, Sy-Ana-x 19QQ9 i qN42l,,n , Scianl cienc Wnre rrna c r no o Spae nd Time: A Critical Reflection on Robert Putnam's Making Democracy 141-k." Ame1Srican .00.iA;r7 rml wr 121'.viez1 90 (2): 38997. VPAASP (Vivekananda Palli Agragarmr. Seva Prat"i-hn). 1994A. "Resettlement and Rehabilitation Action Plan-Samaleswari OCP, x4_1,anaA, Coa-4h AS T 4- 11 C-nl TA;, + I q.X #. V- --Ar 10__A_ "_a_Econ Comic Study on ProjctJ Affected . 177TV. Vel ine SJocio LV I ILIt iJ-VJ V 1U t I 1tALt ;t Families of Samaleswari OCP-Mahanadi Coalfields Ltd." Coal India Ltd., .al t.i"a. Processed. 3aJcrren, Mark. D. 1998. S.AoJ.mlLuniy nu_L.AU -ar SJli,U;l Dower." American Behavioral Scientist 42 (1): 78-92. Whittington, Keith E. 1998. "Revisiting Tocqueville's America." American L)ftJLULI- JLtLL - .1L ). . J. VWVolcockU, IvMIlichael. 1970. "ocUial CapJit1 CrIdU 1LUI Lll ILL DeVeIUFprLLI. Towards a Theoretical Synthesis and Policy Framework." Theory and '0- 17 /I%\. 1 I 'l to Society 27 (2):. TA7T_ I~~ 101 fAl T 2:. T 'vAvUIIUdlBank. 1991. kJelloel d Poverty ir, nIr.da; Issues and Opportunities Concerning Women in the Indian Economy." Report 8072-IIN. V lvTvasingi(, Li.C. 1rocesseu. i996. "India: Coal Sector Environmental and Social Mitigation Project-Staff Appraisal Report." Report 15405-IN. Washington, D.C. Processed. 9. l97. "India: Coal Sector Rehabilitation Project-Staff Appraisal Report." Report 16473-IN. Washington, D.C. Processed. Annex 1 Instruments of the Social Capital Assessment Tool Thivus armlex presents the ,1 4-l txt- of the in er-.,iew guides and the ques- I 1L3 JULA ~CO MeL LILC ICe UCA L~I IC MtLJILVUI tionnaires of the Social Capital Assessment Tool. In addition, information is FluVILLdedLLaL th a Ut Ucanbe us ui for adaFt:. h- -l.n the SOCAT, such as a guide for selecting and training interviewers. The annex consists of the following sections: Page ~~ A ~ ~ ~~ -- J A ~~~~~- T ,f .J~ 1C 1A. Coml-l-unty PloitU andAsse,: Maippingr-In.erview Gjuide . .. 1Jio lB. Community Questionnaire . . ............................. 158 1C. Household Questionnaire . . .............................. 184 1D. Organizational Profile Interview Guides ................... 207 1E. Clrgni7zfinnql Prnfile Scroredieet . ------- 216 1F. Selection Criteria and Terms of Reference for Interviewers (Panama pilot test) ........ ................. 224 1G. Training Plan for Field Workers (Panama pilot test) ..... ..... 226 The CD-ROM enclosed with this book contains an electronic file of annex 1. Separate files of the five instruments of the Social Capital Assessment Tool (annexes 1A to lE) are also included. Each instrument is available as a PDF file and as a Microsoft Word 2000 file. The latter makes it possible for the user to adapt and modify the instruments as needed and to print questionnaires ready for use. To start the CD-ROM, insert it in your computer's CD drive. If an index file does not open automatical- ly, click twice on the "My Computer" icon on your desktop, select your CD drive, and click twice on the file "Index.htm." An instructional tool- kit for the application of the SOCAT is also available from the World Bank (contact information can be found at www.worldbank.org / socialdevelopment). 152 INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 153 Annex 1A LommL.unltly A roile A,xnd A,&sset Miappn1g1- Interview Guide The community profile is elicited through a series of group interviews conducted in the community during the initial days of field work. The community profile allows the research team to become familiar with community characteristics and issues relating to social capital for refer- ence in later phases of the data collection. The group interviews establish a consensus definition of the "community" in which the research takes place. This definition will be used throughout the community profile exercise and will serve as reference for the interviews of the household survey. It will also define the catchment area of institutions for the orga- nizational profile. Several participatory methods are used to develop the community profile. In addition to a focus group format, the data collection includes a community mapping exercise followed by an institutional diagramming exercise. The primary data source material generated by these interview- ing, mapping, and diagramming exercises are: * Community maps, indicating location of community assets and services * Observational notes of group process and summary of issues dis- cussed * List of positive characteristics of community assets and services * List of negative characteristics of community assets and services * List of all formal and informal community institutions * Case study of community collective action * Institutional diagrams (Venn) of relative impact and accessibility * Institutional diagrams (web) of institutional network relationships Between two and eight group interviews should be conducted in each community. Each group should have 5-12 participants. At least two group interviews should be carried out with women and men separately. Groups may be stratified on other sociodemographic characteristics that may be important within the community context, such as age or ethnici- ty. Mixed group interviews can also be conducted to assess levels of con- sensus, but these should be in addition to separate groups. Each group should have a moderator and two observers. The moder- ator's role is to facilitate the discussion, probe on key issues, elicit com- ments from all participants, and focus the discussion on the issues of interest without seeming to interrupt or ignore extraneous comments from participants. The observers' role is to take notes on the content of the discussion and process of group dynamics. 154 UNDERSTANDING AND MEASURING SOCIAL CAPITAL The team should have the following materials available: interview guide, pads of notepaper, writing pens, flip-chart paper, markers (sever- al colors), colored paper circles of different sizes, tape, scissors. 1. DEFINITION OF COMMUNITY AND IDENTIFICATION OF COMMUNITY ASSETS Bring large sheets of paper and several color markers. Ask the group to draw a map of their village or neighborhood that shows the settlement pattern, sites for productive activities, and locations of various assets and services in the commu- nity. A second group may be asked to make modifications to the map developed by the first group or, if they prefer, draw their own. The map is a key reference point for the discussion and should be used throughout the interview process to stimulate discussion, identify critical issues, clarify discussion points, and so on. 1.1 How do you define this village/ neighborhood? (Probe on geographical boundaries, place names, and other reference points. Establish consensus on the geopolitical definition of "communi- ty" for later use in the household survey.) 1.2 Where is/ are the... ...primary school? Secondary school? Childcare centers? Other schools? ...health services (both formal and informal)? ...sources of water? ...waste and garbage disposal sites? ...sources of electric lighting? ...public telephones? ...main streets/roads? ... principal means of transportation? ...markets, shops, and other commercial establishments? ...churches (places of worship)? ... cultural and recreational areas? (Where do you spend your free time?) ...areas that are less safe? RURAL:.. .irrigation systems? 1.3 How many years has this village/neighborhood been in exis- tence? Has the village/neighborhood grown, gotten smaller, or stavpd the same in the last five years-? Who are the people most likely to come into or leave the community? (In the case of significant in- or out-migration, have the group draw a Cecnlld ma7p CshoflwinY pfattePrns nfmigrnftinn new71 COsftlOment alnd pY"an- sion of community boundaries and land use.) INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 155 2. COLLECTIVE ACTION AND SOLIDARITY 2.1 People from the same village/neighborhood often get together to address a particular issue that faces the community, fix a prob- lem, improve the quality of life, or something similar. Which of the following issues has your village/neighborhood tried to address in the last three years? (Probe: education, health, public services, roads and transportation, markets, credit, recreational and cultural resources, security, child care, irrigation, agricultural services.) 2.2 Do you think that everyone in this village/neighborhood has equal access to _? (Probe: same services as mentioned under 2.1) Is this also true for the poorest members of the community? 2.3 Have there been any efforts by the community to improve the quality of the (service or benefit) or overcome a problem? Can you describe one instance in detail? (Refer to this case study for specifics of the following questions.) Were there community groups that played an important role? What kinds of responses did you get from the local government? From other organiza- tions? From the rest of the community? What kinds of obstacles did you have to deal with? What was the outcome of the effort? (Probe for locus of leadership, resources tapped, sources of resistance, who benefited or suffered from the outcome, the kind offollow-up that occurred as a result of the effort, and the mechanisms employed to ensure sustainability of the effort.) 2.4 Has this village/neighborhood ever attempted to make improve- ments but failed? Why do you think the attempt failed? What would you have done differently to make the effort more successful? (Probefor constraints on collective action; identify the roles of govern- ment, community organizations, and secondary institutions in influ- encing outcomes; and discuss the relationship between the community, representative organizations, local government, and other civil society actors.) 3. COMMUNITY GOVERNANCE AND DECISIONMAKING 3.1 Who are the main leaders in this community? (Probe formal and informal leadership.) 156 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 3.2 How do they become leaders? How are new leaders selected? 3.3 How are decisions made within this community? What is the role of the community leaders? How are community members involved? (Probe on role of traditional leaders, informal leaders, elites.) 4. LIST OF COMMUNITY INSTITUTIONS 4.1 What are the groups, organizations, or associations that function in this village/ neighborhood? Have the group list all the organizations,formal and informal, that exist in the community. Make sure all different types of organizations are included (agriculture, credit, religious, recreational, health, education, etc.) and that the list is as complete as possible. Have the group go through the list and identify which institutions are most important in meeting the community's needs. Make sure the list is written with plen- ty of space between each item. 4.2 Which groups play the most active role in helping improve the well-being of community members? 4.3 How did this group or organization get started (government ini- tiated, through government donations, NGO donations, grass- roots initiative, etc.)? 4.4 How are the leaders selected (election, appointment, inheri- tance)? How stable is the leadership (frequent or sudden changes, normal progressive change, or never changes)? Is lead- ership generally harmonious or conflictive? 4.5 How are decisions made within these groups or organizations? 5. RELATIONSHIPS BETWEEN ORGANIZATIONS AND THE COMMUNITY Venn diagram: Cut out (ahead of time) vaver circles of three different sizes and lay them out. Ask the group to place the largest circles next to the most important organizatinns.; the middle-sized circles next to the less important organizations, and the smallest circles next to the least imnnrtant orrOni7ntionn Write the name nf the organni7Otinlf in each - --- 0-6----------01--- -. circle. Observers should record the group's reasoning as to why organi- zatinotc nre cafePgnri7pd as mnre nr less ininartnnt INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 157 Draw a relatively large square in the center of theflip-chart paper. Tell the group that this square represents themselves. Have tne group place the organization-labeled circles in or around the square at the center. The closer they are to the center square, the more accessible the particu- lar organization is to the commnunity. Let the group discuss among themselves and facilitate as necessary. Record the resulting diagram and reasoning behind the group's discussion on each organization. 5.1 Of the organizations on this list, which are most important? Which are least important? Which are of medium importance? 5.2 Of the organizations on this list, which ones are most accessible to the community? Which are least accessible? Which are some- what accessible? 6. INSTITUTIONAL NETWORKS AND ORGANIZATIONAL DENSITY Flowchart diagram: Have ready a sheet of flip-chart-sized paper and markers. Facilitate a discussion among the group regarding the rela- tionships among the identified organizations, community leaders, and the community. Probe on local government institutions, nongovern- mental organizations, base organizations, and other civil society actors. Ask the group to draw each actor and, using arrows or other appropri- ate symbols, indicate the relationship among them. Probe links among all organizations. 6.1 Which organizations work together? How do they work togeth- er (hierarchically, collaboratively)? 6.2 Are there any organizations that work against each other (com- pete or have some sort of conflict)? Which ones and why? 6.3 Some groups may share the same members and some groups have different members. Which organizations have the same or similar membership? 6.4 Are there organizations that share resources? 158 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Annex 1B Co l. a .unitIyr Que. stionnai:e r 1. COMMUNITY CHARACTERISTICS l.i How many years has the community been in existence? More than20 years [ ] 1 Between 10 and 20 years [ ] 2 Fewer than 10 years [ 1 3 1.2 How many households are in this community? Fewer than 25 [ ] 1 Between 25 and 49 [ ] 2 Between50 and 99 [ ] 3 Between 100 and 249 [ ] 4 More than 250 [ ] 5 1.3 In the last three years, the number of people living in this com- munity has: Increased [ ] 1 Decreased [ ] 2 Remained the same [ ] 3 1.4 What are the two main reasons for the increase, decrease, or lack of change? (a) (b) 1.5 What are the two principal economic activities for men in this community? (a) (b) 1.6 What are the two principal economic activities for women in this community? INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 159 (a) _ (b) 1.7 In the last three years, availability of employment has: Improved [ ] 1 Worsened [ ] 2 Remained the same [ 3 1.8 What is the main route that inhabitants use to reach this commu- nity, both during rainy season and dry season? (a) Rainy (b) Pry Paved road [ ] 1 [ 1 Dirtroad [ ] 2 [ 1 2 Mixed paved and dirt [ ] 3 [ ] 3 Footpath [ ] 4 [ ] 4 Horse trail [ ] 5 [ ] 5 Sea [] 6 [] 6 Other (specify) [ ] 7 [ ] 7 1.9 In the last three years, the roads leading to this community have: Improved [ I 1 Worsened [ ] 2 Remained the same [ ] 3 1.10 The availability of housing in this community is: Adequate [ ] 1 Deficient [ ] 2 1.11 In the last three years, the quality of housing in this community has: Improved [ ] 1 Worsened [ ] 2 Remained the same [ l 3 1.12 What are the two main reasons that housing in the community has improved, worsened, or remained the same during the last three years? 160 UNDERSTANDING AND MEASURING SOCIAL CAPITAL (a) (b) 1.13 In the last three years, the overall quality of life of the people liv- ing in this community has: (consider job availability, safety and secu- rity, environment, housing, etc.) Improved [ ] 1 Worsened [ ] 2 Remained the same [ ] 3 1.14 What are the two main reasons that the quality of life in the com- munity has improved, worsened, or remained the same during the last three years? (a) (b) 1.15 Overall, the level of living of this community may be character- ized as: Wealthy [ ] 1 Well-to-do [ ] 2 Average [ ] 3 Poor ] 4 Very poor [] 5 1.16 Do people in this community generally trust one another in mat- ters of lending and borrowing? Yes [ ] 1 No [ ] 2 1.17 In the last three years, has the level of trust improved, worsened, or stayed the same? Improved [ ] 1 Worsened [ ] 2 Remained the same [ ] 3 INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 161 1.18 Compared with other communities, how much do people in this community trust each other in matters of lending and borrowing? More trust than in other communities [ ] 1 Same as in other communities [ ] 2 Less trust than in other cornmunities [ 1 3 1.19 Do you agree or disagree with the following statement: People here look out mainly for the welfare of their own families and they are not murh connerned with rommiunity welfare. 'trongly agree [ I Agree [ ] 2 Dlisagreep [ 1 3 Strongly disagree [ ] 4 2. PRINCIPAL SERVICES 2A. Electricity 2A.1 What fraction of the community has household electrical service? The entire community [ ] 1 Mfost of the coAmmuit- r I About half the community [ 1 3 Less thar. half /I veny fFeW r 1 A No one in the community [ ] 5 (go to section 2B) 2A.2 In the last three years, the electrical service to this community ILhas: 1 u L~~~~~~~~ J1 iImproved [ ] 1 Worsened [ ] 2 2-A.3 CurrIn,ltLy, LIth quality V1 electLicdl serviLe WILItIIll the Il0llIeS of this community is: Very good [ ] 1 Good [ j 2 Average [ 1 3 Poor [ j 4 Very poor [ ] 5 162 UNDERSTANDING AND MEASURiNG SOCIAL CAPITAL 2A.4 What are the two main problems with the electrical service? (a) (b)_ 2B. Public Lighting 2B.1 Does this community have street lights? Yes [ I 1 To r I 2 (go to sc INU ~~~~~~~L J 4. Ik LU DCL1UVt 4.2 Ul , t t rU eeLL LIUC yeaCri, LItC pubIIL 116L1LUtr, sIvi hLCsL. HlIFIU.VC:U L I l Worsened [ 1 2 T) :__ 1 3 A_ _-- r I n ixemHuainec utef samte, 2B.3 C-Urrentl-y, Lite quality of p-ubiuc ligling service is: very good [ 1 ; Good [ ] 2 Average 3 ] 3 Poor [ ] 4 'Very poor [ j 5 2B.4 What are the two main problems with the public lighting in this community? (a) (b) 2C. Drinking Water 2C.1 What part of the community has pipe-borne water? The entire community [ ] 1 Most of the community [ ] 2 About half the community [ 1 3 Less than half/ very few [ ] 4 No one in the community [ ] 5 INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 163 2C.2 What part of the community has access to public standpipes? The entire community I ] 1 Most of the community [ 1 2 About half the community [ ] 3 Less than half/verv few [ 1 4 No one in the community [ ] 5 2C.3 In the last three years, potable water service has: Improved [ ] 1 Worsened [ 1 2 Remained the same [ ] 3 2C.4 Currently, the potable water service is: Very good [ I 1 Cnod [ 1 9 Average [ ] 3 Poor [ ] 4 Very poor [ ] 5 2C.5 What are the two main problems with the potable water service? (a) _ (b) 2D. Home Telephonie Service 2D.1 What fraction of the community has home telephone service? The entire community [ ] 1 Most of the com., mun,it;Y r I 2) About half the community [ ] 3 L,sO .... ..hlfI very fevv [ ] I No one in the community [ ] 5 (go to section 2E) 2D.2 Currently, the home telephone service is: Very good [ ] 1 Good [ j 2 Average [ ] 3 Poor [ ] 4 very poor [ ] 5 2E ('nsirunioatin qornioc 2E 1 Tonies this cnmmiiihf v ixp biiihlic telephones Y-es r I 1 No [ ] 2 (go to question 2E.3) 2E.2 How many public telephones are in this community? (go to question 2E.4) 2E.3 What is the distance from this community to the nearest public Dsistance (in waLVai 1g .it) 2E.4 Ir. the lashe years, -thpublic tlpoesevcnthsCAm ')U A T-~ j--., LA~~*~~ j~11~ "t£1 UIC la, L ILLCC Y~_CUZ, LliI: FyUJU%LL1C1AJL M:C Ll V I. Ilt LIU. COMI munity has: Improved [ ] 1 Remained the same [ ] 3 2E.5 Currently, the public telephone service in this community is: Very good [ ] 1 Good [ j 2 Average [ ] 3 Poor L i 4 Very poor [ ] 5 2E.6 What are the two main problems with the public telephone ser- vice in this community? (a) (b) 2E.7 Is there a post office in this community? INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 165 Yes [ ] 1 No [ ] 2 (go to question 2E.9) 2E.8 What is the distance from this community to the nearest post office? Distance (in walking minutes) 2E.9 In the last three years, the mail service in this community has: Improved [ ] 1 Worsened [ ] 2 Remained the same [ l 3 2E.10 Currently, the mail service in this community is: Very good [ 1 1 Good []2 Average [ ] 3 Poor [ ] 4 Veryvpoor [ 1 5 2E.11 What are the two main problems with the mail service in this community? (a) _ (b) 2E.12 What fraction of the commtnity has access to public Internet ser- vice? The entire commiunity [ 1 1 Most of the community [ ] 2 About half the community [ ] 3 Less than half/ very few [ ] 4 No one in Fhe community [ ] 5 (go to question 2E.14) 2E.13 Where are public Internet access services available? Local school [ ] 1 Library [ ] 2 unir . murut; ce n t e r [ ....r 1 3 %..JiiLit L LLy A~i LLIL J ~- 166 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Training center [ ] 4 Internet cafe [ ] 5 Other (specify) [ ] 6 (go to section 2F) 2E.14 What is the distance from this community to the nearest public Internet access service? Distance (in walking minutes) 2F. Sewage 2F.1 What fraction of the community is served by a public sewage system? The entire community [ 1 1 Most of the community [1 2 About half the community [ ] 3 Less than half / very few I l 4 No one in the community [ ] 5 (go to question 2F.6) 2F.2 In the last three years, the quality of the public sewage system in this community has: Improved [ ] 1 Worsened [ ] 2 Remained the same [ ] 3 2F=3 Currently; the public sewage system is: Very good [ 1 1 Good [ ] 2 Average[ Poor [ 1 4 V7ery poor I ] 5 2F.4 A AWhrf are the hvo0. r,n rn prolemc lXAitil theo public seW^7age system in this community? (a) (b) INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 167 2F.5 Do the streets of this community have sufficient sewers and drains to handle excess water and prevpnt flooding w.hen it rnins? Yes ~~~~~[ ] 1 No [ ] 2 2F.6 What other sewage and waste water systems are used in this Communify? a. Latrine [ ] 1 [ ] 2 b. Sep"icrtnnk.s [ I 1 [ 1 2) c. River or sea [ I 1 [ ] 2 d. Other (specify) [ ] 1 [ 1 2) 2 G. G_arbage Co -e.X-r 2G.1 Wh^,.at fractLion of the com.mu.uri4y is ser3 v3 d by a gabg --IIec.L; service? The entire community [ ] 1 MVIos,t of 4the com.ULUML-U; [ I 2 About half the community [ 1 3 e cI t -acIxxa l1/ Vcy LCvV [ ] 4T No one in the community [ ] 5 2G.2 In the last three years, the quality of the garbage disposal in this Colrrlul-u1y hLas: Worsened [ ] 2 IWeICIauuu LItI ,anlte [ j 3 2G.3 . In ute hitmies mta udo inot receive garbage collection service, wnat is the main solid waste disposal method? Burn it [ ] 1 Throw on own lot [ j 2 Throw on others' lots [ 1 3 Throw into river/ sea [ j 4 Bury it [ ] 5 Pay to naul away [ ] 6 Other (specify) [ ] 7 168 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 2H. Public Market 2H.1 Does this community have a public market? Yes [ ] 1 (go to question 2H.3) No [ ] 2 2H-.2 The war2lkingf diistancfro rnr. t-he rnnni,ih iy t-oth nearest- markt-i is: Distance (in walking minutes) (go to section 21) 2H.3 The market is open: Every day [ 1 Com.e Aays ofl 1-VeV week [ I 2. One day per week [ 1 3 C-Ulher (speciay [l ] 4%r S-/Ut I] \ CLIJy) L I H.4 I las ULC ICILt LIL ycars, tLh qLuIIliLy aniLD1 ViCe of LisL 1rr.airt I Las. T ~~~~~~~~r 1 1 ' ImplUvedt [ ] 1 Worsened [ 1 2 Remained the samne [ ] 3 Hr.5 rHow mriaiy people in the CuLiiuIiLty use the IltdfKel? Tne entire community [ I 1 Most of the community [ ] 2 About nalf tne community [ j 3 Less than half/very few [ ] 4 21. Transportation 21.1 Is this community served by a public transport system? Yes [ ] 1 (go to question 2I.3) No [ ] 2 21.2 The walking distance to the nearest community with public transportation is: Distance (in walking minutes) (go to question 2I.7) INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 169 21.3 Public transportation is available: Every day [ ] 1 ~UIL~ QYD L. -[ I J'.JII d~%ays of the L VVe [ ] 2 One day per week [ 1 3 O.er (spec:.y r ] A '.TJA T- LI--I - ns 2.I.4± IIL ULh lasL LtILr yYCal, the qualIt- andU sri ofL ubi tran s portation has: Improved [ ] 1 vAITrseneU [ r 2 Remained the same [ ] 3 21.5 Public transportation is used by: The entire community [ ] 1 Most or the community [ j 2 About half the community [ ] 3 Less thanhalf/very few [ 4 No one in the community [ ] 5 2I.6 What two main changes can be made to improve public trans- portation to this community? (a) CD) 21.7 What other types of transportation do people in this community use to go to neighboring communities? (List the two most impor- tant ones). (a) (b) Walking 1 Bicycle 2 Horse 3 Canoe/boat 4 Car 5 170 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 21. Recreation 2J.1 Does this community have sports fields or recreational areas? Yes [ ] 1 1iT r I 1 '/IT OX 1NU 0 LI ( - IgoU to IuesLIU LnJ.J) 2-J .2 ll ,t las.IC L tfLtf yeCalr, LI .It LiIUo LIoI t hie LI pO fULte ldsU ad letILt- ational areas has: Improved [ ] 1 vvorseneu [j2 Remained the same [ ] 3 2J.3 Does this community have separate children's play areas? Yes [ ] 1 No [ j 2 (go to section 2K) 2j.4 In the iast three years, tne condition of tnese chiidren's piay areas has: Improved [ ] 1 worsened [ ] 2 Remained the same [ ] 3 2K. Security 2K.1 Does this community have a security or police force? Yes [ ] 1 No [ ] 2 (go to section3) 2K.2 This service is provided by: The police [ ] 1 The community [ ] 2 A private company [ ] 3 2K.3 This security service is provided to: The entire community [ ] 1 Most of the community [ ] 2 INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 171 About half the community [ ] 3 Less than half/ very few [ ] 4 2K.4 In the last three years, the quality of the security service has: Improved [ ] 1 Worsened [ ] 2 Remained the same [ ] 3 3. LABOR MIGRATION 3.1 Are there members of this community who go to other places to work during certain periods of the year? Yes [ ] 1 No [ ] 2 (go to question 3.6) 3.2 Do more women than men leave to work? Do more men than women leave to work? Or equal numbers of women and men? More women than men [ I 1 More men than women [ ] 2 Equal numbers [ ] 3 3.3 Where do they go to work primarily? To a city in this region [ ] 1 To a city in another region [ ] 2 To a city in another country [ ] 3 To a rural area in this region [ ] 4 To a rural area in another region [ ] 5 To a rural area in another country [ 1 6 3.4 What are the two principal jobs women leave for? (a) (b) 3.5 What are the two principal iobs men leave for? (a' ~ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 172 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 3.6 Are there people from other communities who come to work in this community? Yes [ ] 1 No [ ] 2 (go to section 4) 3.7 What are the two principal jobs they come for? (a) (b) 4. EDUCATION 4A. Preschool 4A.1 Does this community have a public preschool? Yes [ ] 1 (go to question 4A.3) No [ 1 2 4A.2 How far from the community is the nearest nublic nreschool? Distance (in walking minutes) (go to section 4B13 4A.3 Is the number of nreschools in this commlinitv sufficient tn serve the number of young children in the community? Yes [ I 1 No [ ] 2 4A.4 IsT the niimh*r nf t-ercherc in these nrecrkolsc csui.frien for the number of children? Yes [ ] 1 No [r 1 AA ~~~~~~~~~~~ 4A5 -se pthy sical condit-on, of thle pecolias: v7ery gooA r I 1 %A- ~.U L I I Good [ ] 2 Average L[ I Poor [ I 4 Very poorr I ] vt:1-y pUUi. [ I J INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENTT TOOL 173 4A.6 What percentage of young children attend public preschools? All children [ I 1 (go to section 4B) Most children [ 1 2 About half of the children [ ] 3 Less than half [ 1 4 Very few/none [ ] 5 4A.7 What are the two principal reasons that young children from this commuinity do not attend uiblir nreschool? (a) aB. Pr4tma" rchonnl 4B.1 Doecs this r-rO-.nitY hwave a public rimnruT choo!? Y.es [ ] 1 (goto quesaon 4B.3) No [ 2 4B.2 How far from the commtnity is the nearest public primary D"istance (in w^al"ing M. inte)m1+-1scH;n C serve the number of school-age children in the community? Yes [ ] 1 1sT,-, I 1 ' No L ] I In.' 1 LIMt ILULIIUt:1 VI ItLIUCILt:1 III L1MLt:2t: 2LLUVI2 32UI1%1t1,:LL LIL LIMt ILUIIL- ber of students? Yes [ ] 1 1T, T r I 'n INo [ ] 2. 4B3.5 T he pnysicaI conuition oi tne pririiary schioi is: Very good [ j 1 Good [ ] 2 Average [ j 3 174 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Poor [ ] 4 ClJery poor [ ] AB.6 X.Alhat percent-age of efigib-Ale sho-g hlrnaedpbi primary schools? All children [ ] 1 (go to section 4C) IVIVO L tA I%A1 VI t 1I~ About half of the children [ ] 3 Tess ~1_an I-a 1^_ r 1 A Lt:D: LitAl I L 11 L I Very few/none [ ] 5 4B.7 What are the two principal reasons that school-age children from tuhis commulnuity do not attend public primrary school? (a) (o) 4C. Secondary School 4C.i Does tnis community nave a public secondary scnooi? Yes i (go to question 4C.3) No [ ] 2 4C.2 How far from the community is the nearest public secondary school? Distance (in walking minutes) (go to section 4D) 4C.3 Is the number of secondary schools in this community sufficient to accommodate the number of secondary-school-age students in the community? Yes [ 1 1 No [ ] 2 4C.4 Is the number of teachers in the secondary schools sufficient for the number of students? Yes [ ] 1 No [ ] 2 INSTRUMENTISN OFr THE SOCIAL CAPITAL ASS-OES3OSMENT TOOL 175 4C.5 The physical condition of the secondary school is: Very good [ ] 1 L~ I Average [ ] 3 Poor [1 A Very poor [ ] 5 4C.6 What percentage of secondary-school-age children attend public secondar,y schools? All Ghidre [ ]AT(o-oIeci n4D Most children [ 1 2 AoUL t hlf tU. LAIILIAICIl [ I 3 Less than half [ ] 4 VCry itvv / none [ I 4C~.I VVILL LI are thLe tvvu pinucilF rCaonsII that seoltdaly-schoIl-age children from this community do not attend public secondary (a) 4 D. Adult L-UUGUL.JOn Ar'. r T .1 _ .1. _____ __ __ ___- _ .~1_ - ______' 4LJ.1 is Umere an dautt uteracy camIpagn or programt 1or UIe collunlLuty f Yes [ I 1 No [ ] 2 4D.2 Are there job training programs for this community? Yes [ ] 1 No [ j 2 5. HEALTH 5.1 What are the three principal health problems affecting children under six years of age in this community? (a) 176 UNDERSTANDING AND MEASURING SOCIAL CAPITAL (b) (c) 5.2 What are the two principal health problems affecting adult men in this community? (a) (D) 5.3 What are the two principal health problems affecting adult women in this community? (a) (b) 5.4 Does this community have a health clinic or hospital? Yes [ ] 1 (go to question 5.6) No [ ] 2 5.5 How far is the nearest public health clinic or hospital? Distance (in walking minutes) (go to question 5.7) 5.6 Does the health clinic or hospital regularly have sufficient: Sufficient Insufficient None a. Basicmedicines [ ] 1 ] 2 [ ] 3 b. Equipment/instruments [ ] 1 [ ] 2 [ ] 3 c. Patientbeds [ ] 1 [ ] 2 [ 1 3 d. Ambulances [ ] 1 [ ] 2 [ ] 3 e. Physicians [ ] 1 [ ] 2 [ ] 3 f. Nurses [ ] 1 [ ] 2 [ ] 3 g. Otherhealthstaff [ ] 1 [ ] 2 [ ] 3 5.7 Does this community have a family planning program? Yes I] 1 No [ ] 2 (go to section 6) INSTRUTMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 177 5.8 Who offers the program? Government [ ] 1 NGO [ 1 2 Private facility [ ] 3 Other (specifv) [ 1 4 6. FNVTURANTM FNTAT TISSUES 6.1 Does this co.m .....munit5y have: Yes NTn a. Garbage dumping that contan. mirnates rivers or wells [ 1 1 r 1 2 I. J L J b. Garbage dumping that contaminates-he oceanr 1 1 r] 2 c. Junk yards or scrap heaps [ 2 1 [ ] 2 d. Staindin`g water or cstagn-r.t pools [ ] 1 [ ] 2 e. Slugtehose A.- dm waste in public places [ ] 1 [ ] 2 f. Mech-ics who AU.Mp waste oil in soil or water [ ] 1 [ ] 2 T31u -A.-4-..- r 1 1 r I1 ) g. IJ11 lU.ng LlItC i es [ ] I [ ] h. Clear-cutting or forest burns [ ] 1 [ ] 2 A If., I ~~ ~~~r I 1 r 1 '*) i.IVirLU, E r 1 [ ] 2 j. Other (specify) [ ] 1 [ ] 2 6.2 Overall, the current environmental condition of the community is: Very good [ ] 1 Good [ ] 2 A r I 1 ' AXverage [ ] 3 Poor [ ] 4 very poor [ I 5 6.3 mn tne Iast three years, the environmental conditions in ute coin- munity have: Improved [ ] 1 worsened [ j 2 Remained the same [ ] 3 178 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 6.4 What are the two main actions that could be taken to improve the (a) (b~) 7. AIGR.CLULIURE (ol Ly ILn IrIal areas '7 1 TAI 1 .1 VVlILaIa are LiUe Lhree FUiuL%ipal cu6IILUtural oJr lVLesoc acLIVILXi0 undertaken in this community? (a) (b) (c) 7.2 Where do the inhabitants of this community generally sell their livestock and produce? (List up to three venues by order of impor- tance.) a b c I I l I Community market 1 Market in neighboring areas 2 Domestic middlemen 3 Exporters 4 Public institutions 5 Cooperatives 6 Local stores and shops 7 Other (specify) 8 Only self-consumption/ no outside sales 9 (go to question 7.4) 7.3 What are the two most important problems facing members of this community for getting their products to the market and earning a profit? (a) (b) iNSTRUMENTS OF THE SOCiAL CAPITAL ASSESSMENT TOOL 179 7.4 Do the agricultural workers/producers in this community Yes []11 No [ 1 2 (go to question 7.6) 7.5 Who is the main provider of this technical assistance? (Probe whLeI!t ilt; irzlstttutlu!l is JLUlL Ut pi LUaLt.) 7.6 uoes utis coLrtU11uULy hI'ave afny ype of agricultural CoUperudUVe? Yes [ j 1 No [ ] 2 7.7 Does this community have any institution or person (either in the community or nearby) that provides credit and loans to agricul- tural producers? Yes [ ] 1 No [ L 2 (go to question 7.9) 7.8 vvhat are the three main persons or institutions that provide credit or loans to agricultural producers in this community? a b c National banks 1 Agricultural / development banks 2 Private banks 3 Agricultural credit unions or cooperatives 4 Private individuals 5 Export businesses 6 Packing businesses 7 Producer associations 8 Warehouses or middlemen 9 Other (specify) 10 180 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 7.9 Do the agricultural producers of this community receive loans or credits from individuals or institutions in other cities or regions? Yes [ ] 1 No [ ] 2 7.10 What percentage of the agricultural producers in this communi- ty use loans or credits to support their activities? 7.11 What are the two principal problems facing the agricultural pro- ducers of this community in terms of receiving loans and credits? (a) _ (b) 7.12 In the last three years, the harvests/ yields have: Increased [ I 1 Decreased [ ] 2 Remained the same [ ] 3 7.13 In the last three years, the sales of agricultural/livestock prod- ucts in this community have: Increased [ ] 1 Decreased [ ] 2 Remained the same [ 1 3 8. COMMUNITY SUPPORT 8.1 Which of the following organizations exist in this community? Yes No a. Community development committee [ ] 1 [ ] 2 b. Cooperative (fishing, agriculture, crafts) [ 2 1 [ ] 2 c. Parent-teacher association [ I 1 1 1 2 d. Health committee [ ] 1 [ ] 2 e. Youthgroup [ ] 1 [ ] 2 iNSTRUMENTS OF THE SOCIAL CAPiTAL ASSESSMENT TOOL 181 f. Sports group [ I 1 [ ] 2 g. Cu--ltal group [ ]1 [ I 2 h. Civicgroup [ ] 1 [ 2 rl spe,d,\ r 1 1 r 1 o I.. ILItWL UIL a I ICI Ul OUF JL ULLO J f I Ly based organizations? Yes No T_---1 r I 1 r 1 n a. Local UVl L J I I I b. National government [ I 1 [ ] 2 c. P -olitic [ 1 1 [ 1 2 d. ReLigious organizations [ I 1 [ 1 2 e. Scnool/teachers [ ] I [ ] I f. Nongovenimental organizations [ I 1 [ ] 2 g. Business group [ ] 1 [ ] 2 h. ServiceC;UD [ c l [ I 2 i. Prosperous citizens [ ] 1 [ ] 2 j. The community asawhoie [ w i [ j 2 8.3 'What buildings do people in t'his community reguiarly use for meetings and gatherings? Yes No a. Community center [ 2 1 [ j 2 b. Personal homes [ ] 1 [ 1 2 c. Homes of political leaders [ ] 1 [ j 2 d. Homes of other local leaders [ ] 1 [ ] 2 e. Churches or religious buildings ] 1 [ ] 2 f. Health center/ school 2 1 1 [ ] 2 g. Government buildings [ ] 1 [ ] 2 h. Business/ commercial buildings [ 2 1 [ ] 2 i. Other (specify) I ] 1 [ ] 2 8.4 Which members of the community participate most in solving the issues facing the community? (a) By gender Men [ ] 1 Women [ ] 2 182 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Men and women equally 1 ] 3 lNeither pari.iL-pa.e ] to) By age Youth and adolescents [ ] 1 Adtults I I Older persons [ ] 3 Youth, adults, and elders equally [ ] 4 None participate [ j 5 (c) By employment status Workers [ ] 1 Unemployed or nonworKers [ 1 2 Workers and nonworkers equally [ j 3 Neither participate [ ] 4 8.5 In the last three years, has the community organized to address a need or problem? Yes [ ] 1 No [ ] 2 (go to question 8.8) 8.6 Around what issue(s) did the community organize? (a) (b) 8.7 Was/were the initiative(s) successful? Yes No Ongoing a. Initiative #1 [ ] 1 [ ] 2 [ ] 3 b. Initiative #2 [ ] 1 [ ] 2 [ ] 3 8.8 What are the two main problems or needs that community mem- bers feel must be addressed or solved? (a) (b) INSTIUMENTS OF --E SOCIAL CAFITAL ASSESSMENT TOOL 183 8.9 Are there any specific assistance programs to this community? Yes [ ] 1 No [ ] 2 (go to questlion 8.11) 8.10 What are the twA.o main prograrm.s and the instiftutions hat sup- port them? (a) Program/ institution (b) Program/ institution 8.11 Do any of the following problems exist in this community? If yes, vvnno isn~ t litoMt affcte. or at-risk gr up y agC, 6CgLenC, caste, ethnic group, etc.)? Most C1[1t:LLC,U Yes No group a. Burglaries [ ] 1 [ ] 2 L. DLl~~2-.,-. r I I r I n bi. Robber~ies [ ] 1 [ ] 2 ____ c. Assaults [ I 1 [ ] 2 a ,-ang r 1 1 r I n u %-3 18 I j I I I 2- e. Vandalism [ ] 1 [ ] 2 f. Violent disputes [ I [ I 2 g. Alcohol abuse [ ] 1 [ ] 2 II. u-usiailce (urug) abuse [ I 1 [ i 2 i. Teen pregnancy [ ] 1 [ ] 2 j. Domestic violence [ j 1 [ j 2 k. Child abuse [ ] 1 [ ] 2 ;. Prostitution [ j 1 [ j 2 m.Other problems (specify) [ ] 1 [ ] 2 184 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Annex 1C Household Questionnaire L.engsth of ilerv-ew Timl le " .-ated: Time terminated: 1. IDENTIFICATION OF SELECTED HOUSEHOLD: 1.1. Province / state 1.2. District 1.3. Subdistrict 1.4. Town/village 1.5 Type of area: Urban [ ] rsural nonindigenous [ ] Indigenous Difficult access [ 1.6 Location: Unit Number 1.7 Address of selected household: Community Street House number Other details Interviewer: Supervisor: 2. HOUSING CHARACTERISTICS AND HOUSEHOLD ROSTER 2.1 Type of house (observation only) Individual house [ 1 1 Open roof and patio [ ] 2 Apartment [ ] 3 Room within a larger house [ 1 4 INSTRUMENTS OF THE SOCIAL CAPr11AL ASSESSMENT TOOL 185 Other (specify) [ ] 5 2.2 What construction material is used for the majority of the exteri- -or 1 wa!-f 4the house orbilig -.t -.AL - I ~.f.. . ..IJ 1t -0LJ Ci-d erblock / brick / stone/ concrete/ cement [ ] 1 Fiberglass [r 1 2 Wood [ ] 3 AAdtoe/wattle. ean dAu [ A A Cane/straw/sticks [ ] 5 Im valls L[ ] 6 Other (specify) [ ] 7 2.3 What is the construction material of most of the roof of this mLuu ? hlouse. ' . . I r 1 1 .oUIInLcrte/ L/IclmetL I J 1 Tiles [ ] 2 Meial (zinc, alumitLtUmll, etc.) [ J 3 Wood [ 1 4 Stiraw or ithatd [ I 5 Other (specify) [ ] 6 2.4 What is the construction material of most of the floor of this nouse? Concrete/cement [ j 1 Tiles, brick, granite [ ] 2 'vvood 3 Vinyl [ ] 4 Earth, sand [ j 5 Cane [ ] 6 Other (specify) [ j 7 2.5 How many rooms are used by th is housenoid for sieeping only? 2.6 What type of sanitary services does this household use? Connected to sewage system [ ] 1 Connected to septic tank [ ] 2 186 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Latrine [ 1 3 114 L~~~~~~~~ 1 A2 Other (specify) [ 1 5 2.7 What is the primary source of water for this household? Piped water system [ ] 1 'n -1 ~~ ~~~r I n rrivate-wwell [ I Public well [ ] 3 Open tap or faucet [ I 4 River or stream [ ] 5 Other (specify) [ i 6 2.8 How uoes this nousenoId dispose or most or its garbage? Public garbage service [ I 1 Private garbage service [ ] 2 T nrow in vacant iots [ j 3 Throw in river, stream, ocean [ ] 4 Burn and/or bury [ j 5 Other (specify) [ ] 6 2.9 What type of lighting does this household use? Electricity (public source) [ I 1 Electricity (private source) [ ] 2 Electricity (combination public and private) I 3 Only kerosene, gas, candles [ ] 4 Other (specify) [ ] 5 2.10 This home is Owned and completely paid for [ ] 1 Owned with a mortgage [ ] 2 Rented [ ] 3 Given in exchange for services [ ] 4 Squatter [ 1 5 Other (specify) [ 1 6 NOTE: * List all the people in the household)first and then ask questions 2.12 to 2.20. Trhe household is dlefinedI as all the people usually living together in this dwelling and sharing expenses. 2.11 List names of all individuals iin household. (List hotusehold headfirst, usefirst names only.) 2.12 What is " "'s relationship to household head? (Use code box on the next page.) 2.13 Sex (male = 1, female =2') 2.14 How old is " "? (years) 2.15 What is " "'s marital status? (married = 1, coimmon law = 2, divorced = 3 [go to 2.17], widow(er) = 4 [go to 2.171, never married = 5 [go to 2.171) 2.16 Is ' "'s spouse currently -a member of the household? If yes, use number of spouse. If'no, write 99 2.17 Occupation (Use code box on the next page.) 2.18 Currently employed? (yes = ], no == 2) 2.19 Complete education level? (Use coade box on the next page.) 2.20 How long have you lived in this community? (years) 2.11 2.12 2.13 ,2.14 2.15 2.16 2.17' 2.18 2.1'3 2.20 01 _ ___ __ 02 03 04 05 06 07 08 09 10 188 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Code box for question 2.12 Head 01 Grandchild 08 Uncle /aunt 14 Wife/husband 02 Grandparent 09 Cousin 15 Son! daughter 03 Father-in-law! Other relative 16 Father/mother 04 mother-in-law 10 Children from Sister/brotlier 05 Son-in-law! another family 17 Stepson! daughter-in-law 11 Other relative 18 stepdaughter 06 SRister-in-law Renter 19 Stepfather brother-in-law 12 Other nonrelative 20 cI temofther 07 Nepher^. / niece 13 | ~~~~~Code box for questi;on 2.17 Farmer ; Private sector: Fisherman 2 Unskilled 6 Trade 3 Skiiled 7 Manufacturing: Public sector: Artisan 4 Unskilled 8 Industrial 5 Skilled 9 Other 10 Code box for question 2.19 Illiterate, no schooling 1 Literate, no schooling 2 Primary incomplete 3 Primary complete 4 Secondary incomplete 5 Secondary complete 6 Vocational college 7 University 8 Other 9 INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 189 3. GENOGRAM (Using symbols below, record here family composition, household com- position, organizational affiliation, and level of involvement. An exam- ple is on the next page.) Genogram symbols nli = male jjjj = male informant = deceased male (Th = female = female informant M = deceased female = = legal marriage b.19_ date of birth common-law marriage = divorced d.19_ date of death dissolution of union Composition of households living in the same communitv as the informant's household *-omposition of nousenoids iiving in a community *--- ..... --:different from that of the informant 190 UNDERSTANDiNG AND MEASURING SOCIAL CAPI1AL Example ii -V W T I This example represents tne situation of a divorced woman whose household includes her three children and granddaughter. In a house- hold within the same community live her mother (a remarried widow); her stepfather; her half-sister, currently separated from her husband; and her niece. Another half-sister died some time ago. The respondent's for- mer husband resides in another community. His household consists of his common-law wife, their son, and the respondent s nephew. The respon- dent's brother lives in common-law union with two sons and a child, a girl, who is not a blood relation. INSTRUMENTS OP THE SOCIAL CAPiTAL ASSESSMENT TOOL 191 4. STRUCTURAL SOCIAL CAPITAL Now I would like to ask you some questions about how you feel about fthils -4lg/.igbrod and -ow you taepaf. tecmmn tyf4I4U~ tLL~tVr ftJ I4 4.-t -_J yR.)1 R.IN~ lt &R.L ~Ut'lrlflt"4111y activities. By community, I mean l_ [insert consensus defin- itionf,rom commun.ity prfi. A1 Al. O01 ,.izati,o,, 'Den,si- a,,d (IL etr,6 A. 1 A A - -- yu - 21.ICf1 YUU Utl 1j~ DV11MtU1L: UI YuuiJ- ILUUZICIU1U al 11MULUCI L) Uty C1t, 1ULt}.I, organizations, or associations? (Probe: Who in the household belongs to W!IILfd group?UJ A,-e t1e ufty utL=I X,fOups ut orjur ri-uL UbUL1Ustiorlsb .lhUa yUU or someone in your household belongs to? Code below and record on genogfram.if the household is not a member in ariy group, go to section 4B.) A A TR.. .-… 1 -ILUv2 1 1 - .1- - 4A.2 DJo you consiuti yoursenl/ ioushIoiu iiieiiioer to ve active in tne group, such as by attending meetings or volunteering your time in oLier ways, or are you relanvely mactve? Are you/ nousenold member a leader in the group? Household| Name of [ Type of 1 Degree of 1 member organization organization participation (use roster (use codes (use code coae) t elow) betow) ! _ _ ! __ _ I _ __ ! !Farmers, /fishermen's Type of organization | Farmers' jistisermen's group 1 NGO 10 Cooperative 2 Religious group 11 Traders' association/ Cultural association 12 business group 3 Political group 13 Protessional association 4 Youth group 14 Trade union 5 Women's group 15 Credit/ finance group 6 Parent group 16 Water/ waste group 7 School committee 17 Neighborhood/village Health committee 18 association 8 Sports group 19 Civic group 9 Other 20 192 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Degree of participation Leader 1 Very active 2 Somewhat active 3 Not active 4 4A.3 Which of these groups is the most important to your household? (List up to three by name and code type of organization.) Group 1: _ __ Group 2: [ Group 3: [_] Now I'm going to ask you some questions about the members of these groups. 4A.4 Overall, are the same people members of these three different groups or is there little overlap in membership? Little overlap [ ] 1 Some overlap [ 1 2 Much overlap [ ] 3 Group__ _ 4A.5 Are group members mostly of the 1 1 ] 3 same extended family? Yes 1 No 2| l l l 4A.6 Are members mostly of the same religion?Ye1| l l l Yes 1 No 2 4A.7 Are members mostly of the same l gender? l l Yes INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 193 Group [1 1 2 1 3 4A.8 Are members mostly of the same political viewpoint or do they belong to the same political party? Yes 1 No 2| ll l 4A.9 Do members mostly have the same occupation? Yes 1 No 2 4A.10 Are members mostly from the same age group? Yes 1 No 2| ll l 4A.11 Do members mostly have the samp level of educafion? Yes 1 No 2 - l i ______ _I __l_l _ l 4A.12 How does the group usually make decisions? Group 1 2 3 The leader decides and informs «he nfhor grnou n,nrnbers. 1 The leader asks group members wha-t they thirk and then decides. 2 The groupm n,meber- ho-ld a discussion and decide Other (specify) 4 194 UNDERSTANDING AND Mv.EASURING SOCIAL CAPITAL 4A.13 Overall, how effective is the group's leadership? Group 1 2 3 Very effective 1 SomPwhat Pffpr'ivp 2 Not effective at all 3 4A.14 Do you think that by belonging to this group you have acquired neAT skVills or leamed somethincg valuabhle? 1 2 3 Yes 1 No 2 4B. Networks and Mutual Support Organizations Now I am going to ask you some questions about how the community xc . SA 12 A A A-.4 A -.11- 4B.1 Ifshe UL l F911-Lay schoo1 VI oLfL villagei /ighborhoIod1LV VVnLL ViULL a teacher for a long time, say six months or more, which people mi UiJLi V1U1dre/IeI6iLUr,LUUhU UU YVU LitiLN VVoUIU 6'tL LU6tgL1I Lu take some action about it? Yes No No one in the village! neighborhood would get together [ j 1 [ ] 2 (if yes, go to question 4B.3) Local/ municipal governmnent [ j 1 [ ] 2 Village / neighborhood association [ j 1 [ j 2 Parents of school children [ ] 1 [ ] 2 Time entire village! neighborhood [ ] 1 [ ] 2 Other (specify) [ ] 1 [ ] 2 INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 195 4B.2 Who would take the initiative (act as leader)? 4B.3 If there were a problem that affected the entire village/neighbor- hood, for instance (RURAL: "crop disease"; URBAN: "violence"), who do you think would work together to deal with the situation? Yes No Each person/household would deal with the problem individually [ I 1 [ ] 2 (if yes, go to section 4C) Neighbors among themselves [ j 1 [ ] 2 Local government / municipal political ieaders [ 1 [ ] 2 All community leaders acting together [ ] 1 [ ] 2 The entire village/ neighborhood [ ] 1 [ ] 2 Other (specify) [ ] 1 [] 2 4B.4 Who would take the initiative (act as leader)? 4C. Exclusion 4C.1 Differences often exist between people living in the same vil- lage/neighborhood. To what extent do differences such as the following tend to divide people in your village/ neighborhood? Not at all Somewhat Very much a. Differences in education [ ] 1 [ ] 2 [ J 3 b. Differences in wealth/ material possessions [ ] 1 [ ] 2 [ ] 3 c. Differences in landholdings [ ] 1 [ ] 2 [ ] 3 d. Differences insocialstatus [ ] 1 [ ] 2 [ ] 3 e. Differences between men and women [ ] 1 [ ] 2 [ ] 3 f. Differences between younger and older generations [ ] 1 [ ] 2 [ ] 3 196 UNDERSTANDING AND MEASURING SOCIAL CAPITAL g. Difference between long- rime irinabiranrs anda new settlers [ ] 1 [ 1 2 [ ] 3 n. Difference in political party affiliations [ ] 1 [ ] 2 [ ] 3 i. Differences in religious beliefs [ ] 1 [ ] 2 [ ] 3 j. Differences in etnnic background [ ] 1 [ ] 2 [ ] 3 k. Other differences (specify) [ j 1 [ j 2 [ 3 3 4C.2 Do these differences cause problems?s Yes [ i 1 No [ ] 2 (go to question 4C.5) 4C.3 How are these problems usually handled? Yes No a. People work it out between themselves [ ] 1 [ ] 2 b. Family/ household members intervene [ ] 1 [ ] 2 c. Neighbors intervene [ ] 1 [ ] 2 d. Community leaders mediate [ ] 1 [ ] 2 e. Religious leaders mediate [ ] 1 [ ] 2 f. Judicial leaders mediate [ ] 1 [ ] 2 4C.4 Do such problems ever lead to violence? Yes [ ] 1 No [ ] 2 4C.5 Are there any services where you or members of your household are occasionally denied service or have only limited opportunity to use? Yes No a. Education/schools [ 1 1 [ ] 2 b. Health services/ clinics [ 1 [ 1 2 c. Housing assistance [ ] 1 [ ] 2 d. Job training/employment [ ] 1 [ ] 2 I NS T RU7 M.ENI'll TSl OF T.H E SO CI A L CA PIT A L A S S E SSM E N T TOOL197 e. Credit/finance [ ] 1 [ ] 2 f Trqn-nnrtafqoin [ 1 1 [ 1 9 r---------I L J L ] - g. Water distribution [ ] 1 [ 1 2 h Sqnitfnnn zprusrvic [ i 1 [ 1 9 b ~~~ ~ ~~~L I L. I i. Agricultural extension [ I 1 [ ] 2 . Ius.r.ceordlkic r~esoluio4~n [ 1 1 [ 1 9 k. Security/police services [ ] 1 [ ] 2 4C.6 Do you think that there are other households in this comnmunity thaft ha,vec Urluh a-rccess problems? 4C.7 If yes, what percentage of houselIolds is excluded? Service Others 4Percentage excIULLUe exciAludUe <--roI -1 l... /0-1 25-50%=2 Yes =1 76-99%=4 X T-- n I nnrI I I N U L j Il u U !0 = 0 a. Education/schools [ ] [ ] b. Health services/clinics [ I [ I c. Housing assistance [ ] [ ] d. job training/employment [ L I j e. Credit/finance [ 1 r. Transportation [ I [ I g. Water distribution [ ] h. Sanitation services [ j [ j i. Agricultural extension [ ] j. justice/conflict resolution i k. Security/policy services [ ] 198 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 4C.8 What are the reasons or criteria why some people are excluded rronri mese services? Ies 1 No a. Income level [ ] 1 [ ] 2 b. Occupation [ j 1 [ j 2 c. Social status (class, caste) [ ] 1 [ 1 2 d. Age [ j 1 [ 2 e. Gender [ ] 1 [ ] 2 f. Race/ ethnicity [ j 1 [ j 2 g.Language [ ] 1 [ ] 2 h. Religious beliefs [ ] 1 [ j 2 i. Political affiliation [ ] 1 [ 1 2 j. Lack of education [ ] 1 [ ] 2 4D. Previous Collective Action 4D.1 In the past year, how often have members of this village/neigh- borhood gotten together and jointly petitioned government offi- cials or political leaders with village development as their goal? Never [ 1 (go to question 4D.3) Once [ ] 2 Acouple of times [ ] 3 Frequently [ 1 4 4D.2 Was this action/were any of these actions successful? Yes, all were successful [ ] 1 Some were successful and others not [ ] 2 No, none were successful [ ] 3 4D.3 How often in the past year have you joined together with others in the village / neighborhood to address a common issue? Never [ ] 1 Once [ ] 2 A couple of times [ 1 3 Frequently [ 1 4 4D.4 In the last three years have you personally done any of the fol- lowing things: iNSTRUMENTS OF THEr SOCIAL CAPiTAL ASSESSMENT TOOL 199 Yes No a. V7oted in tbe erec.ons [ 1 1 [ I 2 b. Actively participated in an association [ I 1 [ I 2 c. Made a personal contact withL an iUfluenl. al person r ] L [ ] 2 d. Made the 'media interested ir. a problerr,-..--. [ 1 1 [r I unt a F1iu1c:`i L J i t e. Actively participated in anL iIUULIILCZLIUII carLd.iLFCII [ ] I [ r I2 f. Actively participated in art elec'U - .. ion -pa ig-n ` r I .I r 1 [ 411 ~1tLLI11 dUFd1tlr, I I L I 1- g. Taken part in a protest _ _ r 1 1 1 n rlidItai U1 Uoir1U :ern dloUli L j I L J h. Contacted your elected representative [ I 1 [ L 2 i. Taken part in a sit-in or aisruptiin of government meetings / ornces 2 j 1 [ 1 2 j. Talked with other people in your area about a problem [ ] 1 [ ] 2 k. Notified the court or police about a problem [ ] 1 [ ] 2 i. Made a monetary or in-kind donation [ ] 1 [ ] 2 m. Volunteered for a charitable organization [ ] 1 [ ] 2 4D.5 Have you been approached by someone personally during the last three years who asked you to do any of the following: Yes- No a. Vote inthe elections [ ] 1 [ ] 2 b. Actively participate in an association [ ] 1 [ ] 2 c. Make a personal contact with an influential person [ ] 1 [ ] 2 d. Make the media interested in a problem [ ] 1 [ ] 2 e. Actively participate in an information campaign [ ] 1 [ ] 2 200 UNDERSTANDING AND MEASURING SOCIAL CAPITAL f. Actively participate in I 1 r I 1 ' art eiecL.UL Lca11tp1y,1t [ I 1 L I g. Take part in a protest marcl or demorsirador 1[ ] 1 [r I 2 niiardci o de -- 0 UIIU1-tti L --t ' 1 01UIt L[ JII L IJ h. Contact your elected representative [ j 1 [ I 2 i. Take part in a sit-in or disruption of government meetings/offices [ ] 1 [ ] 2 j. Talk witn otner people in your area about a problem [ 2 1 [ j 2 k. Notify the court or police about a problem I ] 1 [ I 2 1. Make a monetary or in-kind donation [ 2 1 [ j 2 m. Volunteer for a charitable organization [ 2 1 [ j 2 4D.6 ff some decision related to a development project needed to be made in this village/neighborhood, do you think the entire vil- lage / neighborhood would be called upon to decide or would the community leaders make the decision themselves? The community leaders would decide [ ] The whole village/ neighborhood would be called [ ] 2 4D.7 Overall, how would you rate the spirit of participation in this vil- lage / neighborhood? Very low [ ] 1 Low [ ] 2 Average [ ] 3 High [ ] 4 Very high [ ] 5 4D.8 How much influence do you think people like yourself can have in making this village/ neighborhood a better place to live? A lot [ ] 1 Some [ ] 2 Not very much [ ] 3 None [ 1 4 INSTRUMENTS OF THE 'SOCIAL CAPITAL ASSESSMENT TOOL 201 5. COGNITIVE SOCIAL CAPITAL 5A. Solidarity 5A.1 Suppose someone in the village/neighborhood had something unfortunate happen to them, such as a father's sudden death. Who do you think they could turn to for help in this situation? (Record first three mentioned.) a b c No one would help 1 Family 2 Neighbors 3 Friends 4 Religious leader or group 5 Community leader 6 Business leader 7 Police 8 Family court judge 9 Patron/ employer/benefactor 10 Political leader 11 Mutual support group to which s/he belongs 12 Assistance organization to which s/he does not belong 13 Other (specify) 14 5A.2 Suppose your neighbor suffered an economic loss, say (RURAL: "crop failure"; URBAN "job loss"). In that situation, who do you think would assist him/her financially? (Record first three men- tioned.) a b c No one would help 1 Family 2 Neighbors 3 Friends 4 Religious leader or group 5 Community leader 6 Business leader 7 Police 8 Family court judge 9 Patron/ employer/benefactor 10 202 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Political leader 11 Mutual support group to which s/he belongs 12 Assistance organization to which s/he does not belong 13 Other (specify) 14 5B. Trust and Cooperation 5B.1 Do you think that in this village/neighborhood people generally trust one another in matters of lending and borrowing? Do trust [ ] 1 Do nottrust [ 1 2 5B.2 Do you think over the last few years this level of trust has gotten better, gotten worse, or stayed about the same? Better I 1 1 Thesame [ ] 2 Worse [ ] 3 5B.3 Compared with other villages/neighborhoods, how much do people of this village/ neighborhood trust each other in matters of lending and borrowing? Less than other villages/ neighborhoods [ ] 1 The same as other villages/ neighborhoods [ ] 2 More than other villazes /neighborhoods [ 1 3 5B.4 Suppose someone from the village/neighborhood had to go away for a while, along with their family. In whose charge could they leave (RURAL: "their fields": URBAN: "their house")? (Record first three mentioned.) a b c Ofther fnmily mpmber 1 Neighbor 2 Anyone from, tfh village/neighborhood for this purpose 3 Other (specify) 4 No one 5 INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 203 5B.5 Suppose a friend of yours in this village/neighborhood faced the following alternatives, which one would s/he prefer most?, RUTRAL: Own and farm 10 hectares of land entirely by themselves 1 Own and farm 25 hectares of land jointly with one other person 2 URBAN: Own a patio 10 m2 alone 1 Own a patio 25 m2 that is shared with one other family 2 5B.6 If you suddenly had to go away for a day or two, whom could you count on to take care of your children? (Recordfirst three men- tioned.) a b c Other family member 1 Neighbor 2 Anyone from the village/ neighborhood for this purpose 3 Other (specify) 4 Don't have children 5 5B.7 Do you agree or disagree that people here look out mainly for the welfare of their own families and they are not much concerned with village/ neighborhood welfare? Strongly agree [ ] 1 Agree [ ] 2 Disagree [ ] 3 Strongly disagree [ ] 4 5B.8 If a community project does not directly benefit your neighbor but has benefits for others in the village/ neighborhood, then do you think your neighbor would contribute time for this project? Will not contribute time [ 1 1 Will contribute time [ ] 2 204 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 5B.9 If a community project does not directly benefit your neighbor but has benefits for others in the village/ neighborhood, then do you think your neighbor would contribute money for this pro- ject? Will not contribute money [ ] 1 Will contribute money [ ] 2 5B.10 Please tell me whether in general you agree or disagree with the following statements: Strongly Strongly agree Agree Disagree disagree a. Most people in this village/ neighborhood are basically honest and canbe trusted. []1 [] 2[] 3 []4 b.People are always interested only in their own welfare. [ ] 1 [ ] 2 [ ] 3 [ ] 4 c. Members of this village / neighbor- hood are more trustworthy than others. ! ] 1 [ ] 2 [ ] 3 [ 1 4 d.In this village/ neighborhood, one has to be alert or someone is likely to take advantage of you= I 1 1 [ 1 2 [ 1 3 f 1 4 e. If I have a problem, t-here is always someone tohelpme. ] 1 [ ] 2 [ ] 3 [ ] 4 f T do not npy attention to the opinions of others in th+b lage neighborhood. [ 1 [ ] 2 [ ] 3 [ 1 4 g.Most people!- in this village / neighborhood if VV ilnee it. help 34 if you need it. [ ]1 [ ] 2[1]3 [1]4 INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 205 h.This village/ neigh- borhood has prospered in the iast five years. [ 1 [] 2[ 3 [ 4 i. I feel accepted as a member of this village! neighborhood. []1 [] 2[] 3 []4 j. RURAL: If you lose a pig or a goat, someone in the village would help look for it or would return it to you. URBAN: If you drop your purse or wallet in the neighborhood, someone will see it and returnit to you. [ ] 1 [ ] 2 [ ] 3 [ ] 4 5C. Conflict Resolution 5C.1 In your opinion, is this village/neighborhood generally peaceful or conflictive? Peaceful [ ] 1 Conflictive [ ] 2 5C.2 Compared with other villages/neighborhoods, is there more or less conflict in this village/ neighborhood? More [ ] 1 The same [ ] 2 Less [ ] 3 5C.3 Do people in this village/neighborhood contribute time and money toward common development goals? They contribute some or a lot. [ ] 1 They contribute very little or nothing. [ ] 2 206 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 5C.4 Compared with other villages / neighborhoods, to what extent do people of this village/ neighborhood contribute time and money toward common development goals? They contribute less than other villages/ neighborhoods. [ ] 1 They contribute about the same as other villages/ neighborhoods. [ ] 2 They contribute more than other villages / neighborhoods. [ ] 3 5C.5 Are the relationships among people in this village/neighbor- hood generally harmonious or disagreeable? Harmonious [ ] 1 Disagreeable [ ] 2 5C.6 Compared with other villages/neighborhoods, are the relation- ships among people in this village/neighborhood more harmo- nious, the same, or less harmonious than other villages/neigh- borhoods? More harmonious [ ] 1 The same [ ] 2 Less harmonious 1 1 3 5C.7 Suppose two people in this village/neighborhood had a fairly serious dispute with each other. Who do you think would pri- marily help resolve the dispute? No one; people work it out between themselves [ ] 1 Family! household members [ 1 2 Neighbors [ ] 3 Commninitv leaders [ 1 4 Religious leaders [ ] 5 TiiJuicial leaders [ ] 6 Other (specify) [ ] 7 INSTRUMENTS OF THE SOCIAL CAPIT'AL ASSESSMENT TOOL 207 Annex 1D O;g 1anizationA A IPAroile ILrLCIViC VV Gu.ds The overall objective of the institutional profile is to delineate the rela- tionships and networks that exist among formal and informal institutions operating in the community, as a measure of structural social capital. Specifically, the profile assesses the organizations' origins and develop- ment (historical and community context, longevity, and sustainability); quality of membership (reasons people join, degree of inclusiveness of the organization); institutional capacity (quality of leadership, participa- tion, organizational culture, and organizational capacity); and institu- tional linkages. Between three and six institutions per community should be profiled. The organizations need to be identified through the community inter- views and/ or household survey as key organizations or those having the most impact or influence on community development. For each organizational profiled, interviews need to be carried out with its leadership, members, and nonmembers. Individual interviews need to be conducted with up to three leaders per organization. The inter- views should preferably be face-to-face, but a self-administered written questionnaire may be substituted. Focus group interviews should be car- ried out with members and nonmembers, with each group ideally having between 5 and 12 participants Depending on the size and diversity of the group's membership, anywhere from one to four focus groups should be conducted. Of the nonmembers, effort should be made to conduct two focus groups, one for nonmembers who want to be members and one for nonmembers with no interest in becoming a member. Each focus group should have a moderator and two observers. The moderator's role is to facilitate the discussion, probe on key issues, elicit comments from all participants, and focus the discussion on the issues of interest without seeming to interrupt or ignore extraneous comments from participants. The observers' role is to take notes on the content of the discussion and process of group dynamics. Upon completion of the focus group interview, the moderator and observers should conduct a follow-up debriefing to refine the interview notes and discuss preliminary findings. 1. ORGANIZATIONAL IDENTITY 1.1 Name of organization 1.2 Type of organization 1.3 Membership 208 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 1.4 Location (district, village, neighborhood) 1.5 Names of leaders 2. LEADERSHIP INTERVIEW GUIDE 2A. Origins and Development 2A.1 How was your organization created? Who was most responsible for its creation (e.g., government mandate, community decision, suggestion of outside NGO)? 2A.2 What kinds of activities has the organization been involved in? 2A.3 In what ways has the organization changed its structures and purpose? What is the main purpose of your organization today? 2A.4 As the organization developed, what sort of help has it received from outside? Has it received advice and/or funding or other support from the government? What about from nongovernment sources? How did you get this support? Who initiated it? How was the support given? What benefits and limitations has the organization derived from this support? 2B. Membership 2B.1 Can you tell us about the people involved in your organization? How do they become involved? Are all people in the community involved? If not, why are some members of the community not involved? 2B.2 Why do people join or are willing to serve (as officers/lead- ers /board members) in the organization? Is it hard to convince people to continue being active in the organization? What kinds of renuests/demands do theyr make on the leadership and orpa- nization? 2B.3 Are active members in this organization also members of other nrnyni7,afinn- in the crnmmiinitv/reo'inn? Do npeonle tend on he members of just one organization or join many simultaneously? Car. youexnin .Atrhy? INSTRUMENTS OF THE SOCiAL CAPITAL ASSESSMENT TOOL 209 2C. Institutional Capacity 2C.1 How would you characterize the quality of leadership of this organizataion., in termrs of.. ... .OstabiAh IY ... nwmlber of leaA ers availabili,;? . . .d.iversity / heterogeneit1y of1 1leaerl2ip?UJ ...quality anU sklls VI leadeIrs? ... rela.dlonshLilp 01 IladeUi LU bLstal arLU LU lith LUllUL)Uni y? 2CL.2 1710W IUw wou U y-Lac.dize thle qual't-y o0 particpation in this organization, in terms of... ... attendance at meetings, both internal to the organization and externally with other organizations? ... participation in decisionmaking within the organization ... dissemination or relevant information prior to tne decision? ... informai opportunines to discuss the decision? ... consultation processes with base organizations or with the community? ... broad debate, including opposition positions, and honesty? ... dissemination of the results of the decisionmaking process? ... the number of women, young people, poor people who work in the organization and who occupy positions of responsibility in the organization? ...whether any groups within the community feel excluded from the organization? What groups are they? ...the level of participation of more prosperous families (elites) in the organization? 210 UNDERSTANDING AND MEASURING SOCIAL CAPITAL ...whether elites are sympathetic, supportive, interfering, adver- sarial, or negative influences? 2C.3 How would you characterize the organizational culture of this organization, in terms of... ... the existence and level of knowledge of the procedures and policies'? ...whether the procedures and policies are carried out? Whether there are problems with nonattendance at meetings, theft of property or supplies? ... conflict resolution mechanisms, both within the community and within the organization? ... the nature of conflicts between the organization and commu- nity members? 2C.4 How would you characterize the organizational capacity of this organization, in terms of... ... carrying out specialized activities (e.g., credit, commercializa- tion)? ... supervising and contracting consultants? ...preparing financial reports for banks, donors, and govern- ment? ... reacting to changing circumstances (e.g., price fluctuations, change in government)? ... developing specific plans for the future (instead of reacting to opportunities as they present themselves)? ...reflecting on and learning from previous experiences? 2D. Institutional Linkages 2D.1 How would you characterize your organization's relationship with other community organizations? When do you feel the need to establish collaboration!/ ink- with them? 2D.2 Do you have links with organizations outside the village/neighborhood? r With which ones? What is the nnhire of those links? 2D.3 Do you feel sufficiently informed about other organizations' pro- grams an, d activities? Wbh.t ara your csurrces nf infnrmnri-:n? 2D. Have -ou attem-ted to org-n;-e or n,vwork wzi-h ni-r orgrn2nn.- tions to achieve a mutually beneficial goal? (Ask for which activi- ties.) TIs 4 am -omon c1-s+an among nrza.,;ons n ib vil- lage/neighborhood? (Probe as to reasons why or why not.) 2D.5 Could you describe your relationship with the government? HJave you hOaA expeiece ill_A trin to0 get govCIAL,,en assAOistanc? What was your experience? Which level of government do you find imost- cooperati;Ve Ilocal, district, T-T-,nl) Has __ ov. Il A%A LUL LUjJ I a kLA C~A 01, r..ALI AL, A IIQLrALIU I.) A 1100 LIM[ I ment made particular requests of your organization? 2D.6 Is your organization linked to any government program? Which ~UVLAUICA L /1~L 1111) I1) YULLI VCUL~U4 LLULU tVU IVIAVCU YVV LAIt govern-t-nent programni's' :_ _ologaia-ni.vle ith? Why those particular programs? What sort of role does your uorad11idz1aio pldy Hi UIte p)rolgira tAre LItere certair. cLhlaracter-is- tics of these programs that make it easier for your organization to W UL1L WIlU I ULC f)LU8L1dl b 2D.7 Do you feel s-ufficiently informed about governmllent pluroarni and activities?, What are your sources of information? 2D.8 Have you attempted to give inputs to the government? What were the circumstances, What nave been the results? Wvhat kinds of challenges did you have to deal with? (Probe for any role in planning, operation, and maintenance ojgovernment-sponsored ser- vices.) 2D.9 Has your organization been invited to participate in any of the various government development planning processes? Wnat do you think about these planning mechanisms? 2D.10 In general, how do you assess your organization's actual influ- ence on government decisionmaking at the district level? 212 UNDER5TANDING AND MEASURING SOCIAL CAPITAL 3. MEMBERS INTERVIEW GUIDE 3A. Organizational History and Structure 3A.1 How did this group start? 3A.2 Who have been the leaders of this group? Who are the leaders 7lw Flow anlU wily Ulu Ut le 1due1zbIU~LI Udi1%CUC LIII VVIIL no-w-i wan-whdith edrl.phag over fime? "'llat are the qualities of leadership? 3A.3 Why did you decide to join this group? What kinds of benefits do you get by being a member of this group? 31A.4 How are the leaders of thiis organization selected? How are deci- sions made? To what extent do you feel the organization repre- sents your concerns to tne outside woria ana ro tne government? 3A.5 'vWvhy are some people not members of this organization? 3A.6 How do you feel this organization complements, replaces, or competes with government institutions' activities in the comrnmu- nity? 3A.7 How do you feel this organization complements, replaces, or competes with nongovernmental institutions' activities in the community? 3A.8 What would you do to make this organization more effective? 3B. Institutional Capacity 3B.1 How would you characterize the quality of leadership of this organization, in terms of... ... .stability? ... number of leaders/ availability? ... diversity/heterogeneity of leadership? ... quality and skills of leaders? ...relationship of leaders to staff and to the community? INSTRUM.!ENTS OF THE SOCIAL CAPITAL ASSESSMlENT TOOL21 3B.2 How would you characterize the quality of participation in this organization. in terms of--- attendance at meetings, bnth intearna tn the nrganizatinn and externally with other organizations? ...participation in decisionmaking within the organization? ...dissemination of relevant information prior to the decision? ... informal opportunities to discuss the decision? ... consultation processes with base organizations or with the co.mmurikJ? ... broad debate, includJing --b t- pftoiL-to, and iLont Acssemrina-^rin o-fthe resu-lt-- o-f she Ancnrr1rn.,ang rocess? .t -V.L.LtL of VVomeni, young people, poor peoplie wYIL VVVIo in the organization and who occupy positions of responsibility in M..*whVIh a.Ly grUUps wVILItILn LIt. coJInUILULryf eCCI exAcIluAded frjioL the organization? What groups are they? ... the level of participation of more prosperous families (elites) in ithe u5aniuzauuiin? ~I --1--- ----- . n ... witeUtel Uiit.es aite byIIipatLtt.Lic, sUppUorive, flteleitil., duvtr- sarial, or negative influences? 3B.3 How would you characterize the organizational culture of this oJrgarzaxuu, in1 terms o)r... .m.le exisrence and ievei or knowiedge of procedures and pouicies? ...whether the procedures ana policies are carriea out? vvnetner there are problems with nonattendance at meetings, theft of property or supplies? ... conflict resolution mechanisms, both within the community and within the organization? 214 UNDERSTANDING AND MEASURING SOCIAL CAPITAL ... the nature of conflicts between the organization and commu- ml; remb-ers? ou.± HowUVY VVoL'IQ yoju citaraLctr.e uLIM UoIX&4Ii49&4iILUJ L4p"LU± VI Lftis organization, in terms of... ...carrying out specialized activities (e.g., credit, commercializa- Lion)? ... supef-Vi Utr adtt cUILLIdLLU1g cUIrSUILdILIb .. .preparing financial reports for banks, donors, and govern- ment? ... reacting to changing circumstances (e.g., price fluctuations, cnange in government)? ...developing specific plans for tne future (instead of reacting to opportunities as they present themselves)? ...reflecting on and learning from previous experiences? 4. NONMEMBERS INTERVIEW GUIDE 4A. Group #1: Nonmembers who want to be members 4A.1 In your opinion, do the benefits of this particular organization spread beyond its members? 4A.2 Why are some people not members of this organization? Why are you not a member of this organization? 4A.3 How far do you think this organization complements or com- petes with other community organizations? 4A.4 What is your view about how the organization deals with gov- ernment? (For example, does kinship or party affiliation play a role in determining the relationship?) 4A.5 What is your view about how the organization deals with other organizations that work in the village/ neighborhood? INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 215 4B. Group #2: Nonmembers who do not want to be members 4B.1 In your opinion, do the benefits of this particular organization s-re-ad be-ond its mnemnbers? 4B.2 1AJbhy are some people not memhbersc of 4 4hiso i n? Why are you not a member of this organization? 4B.3 How far do you think this organization complements or com- AD A v a - -- -- A_ A A - - - - - - - - - - 4BU.4T YVL UL Iis yoLuL vicvv about hoL teLIC UIgLrliUI; deaIl VVILth guv- ernment? (For example, does kinship or party affiliation play a role ir.- dJetemlgiring thA el.;nsi? rL~ IULCI iILIL-Ul L6 LI LC: tla CLIU LZIILi'F 4BLJ.J VVItaL tis yuUI vit CILUUUL IIUVV Uho V16he uU.aLlul udcals VVvLiL ULIo I organizations that work in the village / neighborhood? 216 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Annex 1E Organizational Profile Scoresheet 1. TLEADERSHI. 1 A DRo,-atO 1 A.1 1 IA. 1 oes 'I-e - L o L" No [ ] 2 1A.2 Is the amount of time the leaders remain in their position suffi- N/- ~~ ~~ ~~~~r I I i'es L[ I I No [ ] 2 1A.3 Is there the possibility of reelecting successful leaders? Yes [ I 1 No [ j 2 !B. Density/Availability iB.! How many people within the organization have acquired the capability and qualities to be effective leaders? No one possesses these qualities [ ] 1 Few(lto3) [ ] 2 Some (4 to 6) [ I 3 Many (more than 6) [ ] 4 1B.2 How many are put forward for leadership tasks? Only a few are ready to be leaders. [ ] The group of candidates is limited but adequate. [] 2 There is never a lack of candidates (candidates who are prepared, enthused, and available to assume a leadership role). [ ] 3 INSTRUMENTS OF THE SOCiAL CAPiTAL ASSESSMENT TOOL 217 1B.3 How amenable are former leaders to continued participation in she1 orgalu-a9-on T.here are no previous leades the I IL~iSa~IL ~ IVi I a..ItD, UL L organization is new. [ ] 1 ALImost no par.'cia-;o y --'--laer.[ x-i~~IL. LLya LI%IJ.aLIUJIL uy LU1IILCI MaUCIZI~. [ J Some participation by former leaders. [ ] 3 A _ - _ L._ 1__J --. r 1 A ZAcfi;ve particiption aL bLLy forIL.e leaders. L I 4t It L A../IL7.A._: ...~ lC.1 o Lhte leadUerb LerIU t cUmie F1o011 a fe-w groups or iamilies rnar are always the same, or do the leaders represent a wider circle among the comuunity? From few groups [ j 1 From various groups within the community [ j 2 From almost all the groups within the community [ j 3 1C.2 What percentage of those that occupy leadership positions with- in the organization are women? Less than 10%' [ 1 Between 10 % and 25% [ j 2 Between 26% and 50% [ ] 3 More than 50% [ ] 4 !D. Leadership Quality and Skills !D.! in general, how would you characterize the quality ot leadership in this organization in terms of... Defi- Excellent Good Adequate cient a. Education/training [ ] 1 [ ] 2 [ ] 3 [ 1 4 b.Dynamism/vision? [ ] 1 [ ] 2 [ ] 3 [ ] 4 c. Professionalism/ skills? []1[] 2 [] 3 [] 4 d.Honesty/ transparency? []1 [] 2[] 3 []4 218 UNDERSTANDING AND MEASURING SOCIAL CAPITAL IE. Relationship between Leadership, Staff, and Constituency 1E.1 How would you characterize the relationship between the exec- utive director and tne management and technicai stafr? Harmonious, without major probiems [ I 1 Coexisting, with occasional rivalries [ 1 2 Conflictive, with many problems [ j 3 Dysfunctional, without communication or coordination [ i 4 iE.2 What level of acceptance and legitimacy does the leadership have, especially among grassroots organizations whose commu- nities are underrepresented? The leaders are openly accepted and everyone recognizes their legitimacy to represent their interests. [ l 1 The leaders are accepted by the majority of the community; the majority recognize their legitimacy. [ 1 2 The leaders are accepted by a minority of the community members; leaders have little legitimacy. [ ] 3 The leaders are not accepted and do not have legitimacy within the community. [ 1 4 2. PARTICIPATION 2A. Frequency of Meetings 2A.1 Should the frequency with which the organization meets be greater, less, or remain the same? Greater [ ] 1 Less [ 2 The same [ 1 3 2B. Participation in Decisionmaking 2B.1 What have been the hvo most important decisions made in the past year? INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 219 Decision # 1: _ Decision # 2: 2B.2 Thinking about these decisions, did any of the following take place? (Code decision # 1 first, then continue with decision # 2.) 1Topic a. Prior b. Oppor- Tc. Consul-7 dVwide- e. Dis- dissemina- tunity for tation spread semina- tion of in- informal with debate, tion of formation discussion grass- opposing results roots opinions, and frank discussion Yes=1 Yes=l IYes = 1 Yes= 1 Yes= 1 No=2 No= 2 No=2 No=2 No=2 Decision Decision #2 2C. Inclusiveness 2C.1 In the last three meetings, what has been the level of participation of women, of youth, and of the poorest groups? Little! Active Moderate None a. Women [ ] 1 [ ] 2 [ ] 3 b. Youth [ ] 1 [ ] 2 [ ] 3 c. Poor [ ] 1 [ ] 2 [ ] 3 220 UNDERSTANDING AND MEASURING SOCIAL CAPITAL 2C.2 In comparison with earlier meetings, was this level of participa- tion more, less, or tne same? More Less Same a. Women [ ] 1 [ ] 2 [ ] 3 b. Youth [ j 1 [ j 2 [ j 3 c. Poor [ ] 1 [ ] 2 [ ] 3 2C.3 To what degree does the organization truly represent its mem- bers? Highly representative [ ] 1 Somewhat representative [ 1 2 Slightly representative [ ] 3 Not representative at all [ ] 4 2C.4 What percentage of the population in this community feels included as beneficiaries of the organization or feels its interests are represented by the organization? Less than 25% [ ] 1 Between 25 % and 50% [ ] 2 Between 51% and 75% [ ] 3 More than 75% [] 4 2D. Participation by Elites 2D.1 To what degree do the more prosperous families in the commu- nity (those with land, businesses, or professions) attend meet- ings, hold positions, or participate in activities of the organiza- tion? Active [ ] 1 Moderate [ ] 2 Little / none [ ] 3 2D.2 What is the relationship of the more prosperous families toward the organization itself? They are a resource to be counted on, sympathetic and/or supportive. [ ] 1 They could be a resource, demonstrating interest but currently are an interfering element. [ ] 2 INSTRUMENTS OF THE SOCIAL CAPITlAL ASSESSMENT TOOL 221 They could be a resource, but demonstrate no intLeret adCiU .U11eLnly are 11LULdiffere. L I They cannot become a resource and represent an adUversaria1l or negatiLve elertLent. r ] A 3. ORKANILAIIUTNAL C..UL1UKR 3.1 How many rIlerriLWers icnOw tue pruce-uules, I11.1s, andu Ltass V the organization? The majority of members [ ] 1 Some members [ j 2 Few members [ ] 3 3.2 How willing is the organization to confront problems with its members (if they were to happen) sucn as not attending meer- ings, avoiding work, or stealing property belonging to the orga- nization? The organization is very willing to confront damaging behavior on the part of its members. [ ] 1 The organization is sometimes wiiiing to confront damaging behavior on the part of its members. [ ] 2 The organization has iittie capacity to confront damaging behavior on the part of its members. [ ] 3 3.3 For serious cases, do guidelines or rules exist to sanction, fine, or expel the transgressor? Yes [ ] 1 No [ ] 2 4. ORGANIZATIONAL CAPACITY AND SUSTAINABILIrY 4A. Specific Capacities 4A.1 What is the organization's capacity to... Excellent Good Adequate Deficient a. Carry out its specialized tasks (e.g., credit, training, com- mercialization)? [] 1 [] 2 []3 []4 222 UNDERSTANDING AND MEASURING SOCIAL CAPITAL b. Supervise and contract specialized consultants or staff? [ ] 1 [ ] 2 [ ] 3 [ ] 4 c. Prepare fnanc-ial reports for banks, donors, or government? []1 [] 2[] 3 []4 d. Respond in a timely fashion to changes that affect tne organi- zation (e.g., price fluctuations, change of government)? []1 [] 2[] 3 [ 4 e. Develop specific plans for the future (instead of reacting to external oppor- tunities as they present themselves)? [ ] 1 [ ] 2 [ ] 3 [ ] 4 f. Reflect upon and learn from experience (build an institutional memory)? [] 1 [ 2 []3 []4 g. Resolve problems or conflicts with other organizations or social actors? []1 [] 2[ 3 []4 h. Resolve problems or conflicts within the organization? []1 [] 2[] 3 []4 4B. Collective Action and Formulation of Demands 4B.1 Does the organization have clearly defined processes for identi- fying the common needs and priorities of its members? Yes [I] 1 No [ ] 2 INSTRUMENTS OF THE SOCiAL CAPITAL ASSESSMENT TOOL 223 4B.2 In the last three years, have there been petitions or other formal expressions of demand by the 4 fl.n- .floiLty r 11 v,es ~~[ ] 1 No [ ] 2 4B.3 Have there been informal ways for members to express their Ud11 LW ^LU. r 1 1 1ts L[ J I No [ ] 2 4B.4 In what way has the organization addressed these demands? Promotes demands of common interest [ ] 1 Tries to identiry cormuon elernents [ 1 2 Tries to process them one by one [ ] 3 Tnere were rio demands [ j 4 224 UNDERSTANDING AND MEASURING SOCIAL CAPITAL Annex 1F Selection Criteria and Terms of Reference For Interviewers (Panama Pilot Test) Responsibilities The objective of the study is to pilot test the research instruments that are part of the Social Capital Assessment Tool (SOCATj. interviewers select- ed will have the following responsibilities: * Be completely familiar with the content and application of the various data collection methods, including the community profile, the house- hold questionnaire, and the organizational profile. * Be able to answer any questions regarding the study, its objectives, and fieldwork activities. * Select key informants to be interviewed according to the indicated sampling techniques. * Obtain informed consent from all interview subjects prior to their par- ticipation in the study. * Conduct interviews, whether group interviews, individual in-depth interviews, or individual surveys, using the indicated data collection instruments. * Provide in timely fashion the primary source materials generated dur- ing the interviews, whether expanded observation notes, appropriate- ly coded survey questionnaires, or other results of the qualitative inter- views. * Debrief field supervisors about the data collection process and the application of the instruments. Skills The interviewer must have certain qualities and skills to be able to under- take the diverse activities associated with the fieldwork. These include: * Completion of secondary school education, preferably with some uni- versity-level studies in one of the social sciences. * Previous experience conducting household surveys. * Previous experience conducting focus group interviews and/or using participatory research techniques. * Demonstrated ability to document, in detail and precisely, the main issues brought out in individual and group open-ended interviews. * Ability to work in teams, coupled with demonstrated self-motivation. INSTRUMENTS OF THE SOCIAL CAPITAL ASSESSMENT TOOL 225 * Availability to work six days a week, including Saturdays and LIdas _ith long wring-ours. * Availability to work in distant areas, in difficult terrain, under chal- lengLng conditions. Le.eeufn nLIT:unu Tnhe fimai siection of ihe researcn ieaiii will be nidUe d tle conClusuion of the training workshop. The workshop will present the study objectives, and "me contents, justification, and application of tne researcn instru- ments, and include practice applications of the instruments. The team of field supervisors will select rhe people for the position of interviewer. Tne selection criteria include demonstrated ability in the above-mentioned areas as well as: * Diversity among team members tnat appropriately reflects the diver- sity of the research population in terms of gender, ethnicity, and social class. * A complementary skill mix among team members regarding tech- niques for conducting qualitative individual and group interviews, conducting household surveys, drafting expanded field notes, tran- scribing field notes, and applying participatory research methods. Annex 1G: Training Plan for Fiel[d Workers (Panama P'ilot Test), Day 1 8:00 Open ing 15" Introdiuction to the workshop Lecture: Handouts: Review the workshop objectives, agenda, Pens, pads, notebooks, expectations. agenda, markers, maisking tape 30" Introduction of the workshop Group dynamic: participants "Alliterative names" as an ice breaker: name, wthere From, previous research or field expierience; name and alliterative adjective written on paper. 45" Introdluction to the SOCAT study Lecture: Handouts: VVhat is social capital? Memos on social capital, C)verview, definitions, relation to other types of SOCAT study, terms of capital, relationship tc economic development reference for VVhat is the SOCAT stud,y? intervievvers Objective, conceptual framework, structure of the tools, what the study is not. Who will participate? Criteria for selection cf interviewers, review of terms of reference. 9:30 Break; 9:45 Part ]: Cormnunity profile __ _ 15" Overview presentation Lecture: Handouts: Objectives, structure, activities, mate:rials. Community profile intervievw guide 30" Qualitative participatoDry Group discussion: methodologies: How to conduct a group interview; how to enter In-dLepth interview techniques the community; role of the research team with groups (facilitator and observers); using the interview _ guide; importance of transcripts. 30" Qualitative participatory Group discussion: methodologies: How to faci:litate a participatory mapping of - - - Community asset mapping comrnunity assets; identification of the community's asset portfolio; the art of listening. _ 30" Qualitative participatoDry Group discussion: methodologies: How to faci:litate a participatory Vein Venn diagrams diagram; community assessment of the relative importance and accessibility of associations. 20" Qualitative participatory Group discussion: methodologies: How to faci'litate a participatory flowchart Flowchart: diagrams of diagram; commurnity assessment of the organization networks network relationships betwee:n and among associations; assessment of the quality of ______ relaticinships (strong, weak, conflictive, etc.) 1.0" Review Lecture: Handou t: The final product of each communily Examples of transcripts, interview will be a folder containing m.aps, VTenn diagrarns the following: and flowcharits 1. Descriptive surnmary of the interview process (half page transcribed) 2. Expanded notes (integration of the observational and analytical notes of all research team mernbers, transcribed) 3. Methodo]Logicall notes (tran:scribed) it. Community asset map(s) (on 8x11 paper) 5. List of community needs assessment 6. Complete list of community associations 7. Vern diagram (on 8x11 paper) 8. Organizational flowchart (on 8x11 paper) 9. Raw field notes and other support materials 12:00 Lunch 13:00 Part I: Cornmunity profile (cont.) _ 90" Review of the community Group discussion: interview guide, by sections: Review the "why"' of each question and its relationship to social capital measurement; interview products for each section include: A. Definition of the community A. Map drawing and identification of communily assets B. Collective action and solidarity 13. Group interview C. Community governance and C. Grcup interview decisionmaking D. I(lentification of local level D. Drafting list of local-level institutions institutions E. Relationship between the E. Draiwing Venn diagrarm community and local level institutions F. Institutional networks and F. Drawing organizational flowchart organizational density _ _ 14.30 Break 15:00 Part I: Community profile (cont.) ____ ____ 90' Application of the community Role play: Will need enough space interview guide The objective of this exercise is to practice to recreate a comrnnunity interview skills and evaluate participants' interview plus all capabilities for conducting the community relevant materials interviews Assjign the roles of facihtator, ("research team" is observers, and informants. Have the "research responsible) tearn" and the "community memlbers" take 10 minutes to clarify their roles, strategies, responsibilities, etc. 30" Conclusion of the exercise Group discussion: Comments regairding the role play; observations on the roles of facilitator, observer, participants; constructive criticism regarding vways to improve interview techniques, eliciting information, managing group dynamics. 17:00 Finish Homework assignments Readmg: Handl out copies of all * Community interview guide research instruments, * Houselhold questionnaire support memoranda, * Organuzatiorn interview guide and lORs * Memoranda describing the SOCAT study * Terms of reference for intervielwers Writing: * For all participants, draft in their own words an introduction to the study, appropriate for presentation to informants; should include references to the study objectives, the meaning cf social capital, the use of participatory rnethods, and guarantee of confidentiality. * For those who played the role of observer, draft the expanded field notes. Practice: * Explain to sorneone not involved with the stldy the concept of social capital and its importance in the economic development of the comlununty. _ - Day 2 8:00 Review of previous day's work Group discussion: and hom,eworlk assignments; Ask participants to recite their wnltten questions and clarifications explemations of the study and to make comments regarding accuracy of contents, __________________________ style of delivery, clarity of language, etc. 9:00 Part I: FHousehold survey _ 15" Overview Lecture: Objectives, samplmg and interview _____ techniques, structure, materials. 75" Review the first three sections Group discussion: of the household survey Review the "why" of each question, questiormaire: its applFicaticin and coding. FoI the 1. Identification of selected genogram, slhow symbols usecd and household diraw ai sample genograrn. 2. Housing characteristics and household roster 3. Genogram 10:30 Break 90" Application of the household Role play: questionnaire (first thiree sections). In pairs, apply the first three sections of the questionnaire. 30" Review of the exercise Group discussion: In front of thie class, have a third person describe a genograrm from one of the interviews; question and answer pexiod about interview techniques and coding; emphasize research objectives and relatioinship to social capital. 12:30 Lunch 1:30 Part II: Household survey (cont.) ___ _ 120" Review Section 4. Structural Group discussion: social capital Review the "why" of each question, its ________ ___________________________ j