Policy Brief Issue 28 CAN JOB TRAINING DECREASE WOMEN’S SELF-DEFEATING BIASES? EXPERIMENTAL EVIDENCE FROM NIGERIA Authors: Kevin Croke, Markus Goldstein, and Alaka Holla GENDER KEY MESSAGES INNOVATION LAB • Gender-based occupational segregation – where women are The Gender Innovation Lab (GIL) conducts impact evaluations of concentrated in low-paid or low-profit sectors – is a non-trivial development interventions in source of the gender wage gap worldwide, accounting for as Sub-Saharan Africa, seeking much as 50% of the gap in some countries (World Bank 2011). to generate evidence on how There is evidence that women’s biases about their own potential can to close the gender gap in earnings, productivity, assets affect their performance and aspirations. and agency. The GIL team • Through an experiment in Nigeria, we found that an information and is currently working on over 50 impact evaluations in 21 communications technology (ICT) training resulted in university countries with the aim of graduates being 26 percent more likely to work in the ICT sector. building an evidence base with This suggests the potential for trainings to support the development lessons for the region. of and employment in emerging sectors despite initial lack of sector- The impact objective of GIL is relevant skills. However, we observed no average impact of the training increasing take-up of effective on the overall likelihood of being employed in any sector or on earnings. policies by governments, development organizations • We found that job training can increase aspirations among women and the private sector in order who lack the confidence to see themselves as successful. The to address the underlying program’s impact was strongest for women who initially held implicit causes of gender inequality biased against associating women with professional attributes. These in Africa, particularly in terms of women’s economic and women were more likely to switch into the ICT sector after the program social empowerment. The lab than initially unbiased women. aims to do this by producing and delivering a new body of • Job training programs offer a potential opportunity to reduce evidence and developing a occupational segregation by shifting norms about the appropriate compelling narrative, geared sectors for men and women to work in. The results of this study towards policymakers, on suggest that even without explicitly encouraging participants to defy what works and what does social norms, training programs can help individuals overcome self- not work in promoting gender equality. defeating biases. http://www.worldbank.org/en/programs/africa-gender-innovation-lab Globally, women face twin disadvantages in the labor market that contribute to lower earnings. First, women 26% are overrepresented in informal sector employment and unpaid work. The lack of mobility across professions based on gender norms is a significant barrier to the efficient functioning of labor markets. The second, and closely related disadvantage comes from occupational segregation. Some of these disadvantages may stem from biases that women themselves hold about their After the training, women were own potential. Individuals’ biases can affect their 26 percent more likely to aspirations and employment choices, particularly when work in the information and communications technology sector. these biases represent internalized social norms. As new sectors of employment emerge, a key question weeks, while all applicants could post their resumes is whether this pattern is replicated. In 2010, ICT was on a web-based employment network and attend a an emerging industry in Nigeria, with around 400 small job-fair where prospective employers could meet with and medium-sized ICT firms that catered mostly to the job candidates interested in working in the sector. At domestic market. However even as an emerging industry, the end of training, program participants could take it was already male dominated, with government figures an assessment exam which had been recognized by indicating that 67% of those employed in the information the domestic ICT industry as a form of certification to services sector were men. work in business processing activities. In addition to the core competencies, these training providers were THE PROGRAM also required to cover “soft” skills, such as cultural In 2010, the Government of Nigeria and the World sensitivity, teamwork, stress management, and time Bank launched the Assessment of Core Competency management. for Employability in the Service Sector (ACCESS) program in Nigeria with the long-term goal to break HERE’S WHAT WE DID into the international market for information technology Interviews conducted with directors of training centers enabled services. The objective of the program was to and with the target population prior to the intervention equip recent university graduates with sufficient skills to showed that women’s own confidence in seeking work in Nigeria’s ICT sector, and to certify these skills. work in the formal sector is seen as a key factor They expected the training to improve skills in three affecting women’s prospect in ICT. To explore this, we competency areas: communication (oral and written), used a tool developed by psychologists to measure computers, and cognitive skills, which are considered implicit biases, the Implicit Association Test (IAT), a “foundational competencies” for employment in the computer-based sorting task that tries to measure business processing outsourcing (BPO) sector. automatic associations between a group (such as The ACCESS Nigeria IT job skills training program was men or women) and concepts (such as employment, implemented across five cities in Nigeria in 2012. Slots sectors, or professionalism). In this case, the IAT in the program were randomly assigned to applicants. consisted of tests measuring the ease of associations Those offered slots in the program had access to 85 between gender and a number of attributes relevant hours of classroom-based training spread across 10 for women’s labor market participation in Nigeria. For instance, one test measured associations between This gain in ICT employment was only a shift in gender and the concepts of home and office. A second sectors, as overall employment did not significantly test was designed for the urban Nigerian context increased in response to the program. The program and measured associations between gender and the may have provided skills that did not increase the target concepts of office and petty trade. A final test measured population’s general employability but rather just their associations between gender and the concepts of potential in the ICT sector. It is also possible that firms professionalism and unprofessionalism. The computer- outside the sector did not recognize the value of the based pre-assessment provided a platform to collect certification exam that trainees took at the end of the baseline data on the 3,018 applicants of the program. course. This may show a need for trainings to include A self-administered questionnaire on applicants’ socio- more widely-sought after skills that could be attractive economic and demographic backgrounds, education across industries, in addition to sector specific skills. history, and labor market experiences and expectations The switching was more pronounced for women followed the assessment. Three years later, a final who held deep-seated biases against women’s survey was undertaken with applicants over the phone, professionalism and it induced their movement with a response rate of 91 percent. into a currently male-dominated sector, indicating HERE’S WHAT WE FOUND the potential for this program to substitute for confidence in one’s place in the professional world. The program induced switching into the emerging After the training, these women were three times more ICT sector in Nigeria. Given the government’s focus likely to find an ICT-enabled service job than unbiased on developing this sector and its identification of a skills women. The program could have changed how women gap as a major constraint to sectoral growth, this policy viewed their own abilities relative to men’s: women lever has proved somewhat effective in increasing the scored significantly higher on the oral skills that would employment of people with relevant skills in ICT. This have been observable to peers (voice clarity, fluency is encouraging regarding the potential for trainings and vocabulary, message clarity), as opposed to those to support the development of and employment in that would have been tested via computer (such as economically promising sectors despite initial lack of keyboard skills, internet and browsing skills). This may sector-relevant skills. indicate that reinforcing women’s confidence against implicit self-biases may be best achieved through a focus on developing skills easily visible by peers, such as communications. This program contained no special gender focus. Simply by expanding women’s possibility set, it increased their likelihood of taking advantage of jobs in this sector. Because the level of employment in the sector is so low for women, even unbiased women increased their prospects of employment in the sector (by 119%) after they were offered training. Women For women who initially were implicitly biased who entered with a pro-female bias, however, did not against associating women with professional benefit from training. These women may benefit more attributes, the likelihood that the program induced switching into the ICT sector was more than three from other types of programs, such as opportunities to times as large than that of unbiased women. expand their business networks. NEXT STEPS Taking this or a similar program to a larger scale could potentially contribute to the growth of the ICT sector and the economy as a whole. While the program did not impact overall employment levels or earnings, it induced increased mobility across professions and a reduction in occupational segregation, which should ultimately decrease gender wage gaps in the labor market. However, some key areas remain to be explored. Firstly, the results suggest that all the gains in ICT employment were concentrated in Abuja. In the other locations, the training was not effective in increasing employment in this sector. While this could be due to political issues in Nigeria that intervened between survey rounds, further studies should be conducted on increasing ICT employment among women in rural areas or in smaller cities. They may face slightly different, and potentially stronger, challenges to reaching that goal than women in larger cities, notably the absence of local role models or regionally-based networks in the sector. Second, while women are less likely to work in the ICT sector in general, the training program was equally effective for male and female students. Future work should focus on trainings more specifically aimed at encouraging women to enter sectors that are traditionally gender imbalanced, concentrating on the specific constraints faced by female students. In Kenya, the Gender Innovation Lab is currently evaluating an innovative strategy aimed at attracting women to coding boot camps, through a Face-to-Face Campaign that will highlight female role models, increase women’s knowledge on coding boot camps and give participants the opportunity to try out a basic coding class. For more information on this study, see the Policy Research Working Paper: https://openknowledge.worldbank.org/handle/10986/27644 FOR MORE INFORMATION, PLEASE CONTACT Markus Goldstein mgoldstein@worldbank.org afrgenderlab@worldbank.org 1818 H St NW Washington, DC 20433 USA This work has also been funded in part by The Umbrella Facility for Gender Equality (UFGE), a World Bank Group multi- donor trust fund expanding evidence, knowledge and data needed to identify and address key gaps between men and women to deliver better development solutions that boost prosperity and increase opportunity for all. The UFGE has received generous contributions from Australia, Canada, Denmark, Finland, Germany, Iceland, Netherlands, Norway, Spain, Sweden, Switzerland, United Kingdom, and the United States. The first draft of this policy brief was released in September 2018.