Policy Research Working Paper 10197 Gender Differences in Socio-Emotional Skills and Economic Outcomes New Evidence from 17 African Countries Kehinde Ajayi Smita Das Clara Delavallade Tigist Ketema Léa Rouanet Africa Region Gender Innovation Lab October 2022 Policy Research Working Paper 10197 Abstract Using data from 41,873 individuals across 17 African coun- in socio-emotional skills increases at higher education levels. tries and 13 studies, this paper maps data from various Socio-emotional skills are associated with higher earnings, self-reported scales to 10 socio-emotional skills and exam- especially for women. However, the specific skills associated ine gender differences in these skills and their relationship with higher earnings differ by gender. Interpersonal skills with education and earnings. Apart from self-control, the are more strongly correlated with earnings for women than findings show a significant male advantage in self-reported for men, and measures of these skills are often underrepre- skills—men have an aggregate socio-emotional skill level sented, which indicates a key direction for future research. 0.151 standard deviations higher than women, equivalent The paper further examines differences in the relationship to the socio-emotional skill gained over 5.6 years of educa- between socio-emotional skills and earnings by levels of tion. This is robust to controlling for positive self-concept. education and occupation. It discusses the implications Closing the gender gap in education would close 17percent of these results for interventions seeking to hone wom- of this gap. While overall socio-emotional skill and edu- en’s socio-emotional skills for labor market success and cation are positively correlated for both men and women, to address the gender norms that may perpetuate gaps in women do not have a positive correlation with education for socio-emotional skills. some individual socio-emotional skills. The male advantage This paper is a product of the Gender Innovation Lab, Africa Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at cdelavallade@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Gender Differences in Socio-Emotional Skills and Economic Outcomes: New Evidence from 17 African Countries1 Kehinde Ajayi, Smita Das, Clara Delavallade, Tigist Ketema and Léa Rouanet Keywords: gender, socio-emotional skills, earnings, education, Africa JEL Classification: J16, J24, 015 1 The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank Group, its Board of Executive Directors, or the governments they represent. We thank colleagues from the Africa Gender Innovation Lab and participants in the GIL and IFPRI Applied Microeconomics and Development seminar and IPA conference for useful suggestions. Tricia Koroknay-Palicz’s and Dawn McDaniel’s contributions were critical for defining the socio-emotional skills under study and for shaping the framing. We thank David Evans, Michael Frese, Steven Glazerman and Markus Goldstein for their thorough reading of the paper and insightful comments. This work was funded by the Wellspring Philanthropic Fund and the Umbrella Facility for Gender Equality. Ajayi: World Bank, kajayi@worldbank.org; Das: IPA, sdas@poverty-action.org; Delavallade: World Bank, cdelavallade@worldbank.org; Ketema: World Bank, tketema@worldbank.org; Rouanet: World Bank, lrouanet@worldbank.org. 1. Introduction Recent evidence from economics, sociology and psychology has pointed to the importance of socio- emotional skills for explaining economic outcomes and productivity (Heckman et al., 2006; Lippman et al., 2015; Roberts et al., 2007; Woolley et al., 2010). Across regions, occupations, and skill levels, employers consistently report an unmet demand for socio-emotional skills among employees and commonly prioritize skills such as communication, customer handling, teamwork, problem solving, perseverance and self-management (Cunningham & Villasenor, 2016; Kautz et al., 2014). Socio-emotional skills training is already often a component of development programs targeting health, education, and employment: vocational and business training, life skills for maternal health or safe spaces, microcredit, and graduation programs to name a few. Programs teaching socio-emotional skills have real potential as policy levers because they may explain part of the gender gap in economic empowerment; they may affect program participation; and they are more malleable at later ages relative to cognitive skills (Almlund et al., 2011; Cunha et al., 2010). However, socio-emotional skills are numerous and may prove difficult to teach sustainably. Thus, the success of programs teaching socio-emotional skills will require identifying which skills matter most for whom and how this varies with context. Socio-emotional skills is a term often used interchangeably with noncognitive skills, 21st century skills, personality traits, and life skills (Duckworth & Yeager, 2015; Heckman & Kautz, 2013; Sanchez Puerta et al., 2016). However, unlike many of these terms, socio-emotional skills do not include beliefs, preferences, values, and attitudes (e.g. optimism); they are not considered static to a particular individual, and they do not include technical knowledge of media, technology, health, finance, and social issues. Socio-emotional skills focuses on a clear list of associated competencies that are considered malleable and transferable across contexts: the Collaborative for Academic, Social and Emotional Learning (CASEL) defines socio- emotional skills as the set of skills used to “manage emotions, set and achieve positive goals, feel and show empathy for others, establish and maintain positive relationships, and make responsible decisions.” By comparison, soft skills generally encompass socio-emotional skills and personality traits.2 2 Although socio-emotional skills differ from personality traits, we refer to work on gender differences in personality traits to complement our discussion in cases where literature on gender differences in socio-emotional skills is lacking. 2 This paper uses individual-level data on 41,873 respondents from 17 Sub-Saharan African countries to examine gender differences in ten socio-emotional skills and their relationship with education and earnings. It makes four main contributions to the literature. To our knowledge, ours is the first paper to analyze gender differences in several dimensions of socio- emotional skills drawing on a large sample of data from multiple African settings, where limited evidence exists. Only a few key studies have extended analysis of gender differences to developing countries, including some in Sub-Saharan Africa (Lynn & Martin, 1997; Costa et al., 2001; Schmitt et al., 2008), and these studies are mostly confined to measures of the Big Five personality traits3. Unexpectedly, these studies find that a higher human development index ranking is associated with larger gender differences in personality traits; the magnitude of gender differences is higher in European and American contexts and is small to nonexistent in Asian and African countries. Based on Cohen’s D, the overall magnitude of the gender difference in our results is similar to results for personality differences in Sub-Saharan Africa (Costa et al., 2001; Schmitt et al., 2008). Our results indicate that men on average have an aggregate socio- emotional skills level that is 0.151 standard deviations higher than the average level among women, equivalent to 5.6 years of education or a 5 percentile increase in socio-emotional skills levels. This male advantage persists after controlling for education, age, marital status and even measures associated with confidence, such as self-esteem and self-efficacy. However, the range of gender differences across skills and studies is considerable, and most would be categorized as small or moderate based on a meta-analysis of gender differences in skills (Hyde, 2005). A large literature from developed countries has examined whether gender stereotypes are reflected in specific socio-emotional skills gender differences, which we summarize in Appendix Table A1. Eagly and Wood’s (2012) cross-disciplinary social role theory predicts that men have an advantage in agentic skills while women have an advantage in communal skills, though this may vary with local norms and occupational segregation. Focusing on concepts associated with the ten specific socio-emotional skills used in this study, trends in the literature suggest that women score lower than men on self-reported measures of positive self-concept, emotional regulation and higher on measures of self-control, empathy, and interpersonal relatedness. Gender differences are mixed, small in magnitude, or weak in evidence for perseverance, personal initiative, problem solving, expressiveness, and collaboration (see Appendix Table 3 These are five factors that often arise in the analysis of personality traits. They include Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. 3 A1). Here, we confirm a male advantage in agentic skills such as personal initiative. Contrary to expectations, men score higher on the communal skill of teamwork and have no robust advantage in expressiveness. Women are not found to have an advantage in communal skills. The second contribution of this paper is to investigate the relationship between socio-emotional skills and education. Our results are in line with literature demonstrating a positive association between socio- emotional skills and educational attainment (Heckman et al., 2006; Almlund et al., 2011; Taylor et al., 2017; Kraft & Grace, 2015; Mammadov, 2021). While we know that gender inequality in employment outcomes differs across the educational distribution, there is very little evidence on gender differences in the relationship between socio-emotional skills and education. Bridging this evidence gap has important policy implications for what works in school to reduce gender differences in socio-emotional skills and in turn, employment outcomes. We find that gender differences in socio-emotional skills are partially explained by lower education levels among women. Closing the gender gap in education would close about 17% of the socio-emotional skills gender gap. If overall, education is associated with similar gains in overall socio-emotional skills levels for men and women, interestingly, the gender gap in interpersonal skills increases with education attainment. Women may gain less from education than men in terms of interpersonal skills, such that the gender gap in interpersonal skills is only significant for more educated individuals. Alternatively, more highly educated women may encounter post-education social norms or work environments that limit their use and development of interpersonal skills. The third contribution of this paper is to analyze associations between socio-emotional skills and earnings and their gender specificities. Given that one can expect gender-differentiated returns to earnings to vary across skills, it is important to identify which skills matter most for whom to inform the design of future programs teaching socio-emotional skills in Sub-Saharan Africa. While a large literature has documented the link between economic outcomes and socio-emotional skills, particularly in developed countries (Almlund et al., 2011; Heckman et al., 2006; Kautz et al., 2014; Lindqvist & Vestman, 2009; Lippman et al., 2015; Roberts et al., 2007), and more recently in developing countries (Campos et al., 2017; Koop et al., 2000; Krauss et al.,2005, on the impact of personal initiative trainings, Gielnik et al., 2015 on the impact of a STEP training; Groh et al., 2016 in Jordan and Chioda et al., 2021 in Uganda on the impact of a soft skills training), we still lack crucial evidence on gender-specific returns to specific socio-emotional skills in a developing country setting. Grouping socio-emotional skills into one category often results in the use of non-comprehensive measures such that results from various studies are not comparable to each other. 4 While these studies demonstrate the importance of socio-emotional skills, they do not allow insights into programming decisions and how various socio-emotional skills relate to productivity, occupational choice, and discrimination. A seminal study by Heckman et al. (2006) finds that returns to soft skills – utilizing measures of self-esteem and locus of control - are greater for women relative to men in the United States, both for likelihood of employment and earnings. By contrast, a recent study by Exley et al. (2020) highlights conditions under which women’s negotiations may yield negative returns in a US laboratory environment. Additionally, there are several limitations in drawing conclusions from studies based in Western contexts. (i) There is a focus on formal wage employment, rather than the self-employment and informal work common to a developing context. (ii) Studies of entrepreneurship may differ, as self-employment in developing contexts is more often pursued out of necessity and a lack of formal job options, rather than out of desire. (iii) Local social norms and beliefs may affect the cultivation and value placed on particular socio-emotional skills. (iv) Finally, measurement of socio-emotional skills may differ as the behaviors associated with skills may vary with culture. We show that socio-emotional skills are robustly associated with higher earnings for both men and women in Sub-Saharan Africa, and for women, this holds with a wider set of skills. In this way, our results are similar to an examination of nine middle-income countries, where women were found to have a positive correlation between earnings and openness, emotional stability, conscientiousness, and extraversion; while for men, this positive correlation existed for openness alone (Gunewardena et al., 2018). While we do not find evidence of gender differences in correlations between an aggregate index of socio-emotional skills and earnings, we find that positive self-concept is associated with lower earnings for women relative to men, while interpersonal skills are associated with higher earnings. As noted earlier, we focus on socio-emotional skills in our analysis. Finally, this paper contributes to the literature by examining how the relationship between socio- emotional skills and earnings varies by educational attainment, which we interpret as a proxy for the qualification level of the occupation, and by the occupational sector. Few papers have explored this variation. We find that the association between aggregate socio-emotional skills and earnings is higher for more educated women and less educated men, though results differ substantially with the particular skill involved. At higher levels of education, while the relative advantage women have in the association 5 with earnings is highest for teamwork, gender gaps in levels are wider. This calls for education policies that address the gender gap in the acquisition of teamwork and other interpersonal skills. The evidence on the positive relationship between socio-emotional skills and earnings is stronger for non-agricultural self-employment, relative to agriculture and wage employment. The remainder of the paper proceeds as follows. Section 2 presents the conceptual framework for this study. Section 3 introduces the data, presents summary statistics and lays out our empirical strategy. Section 4 presents our main results on gender gaps in socio-emotional skills and on the gender-specific relationships between socio-emotional skills and education on the one hand and socio-emotional skills and earnings on the other hand. Section 5 provides some robustness checks on these results, while Section 6 concludes with policy implications and suggestions for further research. 2. Conceptual Framework and Existing Evidence In this section, we present a simple framework of the relationship between education, socio-emotional skills and economic outcomes, and how these links may vary across gender, in order to provide the economic intuition behind our empirical tests. Figure 1 presents a simplified visualization of how socio- emotional skills, educational attainment and earnings interact. Figure 1: Conceptual Framework: Socio-emotional skills, Education and Earnings 6 2.1 Socio-emotional skills and education In general, education is found to correlate positively with socio-emotional skills, though results may not be monotonic (Heckman et al., 2006). First, there is evidence on school sorting under socio-emotional skills whereby individuals with low endowments in socio-emotional skills are less likely to enter school and more likely to exit (Conti et al., 2010; Acosta et al. 2020; Papageorge et al. 2019). Second, the potential of socio-emotional skills training to improve attendance, educational attainment, and academic achievement is well documented, with important implications for potential earnings (Heckman et al., 2006; Almlund et al., 2011; Durlak et al., 2011; Lippman et al., 2015; OECD, 2015; Taylor et al., 2017; DiPrete & Jennings, 2012). In a study of elementary school children in the United States, Diprete & Jennings (2012) find that girls have higher levels of soft skills, but there is no gender difference in academic returns to these skills. In Uganda, Chioda et al. (2021) show that graduates from a soft-skill intensive training were more likely to have graduated from secondary school and female graduates were also more likely to be enrolled in or to have completed tertiary education. However, the impact of socio-emotional skills on academic outcomes is by no means guaranteed and varies with the particular skill and academic indicator analyzed (Smithers et al., 2018; OECD, 2021). Conscientiousness, and related concepts such as self-discipline, persistence, grit, have been particularly predictive of academic performance (Almund et al., 2011; Duckworth et al., 2007; Duckworth & Segilman, 2005; OECD, 2021; Mammadov, 2021), possibly due to their impact on study habits, effort, and prosocial behavior (Almund et al., 2011). Other mindset interventions have been found to create a belief in malleable intelligence, transform the attribution of setbacks, affirm values, or improve aspirations, and thereby motivate students, increase use of learning strategies, mitigate stereotype threat, encourage belongingness, and reduce absences and disruptive behavior (Yeager & Walton, 2011; Farrington et al., 2012). While self-esteem has had mixed results on academic performance, results may be positive in cases where it is tied to a positive attitude which enhances incentives (Mohanty, 2009), facilitates persistence and initiative (Baumeister et al., 2003), or reduces risky behavior (Heckman et al., 2006; Guerra et al., 2014). However, few other socio-emotional skills have been analyzed in detail, with little to no examination of social skills (Farrington et al., 2012). Though the causal influence of education on socio-emotional skills is promising, demonstrating a rigorous causal relationship is logistically difficult. However, a number of studies suggest the causal mechanisms at play. In the US, Jackson et al. (2020) found that high schools impact students' self-reported socio- emotional development by enhancing social well-being and promoting hard work. Their results suggest that socio-emotional skills can be fostered by schools to improve longer-run outcomes. Schooling offers 7 the opportunity for individuals to develop attention and self-discipline, regulate stress as they navigate academic and social problem solving, develop teamwork and empathy skills via cooperative learning (Guerra et al., 2014; Durlak et al., 2011), and grow self-esteem from academic performance (Baumeister et al., 2003). Moreover, exposure to ideas and peers may affect academic performance, socialization, and forward-looking behavior (Duflo et al., 2011; Villaseñor, 2018; Bernard et al., 2014). Teachers and school leaders can affect students’ socio-emotional skills through their relationships with students, the behaviors they model, and the classroom environments they create (Jones et al., 2013). Both positive and negative impacts have been found on students' complex task performance, growth mindset (Kraft & Grace, 2016), self-efficacy (Blazar and Kraft, 2017), school culture, classroom behavior, peer interactions, emotional support, and motivation (Loeb et al., 2019; Villaseñor, 2018). These teacher effects on socio-emotional skills are comparable to those on academic achievement (Villaseñor, 2018). Thus, in a setting such as sub- Saharan Africa where educational attainment remains lower for girls, girls may have less opportunity to develop foundational skills. Moreover, educational institutions may be a source of instilling social norms and beliefs that affect one’s skill acquisition. Girls may have a poorer sense of belonging (OECD, 2021) and be less encouraged to speak up or interact in an assertive manner and more penalized when doing so. These factors may lead them to practice certain socio-emotional skills, especially interpersonal skills, less than boys (Amanatullah and Morris, 2010), which may affect their individual beliefs and future choice of activities, goal levels, commitment, and persistence (Gielnick et al., 2015). In turn, individuals’ beliefs may also affect their likelihood of using and cultivating particular socio-emotional skills, and thus affect their socio-emotional skills levels in adulthood: if women are less likely to believe their actions will obtain results, it reduces their motivation to develop and hone skills such as perseverance, personal initiative, expressiveness, influence, and negotiation. Correll (2001, 2004) found that if individuals are told that men are better than women at a task, they will assess themselves and select a career accordingly, despite equal ability and aspirations. 2.2 Socio-emotional skills and economic outcomes Conceptually, socio-emotional skills may contribute indirectly to economic outcomes (labor market participation and earnings) through educational outcomes, which in turn contributes to the building of other skills (cognitive or technical). More directly, socio-emotional skills may be key to the development, 8 planning, and realization of goals across contexts4. The evidence base on the relationship between socio- emotional skills and economic outcomes, and how it can differ by gender, is rapidly expanding. Studies based in Western contexts have examined the relationship between soft skills and economic outcomes through many stages of the employment process, though these often focus on personality traits and beliefs rather than socio-emotional skills. In the hiring process, extraversion, conscientiousness, and lower neuroticism are correlated with positive interviews and job recommendations (Roberts et al., 2007). Social skills, self-control, and low irritability have been found to be protective against unemployment (Lippman et al., 2015; Roberts et al., 2007). Job performance has been linked to conscientiousness, regardless of job complexity (Kautz et al., 2014; Almlund et al., 2011; Judge & Ilies, 2002), and to positive emotions, which are associated with creative problem solving (Roberts et al., 2007). Among men in Sweden, soft skills such as persistence, social skills, and emotional stability, as measured in a psychological interview, correlate with employment and wage across occupations. Moreover, these skills mattered more than cognitive ability among unqualified workers and managers (Lindqvist & Vestman, 2009). Recent evidence has expanded to developing contexts and varied socio-emotional skills exogenously to isolate the causal impact of socio-emotional skills on economic outcomes. Personal initiative trainings aimed to instill self-management skills have become recognized as an effective way to build an entrepreneurial proactive mindset among farmers and entrepreneurs and increase profits (Campos et al., 2017; Koop et al., 2000; Krauss et al.,2005). This training was found to be effective for women regardless of their educational background (Campos et al., 2017). This contrasts with the null impact of a soft skills training program on female youth employment among community college graduates in Jordan (Groh et al., 2016). STEP, a program targeting self-efficacy among younger entrepreneurs who have not yet launched their businesses, has shown positive impacts in Uganda, Kenya, and Mexico. STEP students start 34% more businesses one year after the training and 20% more businesses two years after the training, and they create 35% additional jobs two years after the training (Gielnik et al., 2015). In Uganda, Chioda et al. (2021) tested the effects of a skill upgrade by introducing hard skills and soft skills in a 3-week mini- 4 More specifically, developing an action plan may require the self-awareness to set desirable and attainable goals, and the problem solving, decision making and social awareness to anticipate potential obstacles and plan accordingly. Goal attainment may require the emotional regulation and perseverance to transcend obstacles, the expressiveness and respectful communication to share one’s desires and ask for help, the self -control to stay on track with goals, the persuasion and negotiation skills to navigate business relationships and market prices, and the networking and collaboration skills to build market linkages and find resources or opportunities. 9 MBA training to high school students while varying the intensity of the soft skills students received. After three and half years, students in both groups showed an increase in both soft and hard skills, while only training in soft skills was linked to gains in self-efficacy, persuasion, and negotiation. Further, students who received the skill upgrade had substantially higher earnings and were more likely to start enterprises and have enterprises with higher survival rates. The training also led to larger profits and business capital investments. In theory, socio-emotional skills-building policies may have particularly high economic returns for women in Sub-Saharan Africa, as women face more barriers to success in the workforce and binding social norms that limit expressiveness, mobility, time, network formation, and occupational choice (Chakravarty et al., 2017). Perseverance and creativity in problem solving may be essential since women bear more responsibility for daily survival and short-term budgeting, and women may be more mentally taxed and have less financial flexibility (Friedson-Ridenour & Pierotti, 2019; Schilbach et al., 2016). Communication, persuasion, and conflict resolution skills may improve women’s ability to navigate home-based barriers as they request and obtain the support of family members, influence fertility decisions, and negotiate the allocation of household assets and responsibilities. At work, higher levels of relationship building skills, teamwork and personal initiative may be required to expand networks, obtain information and find opportunities, develop skills, navigate inclusion in male-dominated occupations, and find allies for both business and emotional support. Conversely, the relationship between socio-emotional skills and economic outcomes could be weaker for women if they have lower expected returns, or they are less likely to cultivate and practice socio-emotional skills due to restrictions on mobility, socialization, and labor force participation. Evidence on the link between gender, socio-emotional skills, and earnings is still scant. A longitudinal study from the US shows that moving an individual from the 25th to the 75th percentile of non-cognitive ability at age 14 to 21 is associated with males’ and females’ wages being 10 and 30 points higher and their probability of employment being 15 and 40 points higher, respectively (Heckman et al. 2006). In this study, women’s economic outcomes are thus more strongly correlated with non-cognitive skills. Nyhus & Pons (2012) found that personality traits reduced the unemployed portion of the gender-wage gap by 12.% percentage points. A set of studies have documented gender differences in the returns to specific skills, though results differ with culture and occupation. Among Wisconsin graduates, men alone were rewarded for antagonism (the opposite of agreeableness) (Mueller & Plug, 2006), whereas management positions in Australia were associated with non-agreeableness regardless of 10 gender. However, these positions were associated with conscientiousness among men and openness and extroversion among women (Cobb-Clark & Tan, 2011). One might expect that returns to socio-emotional skills may differ across sectors and by education level. Those with lower education levels may have fewer job options with less decision-making power, such that possession of certain socio-emotional skills cannot be used productively. For example, repetitive wage or agricultural work may be expected to require listening, perseverance, and self-control. However, employment that allows decision-making or working with others may observe returns if individuals can use personal initiative to adopt new methods or explore new markets, use negotiation to obtain higher prices or access cheaper inputs, or use interpersonal skills to influence family members or collaborate with other businesses. Gender-differentiated returns to socio-emotional skills may be less prevalent among those with lower education levels or those residing in rural economies, as there may be fewer job options and less gender-based occupational segregation (Das & Kotikula, 2019). Those with lower education levels may face stronger social norms that limit the returns to particular socio-emotional skills. Alternatively, as socio-emotional skills are transferable across occupations and relevant for economic empowerment in the home and workplace, returns to socio-emotional skills may not differ by education or occupation. A few studies have found variation in returns to personality at higher education levels. Gensowski (2018) found that the correlation of personality and earnings is higher for men with a graduate degree than those with a bachelor’s or less. Nyhus & Pons (2005) found that extraversion was less punished, and autonomy was less rewarded among men with some university education, while women were punished for emotional stability at both pre-university/vocational and university education levels. In the United Kingdom, Carneiro et al. (2007) found that returns to social adjustment levels at age 11 were positive and did not vary with education. There is also some evidence that women with high-status occupations faced earnings penalties for being aggressive (Bowles et al., 2001). Three non-experimental papers provide evidence on the relationship between soft skills and earnings in the agriculture sector from Africa. In Malawi, an increase in women’s non-cognitive ability (such as perseverance, passion for work, and optimism) was correlated with higher rates of adoption of valuable cash crops (Montalvao et al., 2017). In Ghana, non-cognitive skills were found to increase technology adoption and technical efficiency on rice farms (Ali et al., 2020). In Côte d’Ivoire, the relative self-esteem 11 of rural spouses determines who is earning income from cash crop agriculture through increased control over household land, with women’s outcomes being more sensitive to self-esteem levels (Botea et al., 2021). 3. Sample Data and Empirical Strategy 3.1 Data set construction We started by compiling data from all baseline surveys conducted by the World Bank’s Africa Gender Innovation Lab (GIL) which included any measures of socio-emotional skills and economic outcomes for both men and women. This approach yielded a database of 10 studies in 8 countries (Benin, the Republic of Congo, Côte d’Ivoire, Ghana, Kenya, Mozambique, Nigeria, and Togo) conducted between 2013 and 2020. To broaden the scope of our analysis, we added data for the two African countries included in the World Bank’s STEP Skills Measurement Program – Ghana and Kenya, as well as data from the Future of Business Survey, which was a co-product of Facebook, OECD and World Bank (which includes respondents from Angola, Benin, Botswana, Côte d’Ivoire, Cameroon, Ethiopia, Ghana, Kenya, Mozambique, Nigeria, Senegal, Tanzania, Uganda, South Africa, and Zambia). Appendix Table A2 summarizes the database of studies included in our sample. Since each study used different scales and measures of skills, it was necessary to align the data under a common framework. We utilized a framework developed by GIL as part of its larger research agenda to examine the importance of socio-emotional skills for women’s economic outcomes in Sub Saharan Africa. To establish this framework, GIL worked with psychologists and used existing literature to develop a list of skills that are malleable, span the range of socio-emotional skills while mapping to existing frameworks5, can be used to examine gender differences, and include higher-order categories that allow for separate treatment arms to rigorously unpack the causal influence of particular socio-emotional skills. We appended the data and mapped each of the survey items to one of these 14 socio-emotional skills. This categorization process was independently conducted prior to the analysis by a psychologist well- versed in the definitions of the skills, in close collaboration with our research team. The resulting dataset includes measures for 10 of the 14 skills. The remaining four skills were not included because of their absence from the 13 studies in our sample: emotional awareness, listening, interpersonal influence, and 5 Several groups and organizations have developed frameworks to classify and organize socio-emotional skills and overlapping concepts. Three commonly used frameworks are CASEL, USAID Youth Power, and Big 5 Personality, and Harvard’s EASEL Lab documents many others. 12 negotiation.6 Appendix Table A3 provides a definition for each of the 10 skills under study as well as example items aggregated to construct each skill measure. The complete list of items combined to construct a given skill measure is available on request. In order to create an index for each skill, we first reversed all socio-emotional skills measures such that higher numbers indicate "higher" socio-emotional skills levels. Second, we standardized each item such that the mean is zero and the standard deviation is 1 for women in each study. We then took an average of the items for a given skill based on the categorization exercise. Finally, we standardized the index to the women’s sample in each study. This allows us to later interpret regression coefficients on gender differences relative to the variation between women. Additionally, we created three aggregated skills measures: Intra that averages all intrapersonal skill indices available in a given study (positive self-concept, emotional regulation, self-control, perseverance, personal initiative and problem-solving and decision-making (PSDM)), Inter that averages all interpersonal skill indices available in a given study (empathy, expressiveness, interpersonal relatedness and teamwork), and All that averages all socio-emotional skill indices available in a given study. Our analysis thus considers 13 measures of socio-emotional skills (10 skill-specific and 3 aggregated measures). 3.2 Sample description Our pooled sample comprises 41,873 individual-level observations from surveys (see Appendix Table A2 for sampling criteria) conducted in 8 Sub-Saharan African countries (Benin, the Republic of Congo, Côte d’Ivoire, Ghana, Kenya, Mozambique, Nigeria, and Togo) or under the Future of Business Survey (15 Sub- Saharan African countries). While all studies target both men and women, and more specifically couples in Côte d’Ivoire and Mozambique, only 42.7% of the respondents are women. Most studies’ respondents are adults and young adults, in both urban and rural areas, with an average age of 36 years old (23 in the Republic of Congo skills project to 49 in the Future of Business project). Very few studies limit sample selection by education level and the average respondent has completed 9.6 years of education (a minimum average of 2.6 years in Mozambique and a maximum of 13.3 years in Nigeria). Among the respondents, 57.3% are married. Of the respondents, 74.5% report earning a positive income, with some variation across studies (43% in Mozambique up to 98% in Togo’s Private Sector Development Project). Respondents earn on average $138 (USD) monthly with the average income varying between $2 in 6 Self Awareness in the framework was renamed Positive Self Concept and Collaboration from the framework was renamed Teamwork, as the latter terms are similar, but better reflected the specific data that was available. 13 Mozambique and $504 in Ghana (see Table 1). The Future of Business Survey is the only study in the data set with no data on earnings. Apart from the STEP surveys, samples used for this analysis are typically not representative of the general population but come from selected subpopulations. This implies that results are not always externally valid, but representative surveys very rarely contain detailed measures of socio-emotional skills. Table 2 shows summary statistics on the 13 socio-emotional skills measures for the pooled sample (Panel A) and at the study-level (Panel B). Overall, 323 items were used to construct the socio-emotional skills measures. Importantly, more than 70% of socio-emotional skills items refer to intrapersonal skills, which reflects the larger effort made to collect data on intrapersonal skills (mostly positive self-concept, personal initiative and PSDM). In comparison, interpersonal skills, and especially empathy and teamwork, suffer from a thinner information base. This translates into a lower number of observations being used for interpersonal skills and a lower reliability in these indices as indicated by systematically lower Cronbach’s alpha across studies compared to intrapersonal skills. Panel B indicates for each study which socio- emotional skills measures are available and the number of items they are based upon with summary statistics when available. Note that each study measures between 2 and 9 specific socio-emotional skills (excluding aggregate measures), the average study measuring 5.7 specific socio-emotional skills: 3.8 intrapersonal skills (out of 6) and 1.8 interpersonal skills (out of 4). 3.3 Estimating gender differences in socio-emotional skill levels To measure gender differences in socio-emotional skill levels, we regress each of the 13 measures of socio- emotional skills on the gender of the respondent controlling for a set of project dummies, using the pooled sample described above. = + + + (1) Where: - is the socio-emotional skill measure for individual i, either skill-specific or aggregated. - is a dummy equal to one if the individual i in study s is a woman. - is a study fixed effect. 14 However, in model (1) the estimated gender difference coefficient might also be capturing differences in age, marital status or education. To account for these confounding factors, we re-run model (1) four times by adding each of these variables one at a time as a control variable. See equation (2) below. ′ = + + + + (2) Where: - is a vector of control variables including age, marital status, education (highest educational attainment), and employment status of individual i.7 We run a weighted regression to account for the fact that sample sizes vary a lot across skills. As can be seen in Table 2, sample sizes vary from 39,885 to 8,260 observations for the individual skills. Without weights, for instance, Empathy would be given more weight in the pooled sample than Perseverance, despite having almost a fifth of the observations. To avoid skills with larger samples contributing more to the analysis, we create sampling weights, which are equal to the inverse of the probability that a given observation is included in our analysis sample. The OLS estimates of give the conditional gender differences for each skill. 3.4 Heterogeneity in gender differences by education To look at the heterogeneity of gender differences in socio-emotional skill levels, we regress each of the 13 measures of socio-emotional skills on the interaction of the gender of the respondent with education variables, years of education and a vector of dummy variables of transitional grades (ever entered lower secondary, ever entered senior secondary and ever entered higher education), using the pooled dataset, in two sets of regressions.8 ′ = + 1 + 2 + 3 × + + + (3) 7 Age bins represents dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5-year gap, 0 otherwise. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Education dummies represent dummy variables equal to 1 if the respondent's highest educational attainment (completed) is 0, 1, ... or 14, where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed high school, 13=completed certificate or dip loma and 14=completed university degree or above, 0 otherwise. Employment is a dummy variable equal to 1 if the respondent is currently working, 0 otherwise. 8 We examine heterogeneity by years of education first and then add a vector of dummy variables for transitioning grades in a different regression to allow us to estimate the relative implications of each set of factors. 15 Where: - is the number of years of education of individual i. - controls for marital status and a vector of dummy variables of age bins equal to one if individual i’s age belongs to the age cohort ranges from 15 to 65 with a 5-year gap. - is a study fixed effect. The coefficient 1 in equations 3 gives the difference in socio-emotional skills levels between less educated women and less educated men. Again, 3 stands for the differential gender gap in socio- emotional skills levels between more and less educated populations. In other words, a significantly positive 3 indicates that gender differences are significantly stronger among more educated people. 1 + 3 estimates the difference in socio-emotional skills levels between men and women among the educated people. 3.5 Correlation between socio-emotional skills and employment outcomes Differences by gender For each of the 13 measures of socio-emotional skills, we run a regression on the full sample of men and women, regressing an employment outcome on the interaction of the socio-emotional skill measure with the gender of the respondent. = + 1 + 2 + 3 × ′ +4 + 5 × + + + (4) Where: - is monthly earnings defined as the inverse hyperbolic sine (IHS) transformation of the respondent’s monthly earnings in US dollars. - is a vector of control variables including age bins/cohorts and respondent’s marital status. - is the number of years of education of individual i. - is a study fixed effect. The coefficient 1 in equation 4 gives the difference in earnings between women and men with a given socio-emotional skills. 2 gives the correlation between the socio-emotional skills and any earnings for men, while 2 + 3 gives the same correlation for women. 3 indicates whether having any earnings is 16 more or less correlated with for women than men. On the other hand, 4 and 4 + 5 shows the correlation between years of education and any earnings for men and women, respectively. Heterogeneity by education To assess differences in returns to socio-emotional skills by education, we regress earnings on the interaction of each of the 13 socio-emotional skill measure with respondent’s years of education. = + 1 + 2 + 3 × + 4 + 5 × ′ + 6 × + 7 × × + + + (5) Where: - is monthly earnings defined as the inverse hyperbolic sine (IHS) transformation of the respondent’s monthly earnings in US dollars. - is a vector of control variables including age bins/cohorts and respondent’s marital status. - is the number of years of education of individual i. - is a study fixed effect. In equation 5, the coefficient 1 gives the difference in earnings between women and men with a given socio-emotional skills. 2 gives the correlation between educational attainment and any earnings for men, while 2 + 3 gives the same correlation for women. Similarly, 4 and 4 + 5 shows the correlation between socio-emotional skills and any earnings for men and women, respectively. Finally, 6 gives the correlation between socio-emotional skills and any earnings for men with a given years of education while and 6 + 7 gives the same correlation for women. Heterogeneity across types of employment After dividing our sample into two groups based on whether the respondent is self-employed in the non- agricultural sector or wage employed, we re-run equation 4 and assess whether results vary based on employment type. 17 4. Results 4.1 Gender differences in levels of socio-emotional skills Table 3 and Figure 2 report estimates of the average gender difference in socio-emotional skills. Following equation (1), we begin by estimating the gender gap with project fixed effects in Model A. Our preferred specification is reported in Model D and follows equation (2), with additional controls for age, marital status, education.9 We find that women have significantly lower levels of socio-emotional skills, even after controlling for education and demographic characteristics. The gender gap in socio-emotional skills decreases from 0.18 (model A) to 0.151 when adding the full set of controls (model D). Finally, in Model E, we add controls for employment status and the results do not substantially change, indicating that the gender differences in socio-emotional skills we find are not driven by women’s lower likelihood of engaging in paid work. Whereas controlling for education does change some results meaningfully, controlling for employment does not. This suggests that employment does not substantially develop socio-emotional skills in this context. We explore the relationship between socio-emotional skills and employment in more detail in section 4.3 of the paper. To provide some context on the magnitude of the gender gap in our preferred specification in Model D, the conditional gender gap is equivalent to 5.6 years of education. As a comparison, the gender gap in education in our sample, after controlling for age, marital status and project fixed effects, is 0.96 years. Therefore, closing the gender gap in education would close about 17% of the socio-emotional skills gender gap. Additionally, the gender gap is equivalent to a 5 percentile increase in socio-emotional skills levels, based on the women’s distribution. This gap is similar without project fixed effects, indicating that the gap holds across the whole sample and is not only due to within-country variations. One factor to note is that sample sizes differ across the specific individual skills, we therefore take the aggregate measures as our preferred estimates of general differences in skills and conduct robustness checks in section 5 to examine how sensitive gender gaps in individual skills are to skills measures. We confirm that these main results with our aggregate measures are unlikely to be driven by sample selection, using data from the Ghana and Kenya STEP surveys, which include a random sample of respondents drawn from a general urban population. Consistent with our pooled results, we find in the STEP surveys that men have significantly 9 The findings below are not sensitive to additionally controlling for number of children. Because of the reduced sample size, these results are not presented here but available upon request. 18 higher levels of socio-emotional skills, with a larger gender gap in intrapersonal compared to interpersonal skills (we report a full set of results for individual projects in Table 8). With this caveat in mind, looking into intrapersonal skills, interpersonal skills, and individual skills still yields informative conclusions. The conditional gender gap is larger for intrapersonal skills (0.143) than for interpersonal skills (0.104). Apart from self-control, there is a gender gap favoring men for all individual skills, significant at the 1% level (5% level for empathy). It is largest for emotional regulation, personal initiative, problem-solving and decision-making and teamwork; it is smallest for positive self-concept, empathy, expressiveness, and interpersonal relatedness. In addition, the portion of the gender gap in socio-emotional skills which can be associated with education varies across skills: it is largest for positive self-concept, problem-solving, and empathy while lowest for teamwork. We will dig deeper into the analysis of the correlation between gaps in socio-emotional skills and education in the following tables. 4.2 Educational attainment and socio-emotional skills, by gender Having estimated gender differences in socio-emotional skills for the full sample, we now turn to examine heterogeneity by education. Gender inequality in employment outcomes and economic opportunity differs across the educational distribution. Understanding the associated differences in socio-emotional skills could therefore have important policy implications. Table 4 and Figure 3 report estimates of gender- and education-related differences in socio-emotional skills while allowing for different associations between education and socio-emotional skills for men and women, following equation (3). Column 1 in Panel A reports results using our overall aggregate index measure of socio-emotional skills. The coefficient on the women indicator is -0.121, indicating that among uneducated people, women’s level of socio-emotional skills is 0.121 standard deviations lower than men’s on average. This is more than four times the difference associated with completing an additional year of education. Education is associated with similar gains in overall socio-emotional skills levels for men and women, the gender difference in education returns is statistically insignificant (Panel A, Column 1). Columns 2 and 3 in Panel A separately report estimates for aggregate indices of intrapersonal and interpersonal skills. For people with no formal schooling, the gender gap for intrapersonal skills is significant at 0.129, while the gender gap in interpersonal skills is not. Interestingly however, the gender 19 gap in interpersonal skills increases with educational attainment. Women gain less from education than men in terms of interpersonal skills, so that the gap in interpersonal skills is only significant for more educated women (with three years of education or more). Panel B individually reports estimates for each intrapersonal skill, while Panel C similarly reports estimates for each interpersonal skill. Among people with no formal schooling, gender differences are statistically significant at the 1-percent level for perseverance (-0.128), personal initiative (-0.125) problem-solving and decision making (-0.126), while women conversely display significantly higher self-control (0.101) and expressiveness (0.172) levels than men. For each individual interpersonal skill, as well as for emotional regulation and self-control, the gender gap increases with educational attainment. Women’s marginal returns to education are more than half those of men for emotional regulation, empathy and teamwork, while women’s levels of expressiveness do not even correlate with educational attainment. Taken together, this analysis suggests that the gender gap persists even with education. Education actually explains little of the overall gender gap, since both men’s and women’s socio-emotional skills increase with education. However, looking into individual skills, the relationship between socio-emotional skills and education varies by gender. An alternative possibility is that men and women equally acquire these skills, but their persistence over time decreases more for women, potentially due to gender norms. 4.3 Correlations with earnings Our analysis so far has established that men have significantly higher levels of socio-emotional skills and that these skills differences persist even among men and women with formal education. To understand implications for gender differences in economic outcomes, we now turn to examine the relationship between socio-emotional skills and earnings. Figure 4 summarizes the main correlations, displaying regression coefficients from Table 5 with their confidence intervals and significance levels. Average effects We present estimates of equation (4) with monthly earnings as an outcome in Table 5. Here we restrict our sample to the 33,965 individuals for whom we have information on employment and earnings. Panel A indicates that women in our sample have almost 70% lower earnings than men, consistent with Arbache 20 et al. (2010).10 This unconditional gender gap in earnings decreases to 56% after accounting for education and demographics. Both men and women experience positive returns to education. In aggregate, socio-emotional skills are associated with higher earnings for both men and women in similar magnitude. The gender difference in the relationship between socio-emotional skills and earnings is statistically insignificant11 (Panel B, Column1). We find some gender differences when we look separately at intrapersonal and interpersonal skills. For men, earnings correlate significantly with intrapersonal skills (0.081 standard deviations in Panel B, Column 2), while the correlation with interpersonal skills is insignificant (Panel B, Column 3). For women, earnings are significantly correlated with both intrapersonal and interpersonal skills, with coefficients of 0.075 and 0.061, respectively. Our results indicate that a 1 standard deviation increase in interpersonal skills (but not in intrapersonal skills) reduces the gender gap in earnings by about 7.8%. Altogether, we find that one standard deviation increase in socio-emotional skills amounts a 6.8% gain in earnings for men, equivalent to the earnings gain from 2.72 more years of education (0.068/0.025). For women, one standard deviation increase in socio-emotional skills provides an 8.8% increase in earnings, the equivalent earning gain from 5.87 more years of education ((0.068+0.02)/(0.025-0.01)). Our results imply that conditional on similar levels of education for men and women, we would need to increase socio-emotional skills by 9.5 standard deviations (0.647/0.068) to close the gender gap in earnings. Looking at individual intrapersonal skills (Panel C), we find that positive self-concept and perseverance are positively associated with earnings for both men and women, although two important gender differences arise in the magnitudes of these correlations. While high levels of positive self-concept are less associated with earnings for women than for men, perseverance has a significantly higher correlation with earnings for women. Strikingly again, for all other individual and aggregated skills, gender differences are statistically insignificant. These results potentially highlight the positive impacts of overcoming gender norms for women. 10 We estimate a 33% earnings gap when we include men and women with non-zero earnings, which is similar to estimates from other studies. 11 This result holds whether controlling for education levels or not and after controlling for occupation sector, using data from the Ghana and Kenya STEP surveys, which include information on occupation for a random sample drawn from a general urban population. 21 On interpersonal skills (Panel D), we find some positive returns for women only. Expressiveness and teamwork are the only interpersonal socio-emotional skills that are positively associated with earnings and we find this positive association for women but not for men. Future research should study why women get higher returns to these skills in the labor market, and whether this relates to the key differences observed between men’s and women’s networks (World Bank Group, 2019). These are also the two skills that are lower for more educated women, which implies that current education systems tend not to equip women with the socio-emotional skills that could benefit them the most in earnings. By contrast, skills for which the gender gap is lowest have lower returns in earnings (empathy, expressiveness, interpersonal relatedness). Interpersonal relatedness is the only skill with negative returns to earnings for men, and there are no returns for women. We find no significant returns to empathy, although this skill is still more positively associated with earnings for women than for men. Overall, our results suggest that teamwork is the skill most strongly correlated with earnings for women, as well as a skill with high gender gap in levels (although this skill also tends to be less precisely measured in our data). Altogether, women are best rewarded for skills that align with gender stereotypes, but which are also those they gain the least through education, compared to men. The misfit between socio-emotional skills that women gain from schooling (or which schools retain) and those that have highest returns suggests the importance of investing in specific socio-emotional skills trainings to build those particular socio- emotional skills for which women have higher returns (interpersonal skills and perseverance). We also examine differences in employment on the intensive margin, restricting our sample to the 23,387 individuals with positive earnings (Appendix Table A4). Here we again find that in aggregate, socio- emotional skills are associated with higher earnings for both men and women in similar magnitude. The gender difference in the relationship between socio-emotional skills and earnings is statistically insignificant (Panel A, Column1). When we look separately at intrapersonal and interpersonal skills our results differ somewhat. For men, earnings correlate significantly with intrapersonal skills (0.081 standard deviations in Panel A, Column 2), while the correlation with interpersonal skills is smaller but now marginally significant (0.031 standard deviations in Panel A, Column 3). For women, earnings are significantly correlated with both intrapersonal and interpersonal skills, with no significant gender difference. 22 Heterogeneity by education We turn next to analyzing whether correlations between socio-emotional skills and earnings differ by educational attainment. Table 6 and Figure 5 present these results of estimating the specification outlined in equation (5). We find significant gender differences among workers with no formal education. Men with no formal schooling get positive returns to intrapersonal skills and negative returns to interpersonal skills. Women with no formal schooling get no returns to either skill, with significantly lower returns to intrapersonal skills and higher returns to interpersonal skills than men. Education offsets the negative returns to interpersonal skills for men: in contrast to their non-educated counterparts, men with 14 years of education get null returns from these skills. Meanwhile, men with 14 years of education still get positive returns to intrapersonal skills. These results suggest the possibility that non-educated men may be more likely to operate in sectors where intrapersonal skills have positive returns and interpersonal skills have negative ones. Education significantly increases women’s returns from intrapersonal skills, but not for interpersonal skills. Thus, while education increases men’s returns to interpersonal skills, it increases women’s returns to intrapersonal skills. Nonetheless, women with 14 years of education still have positive returns to interpersonal skills. Altogether, education is more transformative for women, and women with the highest education levels receive positive returns to both intrapersonal and interpersonal skills. Looking at specific skills by education level, we find that gender stereotypical skills are more rewarded for non-educated men and women and this pattern lessens with education. In particular, non-educated men and women both get negative returns to emotional regulation and positive returns to self-control. However, non-educated men get positive returns from PSC, while non-educated women get positive returns to perseverance and teamwork and get significantly more returns than men to empathy, expressiveness, and interpersonal relatedness (even though the overall returns for these latter skills are not significant). Moreover, while non-educated men get negative returns from expressiveness and interpersonal relatedness, women do not. Education increases returns to emotional regulation for both men and women, while it decreases returns to self-control. Furthermore, education lowers men’s returns to PSC and increases men’s returns to expressiveness and interpersonal relatedness, effectively reversing the negative returns for these interpersonal skills for men. Both men and women with 14 years of education get positive returns from emotional regulation, perseverance, PSDM and expressiveness. In 23 addition, women with 14 years of education get positive returns from teamwork, positive self-concept, and personal initiative. Heterogeneity across types of employment Finally, we turn to analyzing how these results vary based on employment type, and especially based on whether the respondent is self-employed in the non-agricultural sector or wage employed. Results are reported in Table 7. One caveat to drawing direct comparisons here is that the two subsamples do not derive from the same population. We however observe the following patterns. Women have significantly lower levels of earnings than men in all sectors. While socio-emotional skills are significantly correlated with earnings in non-agricultural self-employment and in wage employment for men, as in our main results, there is no significant difference by gender (Panel A). Socio-emotional skills are not significantly correlated with earnings in wage employment for women. In both sectors, intrapersonal skills appear to have higher returns than interpersonal skills for both men and women. We move next to examining skill specific differences within the non-agricultural self-employment and wage employment sectors. In the non-agricultural self-employment sector (Panel B), positive self-concept, PSDM, and self-control are associated with higher earnings for men. While women yield insignificant returns from positive self-concept and self-control, women’s earnings are positively associated with PSDM, perseverance, personal initiative, and teamwork. Panel C shows correlations for wage workers. Interpersonal relatedness and teamwork are negatively associated with earnings for men, whereas these negative associations are offset for women. Altogether, these results suggest that socio-emotional skills may be especially valuable for non-agricultural self- employment relative to the wage employment sector. 5. Robustness Checks The previous sections have documented gender differences in socio-emotional skills and their associations with economic outcomes. This section reinforces our interpretation of our results, by showing robustness to i) alternative measurement of the gender gap, ii) alternative measurement of skills, iii) differences in self-assessments of skills, iv) controls for school transitions, v) heterogeneity across studies. 24 5.1 Alternative measurement of the gender gap To contrast our results with alternative approaches to quantify differences between subpopulations, we present the standardized difference between men and women’s socio-emotional skills means in Appendix Table A5 in the form of Cohen’s d, reported for the full sample and disaggregated by study as well as by age and education categories. In a large majority of studies, we find higher levels of all socio-emotional skills among men, except for self-control, which is consistent with our main results. A comparison between Table 8, which describes coefficients across studies, and Appendix Table A5 demonstrates that this alternative methodology occasionally yields different results. Some large effect sizes based on regression results fall below the minimal threshold of 0.2 from Cohen’s d results (Hyde, 2005). Similarly, some larger values for Cohen’s d do not result in significant or large effect sizes after controlling for age and education. Nonetheless, the overall pattern of a gender gap predominantly in favor of men remains the same. 5.2 Alternative measurement of skills In our baseline results, we construct skill measures by combining study variables assigned to a given skill. When initial studies include few variables for a given skill, the resulting skill measures combine a low number of contributing items. To address the concern that measures based on less than three items might not provide a robust assessment of the underlying skills (Marsh et al., 1998), we restrict our analysis to socio-emotional skills captured by at least three items in an individual survey. Our sample size reduces to 40,761 observations for the gender differences in the aggregate index. The resulting estimates for gender differences in socio-emotional skills remain largely unchanged (Appendix Table A6, columns 1 and 2). We only observe a change in significance for empathy, for which we keep less than 14% of observations in this robustness check. Now turning to heterogeneity in socio-emotional skills by education in Appendix Table A7, results are also mainly unchanged. The main difference is that we find higher levels of interpersonal skills for women among the non-educated population, notably for expressiveness (similar to Table 4) and interpersonal relatedness. Lastly, we estimate correlations between socio-emotional skills and earnings based on this restricted sample in Appendix Table A8. We find a robust, even stronger, correlation of aggregate and intrapersonal socio-emotional skills with earnings with no gender differences in these correlations (Panel A). As in our main specification, we find that the association between interpersonal socio-emotional skills and earnings is significantly higher for women than for men, for whom the negative correlation with earnings is here significant. Results on the correlation between disaggregated skills and earnings are very similar to those described in subsection 4.3. 25 5.3 Controlling for positive self-concept Higher levels of positive self-concept may bias self-reported measures of socio-emotional skills by creating a gap between self-assessments of skills and actual skill levels. Moreover, this overestimation may be tied to gender norms (Correll, 2004). We indeed find a 0.060 standard deviation difference in positive self- concept, favorable to men (Table 3). On the other hand, positive self-concept may be foundational for the formation of other socio-emotional skills, and some correlation between skills may be expected. Still, we examine whether our previous findings are retained, limited to a sub-sample of 25,551 for which a measure of positive self-concept is available. We find that self-concept has a significant positive association with all other socio-emotional skills for individuals in this sub-sample (Appendix Table A9, column 4). Controlling for self-concept attenuates the gender gap in socio-emotional skills, from -0.151 to -0.115 (Appendix Table A9, columns 1 and 3). The gender gap in intrapersonal and interpersonal skills both fall slightly but remain statistically significant, suggesting that the difference in skills we observe does not fully stem from differences in self-concept but partly reflects an underlying difference in skills. For interpersonal skills, the only robust gender difference is found for teamwork, while for other individual skills, the gender difference is not significant. The fact that coefficients change signals the importance of considering gender differential biases in self-reported measures of socio-emotional skills. The gender-specific association between socio-emotional skills and education is widely unchanged once we control for self-concept (Appendix Table A9). We then estimate the correlation between socio- emotional skills and economic outcomes controlling for positive self-concept (Appendix Table A10). Using two-stage residual inclusion estimation, we estimate residuals for each skill after controlling for positive self-concept. We then run our usual specification (Equation 3) using these residuals instead of our initial index. Results on the correlation with earnings are consistent with our main estimation. A one standard deviation change in the aggregate socio-emotional skills measure is associated with a 6.4% increase in earnings for women and a 4.9% increase for men, compared to 8.8% and 6.8% in the main specification. We also find that intrapersonal skills continue to be more strongly associated with earnings than interpersonal skills for men, and that interpersonal skills are more strongly associated with economic outcomes for women than for men. 5.4 Controlling for school transitions To better understand whether the relationships we observe between skills and gender are rather driven by people sorting on socio-emotional skills to enter and remain in formal education, or by people building 26 socio-emotional skills in school, we run the same analysis as in Table 4 and estimate model (3) but now augmented with controls for transition years in educational attainment (in the form of dummies for ever entering lower secondary, senior secondary or higher education). In Appendix Table A11, we make the assumption that indicator variables for transition years account for any sorting mechanism which might be at play, and that our coefficients on years of education thus provide a conservative estimate of socio- emotional skills building in formal education. Our main results largely remain. Women’s returns to education overall do not substantially change once we control for this sorting mechanism. Men get lower returns than before from interpersonal skills, but they are still positive and statistically significant. Overall, men with higher socio-emotional skills are less likely to enter lower secondary but more likely to transition into senior secondary. We do not find any significant correlation between women’s level of socio-emotional skills and their transition into higher levels of education, which suggests that there is less sorting into education based on socio-emotional skills for women. 5.5 Heterogeneity across studies To uncover differential correlations between socio-emotional skills and economic outcomes across samples, we report estimates of correlations with earnings disaggregated at the study-level in Table 8. We find significantly lower aggregate levels of socio-emotional skills among women in all but three of the thirteen studies (Panel A). The three exceptions (Cote d’Ivoire PSAC, Facebook, and Togo PI) include samples restricted to working individuals – as farmers, business entrepreneurs, or employees, which may explain the statistically insignificant differences in socio-emotional skills levels among men and women in the selected samples. Altogether, gender differences in socio-emotional skills levels disfavor women across all studies and skills except for self-control, of which women display significantly higher levels in both the Mozambique and Togo PI studies (selecting farmers and entrepreneurs), and except for emotional regulation and PSDM being higher for women in the Facebook study. Turning to correlations with earnings (Panel B), we generally find insignificant gender differences except for two cases in which women display a significantly higher correlation between aggregate socio-emotional skills and earnings than men (in the Ghana GADCO and Kenya STEP studies). The overall female advantage in the correlation between interpersonal skills with earnings is driven by the Cote d’Ivoire Pro-Jeunes and Kenya STEP studies (both urban samples), along with positive although statistically insignificant gender gaps favoring women in most of the other studies. Most studies indicate high and statistically significant correlations of all types of socio-emotional skills with earnings for women. Taken together, these results indicate 27 consistent patterns of lower socio-emotional skills levels for women but positive associations with earnings across studies, and with interpersonal skills being more strongly correlated with earnings for women than for men. 6. Conclusion This paper investigates gender differences in a wide range of socio-emotional skills, at different education levels, as well as how these skills correlate with earnings and whether this correlation differs for men and women. We exploit a rich and unique combination of datasets encompassing 41,873 individuals across 17 Sub-Saharan African countries. We use standardized measures of socio-emotional skills covering six intrapersonal and four interpersonal skills by aggregating self-reported items across studies. While results may not be applicable to all settings, this study is the first to examine gender differences in a large variety of socio-emotional skills, separating out intra- and interpersonal skills, and for an extensive number of African countries. We find a significant gender gap disfavoring women by about 0.151 standard deviations, across most socio-emotional skills, equivalent to the gap induced by 5.6 years of education. The largest gap is observed in problem-solving and decision-making while the gap in empathy is the smallest. Self-control is the only skill for which women do not exhibit significantly lower levels than men. We hypothesize that gender norms play an important role in explaining gender differences in socio- emotional skills. More work is needed to better understand and estimate this role. The gender gap in socio-emotional skills is only partially explained by lower education levels among women. Closing the gender gap in education would close about 17% of the socio-emotional skills gender gap. Overall, women and men experience similar returns to education in terms of aggregate socio- emotional skills levels. However, we observe differences at a more disaggregated level. Indeed, the gender gap in interpersonal skills, emotional regulation and self-control is wider at higher education levels. In general, these results are suggestive rather than definitive, as two important considerations arise in the measurement of each skill. First, self-reported measures may reflect gender-specific biases in assessments of skills. In complementary work, we are developing observation-based measures of skills, that will help address this issue. Second, this study is based on available data, and measures were not validated for this division of skills. In addition, larger efforts were spent on measuring intrapersonal skills (75% of socio- emotional skills items) than interpersonal skills across studies. Thus, measures of interpersonal skills suffer 28 from relying on a lower number of observations and from lower reliability. This calls for further data collection and analysis on the relative importance of interpersonal skills. Another central contribution of this paper is in highlighting which skills correlate most with economic empowerment. While we lack a source of exogenous variation in socio-emotional skills acquisition and thus cannot infer causality, we do not find evidence of differential associations of our aggregate socio- emotional skills measures with economic outcomes across gender. Socio-emotional skills are robustly associated with higher earnings for men and women. However, digging deeper into separate skills, we find evidence supporting the hypothesis that the specific skills associated with the highest earnings differ for men and women. Notably, while women seem to gain less interpersonal skills than men from education, we find interpersonal skills to be more strongly correlated with earnings for women than for men. More research is needed to understand why women get higher returns from interpersonal skills, and especially teamwork. Given the positive correlation we find between socio-emotional skills - especially interpersonal skills for women - and economic outcomes, taken together with the fact that higher levels of education are associated with higher gender gaps in interpersonal skills, public interventions aiming to equip women with these highly rewarded skills may provide an effective pathway to reduce gender disparities in the labor market. In terms of differences by educational attainment, we find that non-educated women have null returns to either type of socio-emotional skills, which disaggregates into higher returns to interpersonal skills than men and lower returns to intrapersonal skills, which they lack most compared to men. Interestingly, while education increases men’s returns to expressiveness and interpersonal relatedness as well as the gender gap in these skills levels, it increases women’s returns to positive self-concept and teamwork more than men’s. Finally, we examine differences by occupational sector and find that socio-emotional skills are significantly correlated with earnings in non-agricultural self-employment for both men and women, as in our main results, but are not significantly correlated with earnings in wage employment for women. In both sectors, intrapersonal skills appear to have higher returns than interpersonal skills for men. Contrastingly, women in wage employment only get significant returns from interpersonal skills, not from intrapersonal skills. Perseverance and personal initiative are associated with higher earnings for women in non-agricultural self-employment. For wage workers, interpersonal relatedness and teamwork are negatively associated with earnings for men, whereas these negative associations are offset for women. 29 Overall, our results suggest that programs aiming at reducing the gender gap in earnings would benefit from including socio-emotional skills building components specifically targeting interpersonal skills such as teamwork, expressiveness and interpersonal relatedness. 30 References Acosta P., Muller, N., Sarzosa, M. (2020). Adult’s Cognitive and Socioemotional Skills and Their Labor Market Outcomes in Colombia. Revista de economía del Rosario 23(1), 109-148. Ali, D. A., Bowen, D., & Deininger, K. (2020). Personality traits, technology adoption, and technical efficiency: Evidence from smallholder rice farms in Ghana. Journal of Development Studies, 56(7), 1330- 1348. Alloy, L. B., Abramson, L. Y., Hogan, M. E., Whitehouse, W. G., Rose, D. T., Robinson, M. S., ... & Lapkin, J. B. (2000). The Temple-Wisconsin Cognitive Vulnerability to Depression Project: lifetime history of axis I psychopathology in individuals at high and low cognitive risk for depression. Journal of abnormal psychology, 109(3), 403. Almlund, M., Duckworth, A., Heckman, J. and Kautz, T. (2011). Personality Psychology and Economics. National Bureau of Economic Research (NBER) Working Paper, No. 16822. Amanatullah, E. T., and Morris, M. W. (2010). Negotiating gender roles: Gender differences in assertive negotiating are mediated by women’s fear of backlash and attenuated when negotiating on behalf of others. Journal of personality and social psychology, 98(2): 256. Anthony, D., and Horne, C. (2003). Gender and Cooperation: Explaining Loan Repayment in xzzzdcz Micro-Credit Groups. Social Psychology Quarterly, 66(3): 293-302. Arbache, J. S., Kolev A. and Filipiak E., eds. (2010). Gender disparities in Africa’s labor market. Washington, DC: World Bank. Ashraf, N., Bohnet, I. and Piankov, N. (2006). Decomposing trust and trustworthiness. Exp Econ 9, 193– 208. https://doi.org/10.1007/s10683-006-9122-4 Balliet, Daniel, LI, Norman P., Macfarlan, Shane J., & Van Vugt, Mark. (2011). Sex Differences in Cooperation: A Meta-analytic Review of Social Dilemmas. Psychological Bulletin, 137(6): 881-909. https://ink.library.smu.edu.sg/soss_research/1130 Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K. D. (2003). Does high self-esteem cause better performance, interpersonal success, happiness, or healthier lifestyles? Psychological science in the public interest, 4(1), 1-44. Bazelais, P., Lemay, D., and Doleck, T. (2016). How Does Grit Impact College Students' Academic Achievement in Science? European Journal of Science and Mathematics Education, 4(1): 33-43. 2016 Bear, J. B., & Woolley, A. W. (2011). The role of gender in team collaboration and performance. Interdisciplinary science reviews, 36(2), 146-153. Bernard, T., Dercon, S., Orkin, K., & Taffesse, A. (2014). The future in mind: Aspirations and forward- looking behaviour in rural Ethiopia (p. 10224). London: Centre for Economic Policy Research. 31 Blazar, D., & Kraft, M. A. (2017). Teacher and teaching effects on students’ attitudes and behaviors. Educational evaluation and policy analysis, 39(1), 146-170. Bleidorn, W., Arslan, R. C., Denissen, J. J., Rentfrow, P. J., Gebauer, J. E., Potter, J., & Gosling, S. D. (2016). Age and gender differences in self-esteem—A cross-cultural window. Journal of Personality and Social Psychology, 111(3), 396. Botea, I. Donald A. & Rouanet L. (2021) In it to win it? Self-esteem and income-earning among couples. Journal of Economic Behavior and Organization 187 (2021) 488–506 Bowles, H. R., and Flynn, F. (2010). Gender and persistence in negotiation: A dyadic perspective. Academy of Management Journal, 53(4): 769-787. Bowles, S., Gintis, H., & Osborne, M. (2001). The determinants of earnings: A behavioral approach. Journal of economic literature, 39(4), 1137-1176. Buchan, N. R., Croson, R. T. A., and Solnick, S. (2008). Trust and gender: An examination of behavior and beliefs in the investment game. Journal of Economic Behavior & Organization, 68(3-4): 466– 476. https://doi.org/10.1016/j.jebo.2007.10.006 Campos, F., Frese, M., Goldstein, M., Iacovone, L., Johnson, H. C., McKenzie, D., and Mensmann, M. (2017). Teaching personal initiative beats traditional training in boosting small business in West Africa. Science, 357(6357), 1287-1290. Carneiro, P., Crawford, C., & Goodman, A. (2007). The impact of early cognitive and non-cognitive skills on later outcomes. Chakravarty, S., Das, S., and Vaillant, J. (2017). Gender and Youth Employment in Sub-Saharan Africa: A Review of Constraints and Effective Interventions. World Bank Policy Research Working Paper, No. WPS 8245. Charness, G. and Rustichini, A. (2011). Gender differences in cooperation with group membership. Games and Economic Behavior, 72(1): 77–85 Chioda, L., Contreras-Loya, D., Gertler, P., & Carney, D. (2021). Making Entrepreneurs: Returns to Training Youth in Hard Versus Soft Business Skills (No. w28845). NBER Working Paper Series 28845 Christenson, R. and Knezek, G. (2014). Comparative Measures of Grit, Tenacity and Perseverance. International Journal of Learning, Teaching and Educational Research 8(1): 16-30. Cobb-Clark, D. A., & Tan, M. (2011). Noncognitive skills, occupational attainment, and relative wages. Labour Economics, 18(1), 1-13. Correll, S. (2001). Gender and the Career Choice Process: The Role of Biased Self‐Assessments. American Journal of Sociology, 106(6): 1691-1730. 32 Correll, S. (2004). Constraints into Preferences: Gender, Status, and Emerging Career Aspirations. American Sociological Review, 69(1): 93-113. Conti, G., Heckman, J., & Urzua, S. (2010). The education-health gradient. American Economic Review, 100(2), 234-38. Costa, P., Terracciano, A., and McCrae, R. (2001). Gender Differences in Personality Traits Across Cultures: Robust and Surprising Findings. Journal of Personality and Social Psychology, 81(2): 322-331. Credé, M., Tynan, M. C., and Harms, P. D. (2017). Much ado about grit: A meta-analytic synthesis of the grit literature. Journal of Personality and Social Psychology, 113(3), 492– 511. https://doi.org/10.1037/pspp0000102axzxazxxzx Croson, R. and Buchan, N. (1999). Gender and Culture: International Experimental Evidence from Trust Games. American Economic Review, 89 (2): 386-391. Croson, R., and Gneezy, U. (2009). Gender Differences in Preferences. Journal of Economic Literature, 47(2): 448–474. Cross, C. P., Copping, L. T., and Campbell, A. (2011). Sex differences in impulsivity: A meta- analysis. Psychological Bulletin, 137(1), 97–130. https://doi.org/10.1037/a0021591 Cunha, F. Heckman, J. and Schennach, S. (2010). Estimating the Technology of Cognitive and Noncognitive Skill Formation. Econometrica, 78(3), 883–931. Cunningham, W., Torrado, M., & Sarzosa, M. (2016). Cognitive and non-cognitive skills for the Peruvian labor market: Addressing measurement error through latent skills estimations. World Bank Policy Research Working Paper, (7550). Cunningham, W., and Villasenor, P. (2016). Employer Voices, Employer Demands, and Implications for Public Skills Development Policy Connecting the Labor and Education Sectors. The World Bank Research Observer, 31(1): 102–134. Das, S., & Kotikula, A. (2019). Gender-based employment segregation: Understanding causes and policy interventions. World Bank. DiPrete, T. A., & Jennings, J. L. (2012). Social and behavioral skills and the gender gap in early educational achievement. Social Science Research, 41(1), 1-15. Duckworth, A. L., Peterson, C., Matthews, M. D., & Kelly, D. R. (2007). Grit: perseverance and passion for long-term goals. Journal of personality and social psychology, 92(6), 1087. Duckworth, A. L., & Seligman, M. E. (2005). Self-discipline outdoes IQ in predicting academic performance of adolescents. Psychological science, 16(12), 939-944. 33 Duckworth, A. L., and Seligman, M. E. P. (2006). Self-discipline gives girls the edge: Gender in self- discipline, grades, and achievement test scores. Journal of Educational Psychology, 98(1):198– 208. https://doi.org/10.1037/0022-0663.98.1.198 Duckworth, A. L., and Seligman, M. E. P. (2006). Self-discipline gives girls the edge: Gender in self- discipline, grades, and achievement test scores. Journal of Educational Psychology, 98(1):198– 208. https://doi.org/10.1037/0022-0663.98.1.198 Duckworth, A. and Yeager, D. (2015). Measurement Matters: Assessing Personal Qualities Other Than Cognitive Ability for Educational Purposes. Educational Researcher, 44(4): 237–251. Duflo, E., Dupas, P., & Kremer, M. (2011). Peer effects, teacher incentives, and the impact of tracking: Evidence from a randomized evaluation in Kenya. American Economic Review, 101(5), 1739-74. Durlak, J. A., Weissberg, R. P., Dymnicki, A. B., Taylor, R. D., and Schellinger, K. B. (2011). The impact of enhancing students’ social and emotional learning: A meta‐analysis of school‐based universal interventions. Child development, 82(1): 405-432. D'Zurilla, T. J., Maydeu-Olivares, A., & Kant, G. L. (1998). Age and gender differences in social problem- solving ability. Personality and individual differences, 25(2), 241-252. Eagly, A. and Wood, W. (2012). Social role theory. Handbook of theories in social psychology. 2. 458-476. Erol, R. Y., & Orth, U. (2011). Self-esteem development from age 14 to 30 years: a longitudinal study. Journal of personality and social psychology, 101(3), 607. Exley, C. L., Niederle, M., & Vesterlund, L. (2020). Knowing when to ask: The cost of leaning in. Journal of Political Economy, 128(3), 816-854. Farrington, C. A., Roderick, M., Allensworth, E., Nagaoka, J., Keyes, T. S., Johnson, D. W., & Beechum, N. O. (2012). Teaching Adolescents to Become Learners: The Role of Noncognitive Factors in Shaping School Performance--A Critical Literature Review. Consortium on Chicago School Research. 1313 East 60th Street, Chicago, IL 60637. Feingold, A. (1994). Gender Differences in Personality: A Meta-Analysis. Psychological Bulletin, 116(3), 429-456. Forret, M.L., Dougherty, T.W. (2001). Correlates of Networking Behavior for Managerial and Professional Employees. Group & Organization Management, 26(3):283-311. Friedson-Ridenour, S., and Pierotti, R. (2019). Competing priorities: Women’s microenterprises and household relationships. World Development, 121: 53–62. Gentile, B., Grabe, S., Dolan-Pascoe, B., Twenge, J. M., Wells, B. E., & Maitino, A. (2009). Gender differences in domain-specific self-esteem: A meta-analysis. Review of General Psychology, 13(1), 34-45. 34 Gibson, C., Ward, J., Wright, J., Beaver, K., Delisi, M. (2010). Where Does Gender Fit in the Measurement of Self-Control? Criminal Justice and Behavior, 37(8): 883-903. Gielnik, M., Frese, M., Kahara-Kawuki, A., Katono, I., Kyejjusa, S., Ngoma, M., Munene, J., Namatovu- Dawa, R., Nansubuga, F., Orobia, L., Oyugi, J., Sejjaaka, S., Sserwanga, A., Walter, T., Bischoff, K., and Dlugosch, T. (2015). Action and Action-Regulation in Entrepreneurship: Evaluating a Student Training for Promoting Entrepreneurship. Academy of Management Learning & Education, 14(1): 69–94. Groh, M., Krishnan, N., McKenzie, D. and Vishwanath T. (2016). The impact of soft skills training on female youth employment: evidence from a randomized experiment in Jordan. IZA Journal of Labor Development 5(9). https://doi.org/10.1186/s40175-016-0055-9. Gross, J. J., and John, O. P. (2003). Individual differences in two emotion regulation processes: Implications for affect, relationships, and well-being. Journal of Personality and Social Psychology, 85(2): 348– 362. https://doi.org/10.1037/0022-3514.85.2.348 Guerra, N., Modecki, K., & Cunningham, W. (2014). Developing social-emotional skills for the labor market: The PRACTICE model. World Bank Policy Research Working Paper, (7123). Gunewardena, D., King, E. M., & Valerio, A. (2018). More than schooling: Understanding gender differences in the labor market when measures of skill are available. World Bank Policy Research Working Paper, (8588). Hahn, V. C., Frese, M., Binnewies, C., & Schmitt, A. (2012). Happy and proactive? The role of hedonic and eudaimonic well–being in business owners’ personal initiative. Entrepreneurship theory and practice, 36(1), 97-114. Hankin, B. L., & Abramson, L. Y. (2001). Development of gender differences in depression: An elaborated cognitive vulnerability–transactional stress theory. Psychological bulletin, 127(6), 773. Heckman, J., and Kautz, T. (2013). Fostering and Measuring Skills: Interventions that Improve Character and Cognition. National Bureau of Economic Research (NBER) Working Paper, No. 19656. Heckman, J., Stixrud, J., and Urzua, S. (2006). The Effects of Cognitive and Noncognitive Abilities on Labor Market Outcomes and Social Behavior. National Bureau of Economic Research (NBER) Working Paper, No. 12006. Hyde, J. 2013. Gender Similarities and Differences. Annual Review of Psychology, 65(3): 1–3.26. Hyde, J. S., Mezulis, A. H., & Abramson, L. Y. (2008). The ABCs of depression: integrating affective, biological, and cognitive models to explain the emergence of the gender difference in depression. Psychological review, 115(2), 291. Hyde, J. (2005). The gender similarities hypothesis. American psychologist, 60(6), 581. Jackson, C. K., Porter S. C., Easton J. Q., Blanchard A. & Kiguel S. (2020). School Effects on Socioemotional Development, School-Based Arrests, and Educational Attainment. American Economic Review: Insights, 2 (4): 491-508. 35 Jones, S. M., Bouffard, S. M., & Weissbourd, R. (2013). Educators' social and emotional skills vital to learning. Phi Delta Kappan, 94(8), 62-65. Judge, T. A., & Ilies, R. (2002). Relationship of personality to performance motivation: A meta-analytic review. Journal of Applied Psychology, 87(4), 797. Kanagaretnam, K., Mestelman, S., Nainar, K., & Shehata, M. (2009). The impact of social value orientation and risk attitudes on trust and reciprocity. Journal of Economic Psychology, 30(3), 368-380. Kautz, T., Heckman, J., Diris,R., Weel, B., and Borghans,L. (2014). Fostering and Measuring Skills: Improving Cognitive and Non-Cognitive Skills to Promote Lifetime Success. National Bureau of Economic Research (NBER) Working Paper, No. 20749. Kenney-Benson, G. A., Pomerantz, E. M., Ryan, A. M., and Patrick, H. (2006). Sex differences in math performance: The role of children's approach to schoolwork. Developmental Psychology, 42(1): 11– 26. https://doi.org/10.1037/0012-1649.42.1.11 Kling, K. C., Hyde, J. S., Showers, C. J., & Buswell, B. N. (1999). Gender differences in self-esteem: a meta- analysis. Psychological bulletin, 125(4), 470. Koop, S., De Reu, T., & Frese, M. (2000). Entrepreneurial orientation and personal initiative. Success and failure of microbusiness owners in Africa: A psychological approach, 55. Kraft, M. A., & Grace, S. (2015). Teaching for tomorrow’s economy? Elementary teacher effects on complex tasks, grit, and growth mindset. Working Paper. Kraft, M. A., & Grace, S. (2016). Teaching for tomorrow’s economy? Teacher effects on complex cognitive skills and social-emotional competencies. NBER Working Paper. Krauss, S. I., Frese, M., Friedrich, C., & Unger, J. M. (2005). Entrepreneurial orientation: A psychological model of success among southern African small business owners. European Journal of work and organizational psychology, 14(3): 315-344. Leaper, C., and Ayres, M.M. (2007). A Meta-Analytic Review of Gender Variations in Adults’ Language Use: Talkativeness, Affiliative Speech, and Assertive Speech. Personality and Social Psychology Review, 11(4):328-363. Lindqvist, E. and Vestman, R. (2009). The Labor Market Returns to Cognitive and Noncognitive Ability: Evidence from the Swedish Enlistment. IFN Working Paper No. 794. https://ssrn.com/abstract=1367603 or http://dx.doi.org/10.2139/ssrn.1367603 Lippman, L., Ryberg, R., Carney, R., and Moore, K. (2015). Workforce Connections: Key “Soft Skills” that Foster Youth Workforce Success: Toward a Consensus Across Fields. Child Trends. Loeb, S., Christian, M. S., Hough, H. J., Meyer, R. H., Rice, A. B., & West, M. R. (2018). School Effects on Social-Emotional Learning: Findings from the First Large-Scale Panel Survey of Students. Working Paper. Policy Analysis for California Education, PACE. 36 Lynn, R., and Martin, T. (1997). Gender Differences in Extraversion, Neuroticism, and Psychoticism in 37 Nations. The Journal of Social Psychology, 137(3): 369-373. Mammadov, S. (2021). The Big Five Personality Traits and Academic Performance: A Meta- Analysis. Journal of Personality. Marsh, H. W., Hau, K. T., Balla, J. R., and Grayson, D. (1998). Is more ever too much? The number of indicators per factor in confirmatory factor analysis. Multivariate behavioral research, 33(2): 181-220. McClure, E. B. (2000). A meta-analytic review of sex differences in facial expression processing and their development in infants, children, and adolescents. Psychological Bulletin, 126(3): 424– 453. https://doi.org/10.1037/0033-2909.126.3.424 McRae, K, Ochsner, K., Mauss, I., Gabrieli, J., Gross, J. (2008). Gender Differences in Emotion Regulation: An fMRI Study of Cognitive Reappraisal. Group Processes & Intergroup Relations, 11(2):143-162. Mengel, F., (2020). Gender differences in networking. The Economic Journal, 130(630): 1842– 1873. https://doi.org/10.1093/ej/ueaa035 Mestre, V., Samper, P., Frías-navarro, D., and Tur-Porcar, A. (2009). Are Women More Empathetic than Men? A Longitudinal Study in Adolescence. The Spanish journal of psychology, 12(1): 76-83. 12. Miller, A., Worthington, Jr., E., and McDaniel, M. (2008). Gender and Forgiveness: A Meta–Analytic Review and Research Agenda. Journal of Social and Clinical Psychology, 27(8): 843-876. Mohanty, M. S. (2009). Effects of positive attitude on happiness and wage: Evidence from the US data. Journal of Economic Psychology, 30(6), 884-897. Montalvao, J., Frese, M., Goldstein, M., and Kilic, T. (2017). Soft Skills for Hard Constraints: Evidence from High-Achieving Female Farmers. World Bank Policy Research Working Paper, No. WPS 8095. Mueller, G. and Plug, E. (2006). Estimating the Effect of Personality on Male-Female Earnings. Industrial and Labor Relations (ILR) Review, 60(1): 3-22. Mulac, A., Bradac, J. J., & Gibbons, P. (2001). Empirical support for the gender-as-culture hypothesis: An intercultural analysis of male/female language differences. Human Communication Research, 27(1), 121- 152. Nakhaie, M. R., Silverman, R. A., & LaGrange, T. C. (2000). Self-control and social control: An examination of gender, ethnicity, class and delinquency. Canadian Journal of Sociology/Cahiers canadiens de sociologie, 35-59. Newman, A., Obschonka, M., Schwarz, S., Cohen, M., & Nielsen, I. (2019). Entrepreneurial self-efficacy: A systematic review of the literature on its theoretical foundations, measurement, antecedents, and outcomes, and an agenda for future research. Journal of Vocational Behavior, 110, 403-419. 37 Newman, M. L., Groom, C. J., Handelman, L. D., & Pennebaker, J. W. (2008). Gender differences in language use: An analysis of 14,000 text samples. Discourse Processes, 45(3), 211-236. Nolen-Hoeksema, S. (2001). Ruminative coping and adjustment to bereavement. In M. S. Stroebe, R. O. Hansson, W. Stroebe, & H. Schut (Eds.), Handbook of bereavement research: Consequences, coping, and care (pp. 545–562). American Psychological Association. https://doi.org/10.1037/10436-023 Nolen-Hoeksema, S. (2012). Emotion Regulation and Psychopathology: The Role of Gender. Annual Review of Clinical Psychology, 8(1): 161-187. Nyhus, E. K., & Pons, E. (2005). The effects of personality on earnings. Journal of economic psychology, 26(3), 363-384. Nyhus, E. K., & Pons, E. (2012). Personality and the gender wage gap. Applied Economics, 44(1), 105-118. O’Brien Ed., Konrath S.H., Gruhn D., and Hagen A.L. (2013). Empathic concern and perspective taking: linear and quadratic effects of age across the adult life span. The Journals of Gerontology: Series B, 68(2), 168–175. https://doi.org/10.1093/geronb/gbs055 Organisation for Economic Co-operation and Development. (2021). Beyond Academic Learning: First results from the survey of social and emotional skills. OECD Publishing. Organisation for Economic Co-operation and Development. (2015). Skills for Social Progress: The Power of Social and Emotional Skills. OECD Skill Studies. Onyeizugbo, E. U. (2003). Effects of gender, age, and education on assertiveness in a Nigerian sample. Psychology of women quarterly, 27(1): 12-16. Organisation for Economic Co-operation and Development. (2015). Skills for social progress: The power of social and emotional skills. OECD Publishing. Paola, M., Gioia, F., and Scoppa, V. (2018). Teamwork, Leadership and Gender. IZA Discussion Papers, No. 11861. Park, G., Yaden, D.B., Schwartz, H.A., Kern, M.L., Eichstaedt, J.C., Kosinski, M., et al. (2016). Women are Warmer but No Less Assertive than Men: Gender and Language on Facebook. PLoS ONE, 11(5): e0155885. https://doi.org/10.1371/journal.pone.0155885 Papageorge, N. W., Ronda, V., & Zheng, Y. (2019). The economic value of breaking bad: Misbehavior, schooling and the labor market (No. w25602). NBER Working Paper Series 25602. Roberts, B., Kuncel, N., Shiner, R., Caspi, A., and Goldberg, L. (2007). The Power of Personality: The Comparative Validity of Personality Traits, Socioeconomic Status, and Cognitive Ability for Predicting Important Life Outcomes. Perspex Psychol Sci.; 2(4): 313–345. Robins, R. W., Trzesniewski, K. H., Tracy, J. L., Gosling, S. D., & Potter, J. (2002). Global self-esteem across the life span. Psychology and aging, 17(3), 423. 38 Robitschek, C., & Cook, S. W. (1999). The influence of personal growth initiative and coping styles on career exploration and vocational identity. Journal of Vocational Behavior, 54(1), 127-141. Runyan, R. C., Huddleston, P., & Swinney, J. (2006). Entrepreneurial orientation and social capital as small firm strategies: A study of gender differences from a resource-based view. The International Entrepreneurship and Management Journal, 2(4), 455-477. Sanchez Puerta, M. L., Valerio, A., & Bernal, M. G. (2016). Taking stock of programs to develop socioemotional skills: A systematic review of program evidence. The World Bank. Schilbach, F., Schofield, H., and Mullainathan, S. (2016). The Psychological Lives of the Poor. American Economic Review, 106(5): 435-440. Schmitt, D., Realo, A., Voracek, M., and Allik, J. (2008). Why Can’t a Man Be More Like a Woman? Sex Differences in Big Five Personality Traits Across 55 Cultures. Journal of Personality and Social Psychology, 94(1): 168–182. Shulman, E. P., Harden, K. P., Chein, J. M., & Steinberg, L. (2015). Sex differences in the developmental trajectories of impulse control and sensation-seeking from early adolescence to early adulthood. Journal of youth and adolescence, 44(1), 1-17. Silverman, I. W. (2003). Gender differences in delay of gratification: A meta-analysis. Sex roles, 49(9-10), 451-463. Skaalvik, S., & Skaalvik, E. M. (2004). Gender differences in math and verbal self-concept, performance expectations, and motivation. Sex roles, 50(3), 241-252. Smithers, L. G., Sawyer, A. C., Chittleborough, C. R., Davies, N. M., Smith, G. D., & Lynch, J. W. (2018). A systematic review and meta-analysis of effects of early life non-cognitive skills on academic, psychosocial, cognitive and health outcomes. Nature human behaviour, 2(11), 867-880. Steinberg, L., Graham, S., O’brien, L., Woolard, J., Cauffman, E., & Banich, M. (2009). Age differences in future orientation and delay discounting. Child development, 80(1), 28-44. Stuijfzand, S., De Wied, M., Kempes, M., Van de Graaff, J., Branje, S., and Meeus, W. (2016). Gender differences in empathic sadness towards persons of the same-versus other-sex during adolescence. Sex roles, 75(9-10): 434-446. Sunstein, C. R., & Hastie, R. (2014). Making dumb groups smarter. Harvard business review, 92(12), 90-98. Taylor, R. D., Oberle, E., Durlak, J. A., and Weissberg, R. P. (2017). Promoting positive youth development through school‐based social and emotional learning interventions: A meta‐analysis of follow‐up effects. Child development, 88(4): 1156-1171. Toussaint, L. and Webb, J. (2005). Gender Differences in the Relationship Between Empathy and Forgiveness. The Journal of Social Psychology, 145(6): 673-685. 39 Twenge, J. M., & Campbell, W. K. (2001). Age and birth cohort differences in self-esteem: A cross-temporal meta-analysis. Personality and social psychology review, 5(4), 321-344. Villaseñor, P. (2018). The different ways that teachers can influence the socio-emotional development of their students: A literature review. World Development Report. Vugt, M.V., Cremer, D.D., and Janssen, D.P. (2007). Gender Differences in Cooperation and Competition: The Male-Warrior Hypothesis. Psychological Science, 18(1):19-23. Wagner, H. L., Buck, R., & Winterbotham, M. (1993). Communication of specific emotions: Gender differences in sending accuracy and communication measures. Journal of Nonverbal Behavior, 17(1), 29- 53. Willer, R., Wimer, C., and Owens, L.A. (2015). What drives the gender gap in charitable giving? Lower empathy leads men to give less to poverty relief. Social Science Research. http://dx.doi.org/10.1016/j.ssresearch.2014.12.014 Wilson, F., Kickul, J., & Marlino, D. (2007). Gender, entrepreneurial self–efficacy, and entrepreneurial career intentions: Implications for entrepreneurship education. Entrepreneurship theory and practice, 31(3), 387-406. Woolley, A., Chabris, C., Pentland, A., Hashmi, N., and Malone, T. (2010). Evidence for a Collective Intelligence Factor in the Performance of Human Groups. Science 330(6004): 686-688. Yeager, D. S., & Walton, G. M. (2011). Social-psychological interventions in education: They’re not magic. Review of educational Research, 81(2), 267-301. 40 Figures and Tables Figure 2: Gender differences in socio-emotional skills Note: Each skill measure is displayed on the y-axis. The regression coefficients on Women in Table 3 (Model D) are displayed with their confidence intervals and significance levels. 41 Figure 3: Education heterogeneity for gender differences in socio-emotional skills Note: Each year of education is desplayed on the X-axis. The graph desplays the gender difference on socio-emotional skills at different years of education based on the estimation results of equation (3). 42 Figure 4: Correlations with earnings Note: Each skill measure is displayed on the y-axis. The regression coefficients of Table 5 (“SE skills” for men’s correlations and “SE skills +Women*SE skills” for women’s correlations) are displayed with their confidence intervals and significance levels. Dark markers indicate the correlation between women’s socio-emotional skills and their earnings, while light ones are for men. 43 Figure 5: Education heterogeneity in the correlation between socio-emotional skills and earnings Note: Each year of education is displayed on the X-axis. The graph shows the linear relationship between socio-emotional skills and earning for each year of education based on the estimation results of equation (5). 44 Table 1: Project Demographics Married Monthly Earnings Project Woman Age Education Paid Work /Cohabitating (USD) Full Sample 0.427 (0.495) 35.54 (12.512) 9.646 (4.756) 0.573 (0.495) 0.745 (0.436) 137.756 (435.165) 41873 41873 41873 41873 41478 33965 Women Sample . 33.467 (11.860) 8.477 (5.239) 0.609 (0.488) 0.675 (0.469) 103.251 (368.243) . 17880 17880 17880 17725 16225 Men Sample . 37.084 (12.760) 10.517 (4.154) 0.546 (0.498) 0.797 (0.402) 169.315 (486.280) . 23993 23993 23993 23753 17740 t-test Women-Men 0.000 0.000 0.000 0.000 0.000 0.000 Benin: Youth Employment 0.63 (0.483) 26.231 (4.544) 3.987 (3.685) 0.687 (0.464) 0.766 (0.423) 41.910 (78.152) 2967 2967 2967 2967 2967 2967 Congo: Skills Development Project for 0.332 (0.471) 23.446 (3.116) 10.668 (1.657) 0.121 (0.326) 0.862 (0.345) 50.897 (71.17) Employability 3984 3984 3984 3984 3982 3978 Côte d'Ivoire: Factory Workers 0.490 (0.500) 24.895 (3.610) 10.179 (3.982) 0.138 (0.345) 0.567 (0.496) 73.968 (133.919) 1126 1126 1126 1126 1126 1126 Côte d'Ivoire: ProJeunes 0.716 (0.451) 30.483 (9.184) 5.685 (5.079) 0.365 (0.482) 0.886 (0.318) 65.817 (46.692) 1289 1289 1289 1289 1289 1289 Côte d'Ivoire: Support Project for the 0.185 (0.388) 46.292 (10.555) 6.253 (4.753) 0.992 (0.088) 0.815 (0.388) 280.135 (635.874) Agricultural Sector (PSAC) 1539 1539 1539 1539 1539 1539 Facebook: Future of Business (FoB) 0.184 (0.387) 49.456 (7.765) 12.987 (1.944) 0.527 (0.499) 0.927 (0.26) . 7756 7756 7756 7756 7756 . Ghana: Impact of Outgrower Contracts on Smallholder Farmers (GADCO) 0.355 (0.479) 45.602 (11.395) 9.678 (3.655) 0.814 (0.390) 0.466 (0.499) 108.950 (280.173) 1464 1464 1464 1464 1464 1464 Ghana: Skills Towards Employability and Productivity (STEP)- Skills Measurement 0.482 (0.500) 30.890 (12.016) 10.081 (2.868) 0.411 (0.492) 0.663 (0.473) 504.492 (953.284) 1922 1922 1922 1922 1922 1811 Kenya: Skills Towards Employability and Productivity (STEP)- Skills Measurement 0.523 (0.500) 29.363 (9.664) 9.991 (3.678) 0.523 (0.50) 0.633 (0.482) 474.013 (886.642) 3822 3822 3822 3822 3822 3613 Mozambique: Impact Assessment Integrated Growth Pole Project (IGPP) 0.568 (0.495) 37.118 (14.645) 2.620 (2.894) 0.906 (0.293) 0.430 (0.495) 2.455 (6.019) 5293 5293 5293 5293 5293 5293 Nigeria APPEALS 0.544 (0.498) 33.043 (7.925) 13.278 (1.365) 0.624 (0.485) 0.757 (0.429) 59.949 (115.800) 5918 5918 5918 5918 5918 5918 Togo: Private Sector Development Project (PADSP)- Personal Initiative 0.525 (0.500) 41.260 (9.574) 8.443 (3.986) 0.812 (0.391) 0.978 (0.147) 145.878 (230.080) 1468 1468 1468 1468 1456 1456 Togo: Youth Employment and Skills Development (AIDE) 0.321 (0.467) 31.226 (3.963) 13.190 (1.466) 0.474 (0.499) 0.820 (0.384) 131.421 (180.740) 3325 3325 3325 3325 2944 2944 Note: The table reports the mean, standard deviation and number of observations for the whole sample as well as by project. Woman is a dummy variable equal to 1 if the respondent is a woman, 0 if a man. Age is a continuous variable for the respondent's age. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed high school, 13=completed certificate or diploma and 14=completed university degree or above. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Number of children is a continuous variable indicating the respondent's number of children. Paid Work is a dummy variable equal to 1 if the respondent earns any positive income. Monthly earnings indicate the respondent's monthly earnings in USD. The Facebook data is from the "Future of Business (FoB)" survey. Although it covers 97 countries across the world, this study includes only 15 Sub Saharan Africa (SSA) countries namely: Angola, Benin, Botswana, Côte d'Ivoire, Cameroon, Ethiopia, Ghana, Kenya, Mozambique, Nigeria, Senegal, Tanzania, Uganda, South Africa, and Zambia. http://documents1.worldbank.org/curated/en/734371558715932769/pdf/Tackling-the-Global-Profitarchy-Gender-and-the-Choice-of-Business-Sector.pdf. 45 Table 2: Descriptive Statistics on socio-emotional skills Panel A: Pooled Sample Data Positive Self Emotional Personal Self Interpersonal All Intra Inter Perseverance PSDM Empathy Expressiveness Teamwork Concept Regulation Initiative Control Relatedness All nbr items 323 228 95 39 31 40 48 54 16 14 26 37 18 mean 0.076 0.067 0.067 0.032 0.038 0.056 0.060 0.068 -0.004 0.038 0.053 0.047 0.069 sd 1.022 1.027 1.006 1.005 1.014 1.012 1.028 1.017 1.010 1.013 1.017 1.014 0.980 obs 41873 41834 33658 25551 22573 39885 22052 31959 14835 8260 18866 26941 13115 Women mean 0 0 0 0 0 0 0 0 0 0 0 0 0 sd 1 1 1 1 1 1 1 1 1 1 1 1 1 obs 17880 17870 15535 12129 9638 17495 8923 14726 7519 4192 9451 12470 6611 Men mean 0.133 0.117 0.124 0.062 0.067 0.099 0.100 0.126 -0.009 0.078 0.107 0.088 0.139 sd 1.036 1.045 1.008 1.009 1.024 1.019 1.045 1.029 1.021 1.026 1.032 1.025 0.955 obs 23993 23964 18123 13422 12935 22390 13129 17233 7316 4068 9415 14471 6504 Panel B: Project-specific Data Benin: Youth nbr items 17 16 1 5 0 7 1 2 1 0 0 0 1 Employment mean 0.124 0.107 0.094 0.032 . 0.104 0.068 0.077 0.044 . . . 0.094 sd 0.984 0.990 0.955 0.994 . 0.973 0.952 0.964 1.013 . . . 0.955 obs 2967 2967 2965 2967 . 2967 2963 2967 2967 . . . 2965 Cronbach's alpha 0.764 0.754 . 0.327 . 0.679 . 0.617 . . . . . Congo: Skills nbr items 12 11 1 3 0 5 0 0 3 0 0 0 1 Development Project for mean 0.108 0.106 0.052 0.069 . 0.126 . . 0.007 . . . 0.052 Employability sd 1.029 1.040 0.920 1.028 . 0.970 . . 0.999 . . . 0.920 obs 3984 3984 407 3984 . 3984 . . 3984 . . . 407 Cronbach's alpha 0.454 0.435 . 0.350 . 0.495 . . 0.636 . . . . Côte d'Ivoire: Factory nbr items 15 11 4 6 1 3 0 1 0 0 1 3 0 Workers mean 0.106 0.109 0.031 0.071 0.050 0.074 . 0.067 . . 0.017 0.029 . sd 1.023 1.015 1.016 0.981 0.981 0.977 . 0.991 . . 1.015 0.999 . obs 1289 1289 1289 1289 1289 1289 . 1289 . . 1289 1289 . Cronbach's alpha 0.696 0.735 0.122 0.722 . 0.385 . . . . . 0.246 . Côte d'Ivoire: ProJeunes nbr items 85 47 38 0 9 7 10 13 8 10 6 12 10 mean 0.095 0.097 0.075 . 0.055 0.043 0.090 0.038 0.097 0.048 0.065 0.059 0.067 sd 1.066 1.031 1.079 . 1.059 1.050 1.039 1.057 0.945 1.050 1.030 1.063 1.066 obs 1126 1126 1126 . 1126 1126 1126 1126 1126 1126 1126 1126 1126 Cronbach's alpha 0.939 0.879 0.919 . 0.689 0.718 0.786 0.681 0.812 0.778 0.632 0.843 0.825 Côte d'Ivoire: Support nbr items 5 5 0 3 0 0 0 2 0 0 0 0 0 Project for the mean 0.035 0.035 . -0.052 . . . 0.087 . . . . . Agricultural Sector (PSAC) sd 0.967 0.967 . 1.012 . . . 0.917 . . . . . obs 1539 1539 . 1539 . . . 1518 . . . . . Cronbach's alpha 0.392 0.392 . 0.304 . . . 0.415 . . . . . Facebook: Future of nbr items 8 7 1 0 1 2 1 3 0 0 0 1 0 Business (FoB) mean 0.029 0.010 0.060 . -0.057 0.036 0.047 -0.022 . . . 0.060 . sd 1.059 1.064 1.018 . 1.005 1.040 1.092 1.064 . . . 1.018 . obs 7756 7724 6125 . 7059 7338 6414 6645 . . . 6125 . Cronbach's alpha 0.680 0.648 . . . 0.374 . 0.276 . . . . . Ghana: Impact of nbr items 6 6 0 0 0 2 4 0 0 0 0 0 0 Outgrower Contracts on mean 0.085 0.085 . . . 0.056 0.093 . . . . . . Smallholder Farmers sd 0.977 0.977 . . . 0.977 1.009 . . . . . . (GADCO) obs 1464 1464 . . . 1463 1464 . . . . . . Cronbach's alpha 0.380 0.380 . . . 0.383 0.065 . . . . . . Ghana: Skills Towards nbr items 20 11 9 1 4 2 0 4 0 1 4 4 0 Employability and mean 0.131 0.139 0.077 0.037 0.149 0.059 . 0.163 . 0.070 0.037 0.057 . Productivity (STEP)- Skills sd 1.007 1.007 1.000 0.974 1.004 0.990 . 1.015 . 1.022 0.987 0.987 . Measurement obs 1922 1916 1921 1881 1894 1900 . 1914 . 1852 1919 1902 . Cronbach's alpha 0.620 0.496 0.459 . 0.218 0.273 . 0.416 . . 0.121 0.419 . Kenya: Skills Towards nbr items 20 11 9 1 4 2 0 4 0 1 4 4 0 Employability and mean 0.082 0.090 0.044 0.046 0.082 0.018 . 0.099 . 0.026 0.052 0.022 . Productivity (STEP)- Skills sd 0.983 0.999 0.983 0.981 0.996 1.006 . 0.988 . 0.990 1.006 0.993 . Measurement obs 3822 3821 3822 3807 3820 3820 . 3821 . 3814 3822 3820 . Cronbach's alpha 0.642 0.522 0.486 . 0.286 0.327 . 0.423 . . 0.190 0.437 . Mozambique: Impact nbr items 33 29 4 5 0 5 11 5 3 0 0 2 2 Assessment Integrated mean 0.047 0.042 0.039 0.006 . 0.052 0.056 0.045 -0.036 . . 0.044 0.008 Growth Pole Project sd 0.991 0.994 1.002 1.022 . 0.980 0.992 0.990 0.988 . . 1.001 0.989 (IGPP) obs 5293 5293 5293 5293 . 5293 5293 5293 5293 . . 5293 5293 Cronbach's alpha 0.838 0.830 0.117 0.596 . 0.641 0.834 0.725 0.638 . . 0.494 Nigeria APPEALS nbr items 33 22 11 0 10 3 0 9 0 0 3 8 0 mean 0.097 0.098 0.077 . 0.084 0.065 . 0.110 . . 0.075 0.062 . sd 1.052 1.046 1.049 . 1.036 1.041 . 1.049 . . 1.050 1.033 . obs 5918 5918 5918 . 5918 5918 . 5918 . . 5918 5918 . Cronbach's alpha 0.941 0.911 0.901 . 0.837 0.604 . 0.847 . . 0.719 0.900 . Togo: Private Sector nbr items 44 37 7 9 2 1 13 11 1 2 2 3 0 Development Project mean 0.058 0.072 -0.004 0.099 0.032 0.061 0.100 0.167 -0.094 0.025 -0.040 0.008 . (PADSP)- Personal sd 1.026 1.024 1.014 0.997 0.985 1.001 1.001 1.010 1.144 1.032 0.994 1.030 . Initiative obs 1468 1468 1468 1467 1467 1464 1468 1468 1465 1468 1468 1468 . Cronbach's alpha 0.847 0.861 0.391 0.816 0.518 . 0.859 0.455 . 0.417 0.116 0.146 . Togo: Youth Employment nbr items 25 15 10 6 0 1 8 0 0 0 6 0 4 and Skills Development mean 0.081 0.013 0.143 0.009 . -0.005 0.040 . . . 0.077 . 0.146 (AIDE) sd 1.034 1.064 0.958 1.011 . 1.065 1.039 . . . 0.991 . 0.959 obs 3325 3325 3324 3324 . 3323 3324 . . . 3324 . 3324 Cronbach's alpha 0.695 0.577 0.542 0.435 . . 0.400 . . . 0.421 . 0.505 Note: nbr items stands for the number of items aggregated to construct the corresponding skills measure. sd stands for standard deviation and obs for the number of observations. Empty cells indicate that the project does not have data for the specified skill. PSDM stands for Problem Solving and Decision Making. The Facebook data is from the "Future of Business (FoB)" survey. This study includes only 15 Sub Saharan Africa (SSA) countries namely: Angola, Benin, Botswana, Cameroon, Côte d'Ivoire, Ethiopia, Ghana, Kenya, Mozambique, Nigeria, Senegal, Tanzania, Uganda, South Africa, and Zambia. http://documents1.worldbank.org/curated/en/734371558715932769/pdf/Tackling-the-Global-Profitarchy-Gender-and-the-Choice-of-Business-Sector.pdf. Dark green color indicates if the Cronbach's alpha is above 0.7 while light green color is for values between 0.6 and 0.69. 46 Table 3: Gender differences in levels of socio-emotional skills Model A: Model B: Model C: Model D: Model E: F-test comparing No control +age bins +married +edu dummies +employment coefficient on Woman for Model C and Model D Coef. on Woman Coef. on Woman Coef. on Woman Coef. on Woman Coef. on Woman P-value N All coef. -0.180*** -0.175*** -0.176*** -0.151*** -0.149*** 0.000 41,873 se (0.013) (0.013) (0.013) (0.013) (0.013) R-squared 0.008 0.012 0.012 0.021 0.021 Intra coef. -0.173*** -0.167*** -0.168*** -0.143*** -0.140*** 0.000 41,834 se (0.013) (0.013) (0.013) (0.013) (0.013) R-squared 0.008 0.012 0.012 0.021 0.021 Inter coef. -0.125*** -0.122*** -0.122*** -0.104*** -0.104*** 0.009 33,658 se (0.016) (0.016) (0.016) (0.016) (0.016) R-squared 0.005 0.006 0.006 0.011 0.011 Positive Self Concept coef. -0.092*** -0.086*** -0.086*** -0.060*** -0.057*** 0.091 25,551 se (0.014) (0.014) (0.014) (0.015) (0.015) R-squared 0.004 0.007 0.007 0.014 0.014 Emotional Regulation coef. -0.145*** -0.143*** -0.143*** -0.130*** -0.130*** 0.000 22,573 se (0.018) (0.018) (0.018) (0.018) (0.018) R-squared 0.008 0.011 0.011 0.013 0.013 Perseverance coef. -0.125*** -0.120*** -0.121*** -0.099*** -0.095*** 0.000 39,885 se (0.013) (0.013) (0.013) (0.013) (0.013) R-squared 0.005 0.007 0.007 0.012 0.012 Personal Initiative coef. -0.145*** -0.141*** -0.141*** -0.124*** -0.122*** 0.000 22,052 se (0.018) (0.018) (0.018) (0.018) (0.018) R-squared 0.005 0.007 0.007 0.015 0.015 PSDM coef. -0.193*** -0.186*** -0.186*** -0.160*** -0.160*** 0.000 31,959 se (0.014) (0.014) (0.015) (0.015) (0.015) R-squared 0.011 0.014 0.014 0.021 0.021 Self Control coef. -0.007 -0.005 -0.005 0.027 0.028 0.199 14,835 se (0.021) (0.021) (0.021) (0.021) (0.021) R-squared 0.004 0.007 0.007 0.015 0.016 Empathy coef. -0.083*** -0.085*** -0.084*** -0.051** -0.053** 0.263 8,260 se (0.025) (0.025) (0.025) (0.026) (0.026) R-squared 0.002 0.005 0.005 0.014 0.014 Expressiveness coef. -0.080*** -0.081*** -0.080*** -0.070*** -0.069*** 0.071 18,866 se (0.018) (0.018) (0.018) (0.018) (0.018) R-squared 0.003 0.003 0.003 0.007 0.007 Interpersonal Relatedness coef. -0.088*** -0.087*** -0.087*** -0.062*** -0.067*** 0.023 26,941 se (0.016) (0.016) (0.016) (0.017) (0.017) R-squared 0.002 0.003 0.003 0.010 0.011 Teamwork coef. -0.137*** -0.132*** -0.133*** -0.119*** -0.116*** 0.618 13,115 se (0.026) (0.026) (0.026) (0.027) (0.027) R-squared 0.007 0.009 0.009 0.012 0.013 Note: OLS regression specifications include study fixed effects. All studies have equal weights. Woman is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Age bins represent dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, 0 otherwise. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Education dummies represent dummy variables equal to 1 if the respondent's highest educational attainment (completed) is 0, 1, ... or 14 where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed high school, 13=completed certificate or diploma and 14=completed university degree or above, 0 otherwise. Note that Nigeria & Facebook projects have only categorical variable for education and "completed primary" is coded as 9 and "completed secondary" is coded as 12. Employment is a dummy variable equal to 1 if the respondent is currently working, 0 otherwise. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 47 Table 4: Gender differences in levels of socio-emotional skills-Heterogeneity by education attainment Panel A: Aggregate Skills All Intra Inter (1) (2) (3) Woman -0.121*** -0.129*** -0.016 (0.027) (0.027) (0.030) Education attainment 0.027*** 0.026*** 0.022*** (0.002) (0.002) (0.003) Woman*Education attainment -0.003 -0.002 -0.009*** (0.003) (0.003) (0.003) Observations 41,873 41,834 33,658 R-squared 0.019 0.019 0.009 P-val Woman+Woman*Edu Attain 0.000 0.000 0.363 P-val Edu Attain+Woman*Edu Attain 0.000 0.000 0.000 Mean SE skills for Men 0.133 0.117 0.124 Mean SE skills for Woman 0.000 0.000 0.000 Panel B: Intrapersonal Skills Positive Self Emotional Perseverance Personal Initiative PSDM Self Control Concept Regulation (1) (2) (3) (4) (5) (6) Woman -0.022 -0.036 -0.128*** -0.125*** -0.126*** 0.101*** (0.029) (0.053) (0.028) (0.030) (0.027) (0.034) Education attainment 0.024*** 0.015*** 0.016*** 0.021*** 0.023*** 0.027*** (0.003) (0.004) (0.003) (0.003) (0.003) (0.004) Woman*Education attainment -0.005 -0.009** 0.003 0.001 -0.004 -0.009** (0.003) (0.005) (0.003) (0.003) (0.003) (0.004) Observations 25,551 22,573 39,885 22,052 31,959 14,835 R-squared 0.012 0.012 0.011 0.012 0.019 0.012 P-val Woman+Woman*Edu Attain 0.310 0.358 0.000 0.000 0.000 0.003 P-val Edu Attain+Woman*Edu Attain 0.000 0.085 0.000 0.000 0.000 0.000 Mean SE skills for Men 0.062 0.067 0.099 0.100 0.126 -0.009 Mean SE skills for Woman 0.000 0.000 0.000 0.000 0.000 0.000 Panel C: Interpersonal Skills Interpersonal Empathy Expressiveness Teamwork Relatedness (1) (2) (3) (4) Woman 0.105 0.172*** 0.021 -0.029 (0.074) (0.058) (0.035) (0.035) Education attainment 0.032*** 0.025*** 0.024*** 0.020*** (0.006) (0.004) (0.003) (0.005) Woman*Education attainment -0.016** -0.023*** -0.009*** -0.010** (0.007) (0.005) (0.003) (0.004) Observations 8,260 18,866 26,941 13,115 R-squared 0.011 0.006 0.007 0.011 P-val Woman+Woman*Edu Attain 0.187 0.005 0.696 0.208 P-val Edu Attain+Woman*Edu Attain 0.000 0.435 0.000 0.019 Mean SE skills for Men 0.078 0.107 0.088 0.139 Mean SE skills for Woman 0.000 0.000 0.000 0.000 Note: OLS regression specifications include study fixed effects. All studies have equal weights. Woman is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Educational attainment stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Marital status and age bins are added as controls. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Age bins represent dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, 0 otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 48 Table 5: Correlations between socio-emotional skills and earnings Panel A: Earnings by Gender Earnings No control Column 1 + controls Column 2 + edu Column 3 + Column 2 + edu (Continuous) Woman*edu (dummies) (Continuous) (1) (2) (3) (4) (5) Woman -0.698*** -0.669*** -0.647*** -0.558*** -0.644*** (0.032) (0.031) (0.032) (0.060) (0.032) Education attainment 0.021*** 0.027*** (0.005) (0.006) Woman*Education attainment -0.010 (0.006) Observations 33,965 33,965 33,965 33,965 33,965 R-squared 0.193 0.239 0.239 0.239 0.241 P-val Edu Attain+Woman*Edu Attain 0.002 Mean Earnings for Men 169.3 169.3 169.3 169.3 169.3 Mean Earnings for Woman 103.2 103.2 103.2 103.2 103.2 Panel B: Aggregate Skills Earnings All Intra Inter (1) (2) (3) Woman -0.550*** -0.553*** -0.633*** (0.060) (0.060) (0.061) SE skills 0.068*** 0.081*** -0.017 (0.021) (0.020) (0.026) Woman*SE skills 0.020 -0.006 0.078** (0.030) (0.030) (0.034) Education attainment 0.025*** 0.024*** 0.024*** (0.006) (0.006) (0.007) Woman*Education attainment -0.010 -0.009 -0.009 (0.006) (0.006) (0.006) Observations 33,965 33,957 27,273 R-squared 0.240 0.241 0.283 P-val SE skills+Woman*SE skills=0 0.000 0.000 0.006 P-val Woman + Woman*SE skills=0 0.000 0.000 0.000 Mean monthly earnings for Men 169.3 169.3 185.5 Mean monthly earnings for Women 103.2 102.8 105.6 Panel C: Intrapersonal Skills Earnings Positive Self Emotional Perseverance Personal Initiative PSDM Self Control Concept Regulation (1) (2) (3) (4) (5) (6) Woman -0.494*** -0.883*** -0.628*** -0.504*** -0.558*** -0.628*** (0.059) (0.127) (0.061) (0.068) (0.061) (0.065) SE skills 0.112*** 0.022 0.046** 0.030 0.041 0.043 (0.022) (0.032) (0.022) (0.031) (0.026) (0.027) Woman*SE skills -0.065** -0.039 0.056* 0.036 0.018 -0.044 (0.031) (0.045) (0.030) (0.043) (0.035) (0.038) Education attainment 0.044*** 0.009 0.018*** 0.004 0.032*** -0.011 (0.006) (0.011) (0.007) (0.009) (0.007) (0.009) Woman*Education attainment -0.026*** 0.008 -0.006 -0.014 -0.015** -0.007 (0.007) (0.012) (0.006) (0.009) (0.007) (0.009) Observations 24,835 15,691 32,359 15,726 25,481 14,817 R-squared 0.284 0.203 0.248 0.281 0.290 0.354 P-val SE skills+Woman*SE skills=0 0.036 0.584 0.000 0.0240 0.0120 0.987 P-val Woman + Woman*SE skills=0 0.000 0.000 0.000 0.000 0.000 0.000 Mean monthly earnings for Men 202 272 160.8 81.16 203 55.01 Mean monthly earnings for Women 120.3 160 99.25 42.55 109.6 31.04 Panel D: Interpersonal Skills Earnings Empathy Expressiveness Interpersonal Teamwork Relatedness (1) (2) (3) (4) Woman -0.686*** -0.832*** -0.555*** -0.668*** (0.177) (0.124) (0.072) (0.069) SE skills -0.044 -0.001 -0.067** -0.016 (0.040) (0.029) (0.027) (0.042) Woman*SE skills 0.109* 0.056 0.058 0.136*** (0.060) (0.040) (0.037) (0.053) Education attainment 0.024* 0.014 0.035*** -0.008 (0.014) (0.010) (0.008) (0.011) Woman*Education attainment -0.028 -0.001 -0.019** -0.011 (0.018) (0.011) (0.008) (0.009) Observations 7,930 18,154 20,943 12,728 R-squared 0.245 0.162 0.324 0.318 P-val SE skills+Woman*SE skills=0 0.143 0.053 0.728 0.000 P-val Woman + Woman*SE skills=0 0.002 0.000 0.000 0.000 Mean monthly earnings for Men 452.1 257.8 209.5 73.87 Mean monthly earnings for Women 275.2 155.8 117.8 28.52 Note: OLS regression specifications include study fixed effects. Panel B, C and D include socio-emotional skills as controls. All studies have equal weights. Women is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Earnings is the inverse hyperbolic sine (IHS) transformation of the respondent's monthly earnings in US dollars. Note that only business profit rather than total earnings is reported for Nigeria. Education, age bins and marital status are added as controls. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Age bin represents dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, o otherwise. Married is added as a control and defined as a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 49 Table 6: Correlations between socio-emotional skills and any earnings by education Panel A: Aggregate Skills Earnings All Intra Inter (1) (2) (3) Woman -0.548*** -0.550*** -0.641*** (0.060) (0.060) (0.061) Education attainment 0.025*** 0.025*** 0.023*** (0.006) (0.006) (0.007) Woman*Education attainment -0.010 -0.010 -0.008 (0.006) (0.006) (0.006) SE skills 0.109** 0.139*** -0.103* (0.049) (0.049) (0.053) Woman*SE skills -0.079 -0.116* 0.130** (0.060) (0.060) (0.063) Education attainment*SE skills -0.004 -0.006 0.008 (0.005) (0.005) (0.005) Woman*Education attainment*SE skills 0.012* 0.013** -0.004 (0.006) (0.006) (0.007) Observations 33,965 33,957 27,273 R-squared 0.240 0.241 0.283 P-val SE skills+Edu Attain*SE skills 0.02 0.003 0.052 P-val SE skills+14*Edu Attain*SE skills 0.098 0.058 0.713 P-val SE skills+Woman*SE skills 0.381 0.510 0.425 P-val SE skills+Woman*SE skills+Edu Attain*SE skills+Woman*Edu Attain*SE skills 0.230 0.347 0.309 P-val SE skills+Woman*SE skills+14*Edu Attain*SE skills+14*Woman*Edu Attain*SE skills 0.000 0.001 0.017 P-val Edu Attain*SE skills+Woman*Edu Attain*SE skills 0.064 0.092 0.280 Panel B: Intrapersonal Skills Earnings Positive Self Personal Emotional Regulation Perseverance PSDM Self Control Concept Initiative (1) (2) (3) (4) (5) (6) Woman -0.493*** -0.902*** -0.632*** -0.505*** -0.565*** -0.631*** (0.059) (0.127) (0.061) (0.068) (0.061) (0.065) Education attainment 0.045*** 0.007 0.017** 0.004 0.031*** -0.011 (0.006) (0.011) (0.007) (0.009) (0.007) (0.009) Woman*Education attainment -0.026*** 0.010 -0.005 -0.014 -0.014** -0.007 (0.007) (0.012) (0.006) (0.009) (0.007) (0.009) SE skills 0.244*** -0.163 0.004 0.029 -0.016 0.177*** (0.046) (0.109) (0.053) (0.063) (0.052) (0.054) Woman*SE skills -0.226*** -0.003 0.094 0.001 0.045 -0.105 (0.058) (0.126) (0.063) (0.072) (0.063) (0.065) Education attainment*SE skills -0.015*** 0.017* 0.004 0.000 0.006 -0.015** (0.005) (0.010) (0.005) (0.006) (0.005) (0.007) Woman*Education attainment*SE skills 0.019*** 0.000 -0.004 0.005 -0.002 0.003 (0.007) (0.012) (0.006) (0.009) (0.007) (0.009) Observations 24,835 15,691 32,359 15,726 25,481 14,817 R-squared 0.285 0.204 0.248 0.281 0.290 0.355 P-val SE skills+Edu Attain*SE skills 0.000 0.142 0.865 0.611 0.838 0.001 P-val SE skills+14*Edu Attain*SE skills 0.320 0.104 0.047 0.506 0.077 0.481 P-val SE skills+Woman*SE skills 0.599 0.009 0.003 0.380 0.401 0.043 P-val SE skills+Woman*SE skills+Edu Attain*SE skills+Woman*Edu Attain*SE skills 0.467 0.011 0.001 0.245 0.292 0.057 P-val SE skills+Woman*SE skills+14*Edu Attain*SE skills+14*Woman*Edu Attain*SE skills 0.067 0.098 0.004 0.095 0.038 0.126 P-val Edu Attain*SE skills+Woman*Edu Attain*SE skills 0.326 0.007 0.897 0.354 0.351 0.047 Panel C: Interpersonal Skills Earnings Interpersonal Empathy Expressiveness Teamwork Relatedness (1) (2) (3) (4) Woman -0.683*** -0.829*** -0.563*** -0.660*** (0.178) (0.123) (0.072) (0.070) Education attainment 0.024* 0.013 0.034*** -0.007 (0.014) (0.010) (0.008) (0.011) Woman*Education attainment -0.028 -0.000 -0.018** -0.012 (0.018) (0.011) (0.008) (0.009) SE skills -0.100 -0.263*** -0.224*** 0.087 (0.135) (0.102) (0.055) (0.061) Woman*SE skills 0.252 0.261** 0.170** -0.010 (0.173) (0.121) (0.070) (0.070) Education attainment*SE skills 0.005 0.023*** 0.016*** -0.011 (0.013) (0.009) (0.006) (0.007) Woman*Education attainment*SE skills -0.015 -0.017 -0.010 0.017* (0.017) (0.011) (0.007) (0.009) Observations 7,930 18,154 20,943 12,728 R-squared 0.245 0.162 0.324 0.319 P-val SE skills+Edu Attain*SE skills 0.440 0.0110 0.000 0.166 P-val SE skills+14*Edu Attain*SE skills 0.688 0.0890 0.891 0.318 P-val SE skills+Woman*SE skills 0.162 0.974 0.206 0.028 P-val SE skills+Woman*SE skills+Edu Attain*SE skills+Woman*Edu Attain*SE skills 0.146 0.943 0.215 0.008 P-val SE skills+Woman*SE skills+14*Edu Attain*SE skills+14*Woman*Edu Attain*SE skills 0.816 0.029 0.542 0.016 P-val Edu Attain*SE skills+Woman*Edu Attain*SE skills 0.416 0.304 0.241 0.301 Note: OLS regression specifications control for age and include study fixed effects. All studies have equal weights. Woman is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Earnings is the inverse hyperbolic sine (IHS) transformation of the respondent's monthly earnings in US dollars. Note that only business profit rather than total earnings is reported for Nigeria. Education, age bins and marital status are added as controls. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Age bins are added as controls. Age bins represent dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, 0 otherwise. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 50 Table 7: Correlations between socio-emotional skills and earnings - Heterogeneity along Employment type Panel A: Aggregate Skills Earnings NonAgr SE Wage All Intra Inter All Intra Inter (1) (2) (3) (4) (5) (6) Woman -0.488*** -0.492*** -0.515*** -0.736*** -0.738*** -0.776*** (0.083) (0.082) (0.080) (0.156) (0.156) (0.122) SE skills 0.088*** 0.084*** 0.041 0.135* 0.124* 0.047 (0.030) (0.029) (0.034) (0.077) (0.067) (0.080) Woman*SE skills 0.007 0.002 0.027 -0.072 -0.100 0.058 (0.041) (0.042) (0.043) (0.083) (0.074) (0.086) Education attainment 0.039*** 0.039*** 0.046*** 0.024 0.024 0.009 (0.008) (0.008) (0.008) (0.019) (0.018) (0.017) Woman*Education attainment -0.007 -0.007 -0.012 0.038** 0.038** 0.036*** (0.009) (0.009) (0.009) (0.015) (0.015) (0.014) Observations 10,953 10,950 8,311 14,233 14,232 11,856 R-squared 0.436 0.436 0.501 0.629 0.628 0.674 P-val SE skills+Woman*SE skills=0 0.001 0.005 0.010 0.136 0.551 0.005 P-val Woman + Woman*SE skills=0 0.000 0.000 0.000 0.000 0.000 0.000 Mean monthly earnings for Men 269.9 269.9 304.9 143.7 143.7 172.7 Mean monthly earnings for Women 177.7 176.6 196.3 76.85 76.85 80.64 PANEL B: NON-AGRICULTURAL SELF-EMPLOYMENT Earnings Problem Positive Self Emotional Personal Interpersonal Perseverance Solving & Self Control Empathy Expressiveness Teamwork Concept Regulation Initiative Relatedness DecMaking (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Woman -0.528*** -0.534*** -0.471*** -0.286*** -0.559*** -0.376*** -0.094 -0.497*** -0.485*** -0.510*** (0.081) (0.156) (0.085) (0.097) (0.079) (0.082) (0.188) (0.161) (0.098) (0.095) SE skills 0.114*** -0.015 0.014 0.016 0.075*** 0.053* -0.030 0.046 -0.002 0.018 (0.030) (0.040) (0.032) (0.050) (0.028) (0.028) (0.049) (0.041) (0.033) (0.056) Woman*SE skills -0.088** 0.004 0.098** 0.133** -0.009 -0.052 0.079 -0.002 0.010 0.101 (0.043) (0.058) (0.042) (0.067) (0.041) (0.043) (0.067) (0.056) (0.047) (0.072) Education attainment 0.042*** 0.047*** 0.046*** 0.042*** 0.037*** 0.051*** 0.071*** 0.048*** 0.052*** 0.045*** (0.007) (0.011) (0.009) (0.012) (0.007) (0.011) (0.015) (0.012) (0.008) (0.014) Woman*Education attainment -0.011 0.000 -0.013 -0.034*** 0.006 -0.026** -0.043** -0.015 -0.004 -0.030** (0.010) (0.015) (0.009) (0.013) (0.009) (0.011) (0.019) (0.015) (0.010) (0.013) Observations 8,692 4,611 10,559 6,050 7,939 6,492 3,153 5,033 5,781 4,033 R-squared 0.527 0.474 0.432 0.225 0.554 0.33 0.408 0.438 0.593 0.298 P-val SE skills+Woman*SE skills=0 0.405 0.783 0.000 0.001 0.026 0.989 0.290 0.249 0.796 0.008 P-val Woman + Woman*SE skills=0 0.000 0.001 0.000 0.205 0.000 0.000 0.938 0.003 0.000 0.000 Mean monthly earnings for Men 301.2 521.3 256.5 100.9 343.7 85.11 671.4 484.2 381 64.86 Mean monthly earnings for Women 202 310 171.1 72.84 208.8 57.49 452.1 297.9 267.9 41.15 PANEL C: WAGE EMPLOYMENT Earnings Problem Positive Self Emotional Personal Interpersonal Perseverance Solving & Self Control Empathy Expressiveness Teamwork Concept Regulation Initiative Relatedness DecMaking (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Woman -0.653*** -0.932*** -0.759*** -0.863*** -0.845*** -0.927*** -0.984*** -0.887*** -0.479*** -0.798*** (0.168) (0.238) (0.109) (0.131) (0.177) (0.157) (0.337) (0.231) (0.134) (0.113) SE skills 0.008 0.047 0.163 -0.078 0.096 0.139 0.141 0.155 -0.124** -0.118** (0.067) (0.102) (0.117) (0.081) (0.071) (0.146) (0.129) (0.106) (0.062) (0.051) Woman*SE skills 0.023 -0.087 -0.105 0.145 -0.053 -0.208 0.038 -0.070 0.167** 0.195*** (0.096) (0.110) (0.116) (0.102) (0.081) (0.155) (0.147) (0.112) (0.069) (0.069) Education attainment 0.044** -0.009 0.010 -0.010 0.016 -0.018 -0.002 -0.004 0.016 0.009 (0.022) (0.023) (0.014) (0.020) (0.022) (0.026) (0.028) (0.023) (0.019) (0.016) Woman*Education attainment 0.037** 0.054** 0.033*** 0.026 0.053*** 0.030 0.054* 0.044** 0.015 0.004 (0.018) (0.022) (0.012) (0.017) (0.019) (0.025) (0.033) (0.021) (0.015) (0.013) Observations 11,089 5,406 14,135 8,117 9,871 6,917 2,701 6,934 9,221 7,242 R-squared 0.581 0.71 0.663 0.509 0.674 0.494 0.554 0.658 0.759 0.487 P-val SE skills+Woman*SE skills=0 0.626 0.368 0.047 0.220 0.312 0.271 0.051 0.080 0.242 0.097 P-val Woman + Woman*SE skills=0 0.000 0.000 0.000 0.000 0.000 0.000 0.020 0.001 0.030 0.000 Mean monthly earnings for Men 179.6 269.3 143.9 72.91 171.5 35.34 448.2 261 180.4 83.95 Mean monthly earnings for Women 93.64 151.2 76.56 24.27 75.03 8.467 325 150.7 78.22 25.06 Note: OLS regression specifications include study fixed effects. All studies have equal weights. Woman is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Earnings is the inverse hyperbolic sine (IHS) transformation of the respondent's monthly earnings in US dollars. Note that only business profit rather than total earnings is reported for Nigeria. Education, age bins and marital status are added as controls. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Age bin represents dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, 0 otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 51 Table 8: Summary of Results - Pooled Sample and By Project Positive Self Emotional Personal Interpersonal Project All Intra Inter Perseverance PSDM Self Control Empathy Expressiveness Teamwork Concept Regulation Initiative Relatedness Panel A: Gender differences in levels of socio-emotional skills: Coeff on Women All -0.151*** -0.143*** -0.104*** -0.06*** -0.13*** -0.099*** -0.124*** -0.16*** 0.027 -0.051** -0.07*** -0.062*** -0.119*** Benin: Youth Employment -0.281*** -0.238*** -0.224*** -0.051 . -0.226*** -0.13*** -0.187*** -0.12*** . . . -0.224*** Congo: Skills Development Project for Employability -0.169*** -0.166*** -0.137 -0.104*** . -0.191*** . . -0.019 . . . -0.137 Cote D'Ivoire: Factory Workers -0.153** -0.143** -0.068 -0.108 -0.069 -0.086 . -0.083 . . -0.084 0.008 . Cote D'Ivoire: ProJuenes -0.137** -0.146** -0.102 . -0.081 -0.075 -0.149** -0.072 -0.116** -0.064 -0.114* -0.059 -0.086 Côte D'Ivoire: Support Project for the Agricultural Sector (PSAC) -0.081 -0.081 . 0.018 . . . -0.121* . . . . . Facebook: Future of Business (FoB) -0.001 0.025 -0.061* . 0.112*** -0.006 -0.054 0.058* . . . -0.061* . Ghana: Impact of Outgrower Contracts on Smallholder Farmers (GADCO) -0.106* -0.106* . . . -0.054 -0.141** . . . . . . Ghana: Skills Towards Employability and Productivity (STEP)- Skills -0.18*** -0.211*** -0.083* -0.042 -0.262*** -0.08* . -0.262*** . -0.091* -0.043 -0.04 . Measurement Kenya: Skills Towards Employability and Productivity (STEP)- Skills -0.117*** -0.142*** -0.045 -0.071** -0.164*** -0.016 . -0.145*** . -0.02 -0.072** -0.016 . Measurement Mozambique: Impact Assessment Integrated Growth Pole Project (IGPP) -0.087*** -0.075** -0.077** 0.031 . -0.107*** -0.126*** -0.104*** 0.086*** . . -0.11*** 0.016 Nigeria APPEALS -0.205*** -0.208*** -0.163*** . -0.188*** -0.138*** . -0.222*** . . -0.162*** -0.126*** . Togo: Private Sector Development Project (PADSP)- Personal Initiative 0.029 0.028 0.012 -0.041 -0.024 -0.012 -0.061 -0.228*** 0.258*** -0.021 0.058 -0.015 . Togo: Youth Employment and Skills Development (AIDE) -0.16*** -0.064 -0.227*** -0.071* . 0 -0.118*** . . . -0.142*** . -0.217*** Panel B: Gender differences in levels of socio-emotional skills-Heterogeneity by education Gender Diff for those with no formal education: Coeff on Women All -0.121*** -0.129*** -0.016 -0.022 -0.036 -0.128*** -0.125*** -0.126*** 0.101*** 0.105 0.172*** 0.021 -0.029 Benin: Youth Employment -0.175*** -0.146** -0.144** -0.089 . -0.151** -0.104* -0.094 -0.03 . . . -0.144** Congo: Skills Development Project for Employability 0.102 0.085 -0.59 0.12 . 0.169 . . -0.1 . . . -0.59 Cote D'Ivoire: Factory Workers -0.187 -0.192 -0.054 -0.173 -0.076 0.022 . -0.202 . . -0.129 0.106 . Cote D'Ivoire: ProJuenes 0.201 0.131 0.236 . 0.13 -0.004 0.006 0.194 0.146 0.268 0.095 0.152 0.233 Côte D'Ivoire: Support Project for the Agricultural Sector (PSAC) 0.09 0.09 . 0.069 . . . 0.08 . . . . . Facebook: Future of Business (FoB) -0.595** -0.486** -0.322 . -0.271 -0.334 -0.458 -0.222 . . . -0.322 . Ghana: Impact of Outgrower Contracts on Smallholder Farmers (GADCO) -0.096 -0.096 . . . -0.122 -0.016 . . . . . . Ghana: Skills Towards Employability and Productivity (STEP)- Skills -0.078 -0.167 0.055 -0.187 -0.068 -0.081 . -0.119 . 0.045 0.244 -0.104 . Measurement Kenya: Skills Towards Employability and Productivity (STEP)- Skills 0.055 -0.007 0.12 0.073 -0.003 -0.043 . -0.081 . 0.084 0.066 0.111 . Measurement Mozambique: Impact Assessment Integrated Growth Pole Project (IGPP) -0.105** -0.105** -0.067 0.001 . -0.155*** -0.139*** -0.112*** 0.096** . . -0.117*** 0.041 Nigeria APPEALS 0.747* 0.995** 0.295 . 0.701 0.857** . 1.036** . . 0.569 -0.057 . Togo: Private Sector Development Project (PADSP)- Personal Initiative 0.257* 0.279* 0.061 -0.029 0.355** 0.285* -0.021 -0.189 0.236 -0.12 0.207 0.045 . Togo: Youth Employment and Skills Development (AIDE) 0.05 0.122 -0.082 -0.092 . 0.364 -0.351 . . . 0.12 . -0.213 Gender Diff with each additional year of education: Coeff on Women*Edu Attain All -0.003 -0.002 -0.009*** -0.005 -0.009** 0.003 0.001 -0.004 -0.009** -0.016** -0.023*** -0.009*** -0.01** Benin: Youth Employment -0.025** -0.021** -0.02** 0.008 . -0.017* -0.007 -0.021** -0.019* . . . -0.02** Congo: Skills Development Project for Employability -0.025 -0.023 0.048 -0.02 . -0.034* . . 0.008 . . . 0.048 Cote D'Ivoire: Factory Workers -0.01 -0.009 -0.006 -0.003 -0.005 -0.029** . 0.009 . . 0.006 -0.021 . Cote D'Ivoire: ProJuenes -0.032* -0.026 -0.032* . -0.021 -0.006 -0.014 -0.025 -0.023 -0.031* -0.021 -0.02 -0.03* Côte D'Ivoire: Support Project for the Agricultural Sector (PSAC) -0.023 -0.023 . -0.006 . . . -0.027** . . . . . Facebook: Future of Business (FoB) 0.044** 0.038** 0.02 . 0.028 0.023 0.031 0.021 . . . 0.02 . Ghana: Impact of Outgrower Contracts on Smallholder Farmers (GADCO) -0.002 -0.002 . . . 0.006 -0.013 . . . . . . Ghana: Skills Towards Employability and Productivity (STEP)- Skills -0.011 -0.005 -0.015 0.014 -0.019 0 . -0.015 . -0.015 -0.029* 0.005 . Measurement Kenya: Skills Towards Employability and Productivity (STEP)- Skills -0.018** -0.014 -0.018** -0.015* -0.016* 0.003 . -0.008 . -0.012 -0.015* -0.013 . Measurement Mozambique: Impact Assessment Integrated Growth Pole Project (IGPP) 0.007 0.011 -0.001 0.008 . 0.017* 0.006 0.005 -0.004 . . 0.005 -0.008 Nigeria APPEALS -0.071** -0.09*** -0.034 . -0.066** -0.074** . -0.094*** . . -0.055* -0.005 . Togo: Private Sector Development Project (PADSP)- Personal Initiative -0.023 -0.026* -0.004 -0.001 -0.041*** -0.031** -0.002 -0.004 0.003 0.012 -0.015 -0.007 . Togo: Youth Employment and Skills Development (AIDE) -0.015 -0.014 -0.011 0.001 . -0.027 0.017 . . . -0.019 . 0 52 Table 8 (cont'd): Summary of Results - Pooled Sample and By Project Positive Self Emotional Personal Interpersonal Project All Intra Inter Perseverance PSDM Self Control Empathy Expressiveness Teamwork Concept Regulation Initiative Relatedness Panel C: Earnings Gender Diff in Correlations (Earnings): Coeff on SE skills*Women All 0.020 -0.006 0.078** -0.065** -0.039 0.056* 0.036 0.018 -0.044 0.109* 0.056 0.058 0.136** Benin: Youth Employment -0.062 -0.053 -0.024 -0.064 . 0.055 -0.127 -0.012 -0.030 . . . -0.024 Congo: Skills Development Project for Employability -0.075 -0.072 0.132 -0.046 . -0.078 . . -0.053 . . . 0.132 Cote D'Ivoire: Factory Workers -0.155 -0.196* 0.012 -0.220** -0.102 -0.006 . -0.162 . . 0.063 -0.091 . Cote D'Ivoire: ProJuenes 0.192 0.101 0.263* . 0.101 0.195 0.072 0.072 -0.091 0.209 0.232 0.066 0.361** Côte D'Ivoire: Support Project for the Agricultural Sector (PSAC) 0.055 0.055 . 0.130 . . . -0.021 . . . . . Facebook: Future of Business (FoB) . . . . . . . . . . . . . Ghana: Impact of Outgrower Contracts on Smallholder Farmers (GADCO) 0.328** 0.328** . . . 0.426*** 0.057 . . . . . . Ghana: Skills Towards Employability and Productivity (STEP)- Skills -0.150 -0.223* 0.016 -0.149 -0.095 -0.091 . -0.197 . -0.013 -0.007 0.103 . Measurement Kenya: Skills Towards Employability and Productivity (STEP)- Skills 0.171* 0.096 0.191* 0.082 -0.098 0.089 . 0.129 . 0.113 0.191* 0.096 . Measurement Mozambique: Impact Assessment Integrated Growth Pole Project (IGPP) 0.005 -0.003 0.018 -0.104*** . 0.068** 0.068** 0.055* -0.073** . . 0.051* -0.041 Nigeria APPEALS 0.046 0.030 0.055 . 0.009 0.04 . 0.024 . . 0.045 0.050 . Togo: Private Sector Development Project (PADSP)- Personal Initiative -0.013 -0.010 -0.016 -0.156* 0.049 0.031 -0.014 0.088 -0.031 0.029 -0.090 0.027 . Togo: Youth Employment and Skills Development (AIDE) -0.008 -0.060 0.071 0.082 . -0.180* 0.135 . . . -0.022 . 0.123 Correlations for Men (Earnings): Coeff on SE skills All 0.068*** 0.081*** -0.017 0.112*** 0.022 0.046** 0.030 0.041 0.043 -0.044 -0.001 -0.067** -0.016 Benin: Youth Employment 0.307*** 0.299*** 0.126* 0.200*** . 0.254*** 0.185*** 0.235*** 0.111** . . . 0.126* Congo: Skills Development Project for Employability 0.135*** 0.131*** -0.046 0.082** . 0.319*** . . -0.071** . . . -0.046 Cote D'Ivoire: Factory Workers 0.088 0.084 0.040 0.144 0.024 0.015 . 0.067 . . -0.015 0.110 . Cote D'Ivoire: ProJuenes -0.182* -0.190* -0.153 . -0.141 -0.138 -0.169* -0.108 -0.08 -0.141 -0.102 -0.118 -0.161* Côte D'Ivoire: Support Project for the Agricultural Sector (PSAC) 0.172** 0.172** . 0.180** . . . 0.087 . . . . . Facebook: Future of Business (FoB) . . . . . . . . . . . . . Ghana: Impact of Outgrower Contracts on Smallholder Farmers (GADCO) 0.019 0.019 . . . -0.101 0.178** . . . . . . Ghana: Skills Towards Employability and Productivity (STEP)- Skills -0.006 0.065 -0.092 0.137 0.057 0.000 . -0.017 . -0.025 -0.018 -0.17* . Measurement Kenya: Skills Towards Employability and Productivity (STEP)- Skills -0.027 0.048 -0.101 -0.012 0.094 0.035 . 0.039 . -0.006 -0.082 -0.141* . Measurement Mozambique: Impact Assessment Integrated Growth Pole Project (IGPP) 0.049** 0.053** 0.020 0.100*** . -0.042* -0.074*** -0.039* 0.163*** . . -0.048** 0.103*** Nigeria APPEALS 0.101*** 0.113*** 0.071* . 0.087** 0.078** . 0.133*** . . 0.074** 0.054 . Togo: Private Sector Development Project (PADSP)- Personal Initiative 0.186*** 0.185*** 0.081 0.175*** 0.025 0.109* 0.203*** 0.099* 0.065 0.029 0.128** 0.004 . Togo: Youth Employment and Skills Development (AIDE) 0.025 0.025 0.012 0.006 . 0.055 -0.060 . . . 0.010 . 0.009 Correlations for Women (Earnings): Coeff on SE skills+ SE skills*Women All 0.087*** 0.075*** 0.061*** 0.048** -0.017 0.102*** 0.067** 0.059** 0.000 0.065 0.055* -0.009 0.119*** Benin: Youth Employment 0.245*** 0.246*** 0.102** 0.136*** . 0.310*** 0.058 0.224*** 0.082* . . . 0.102** Congo: Skills Development Project for Employability 0.060 0.059 0.086 0.036 . 0.241*** . . -0.124** . . . 0.086 Cote D'Ivoire: Factory Workers -0.068 -0.112** 0.052 -0.076 -0.078 0.009 . -0.095* . . 0.048 0.020 . Cote D'Ivoire: ProJuenes 0.010 -0.089 0.110 . -0.040 0.057 -0.097 -0.036 -0.171 0.068 0.131 -0.052 0.200* Côte D'Ivoire: Support Project for the Agricultural Sector (PSAC) 0.227 0.227 . 0.310** . . . 0.066 . . . . . Facebook: Future of Business (FoB) . . . . . . . . . . . . . Ghana: Impact of Outgrower Contracts on Smallholder Farmers (GADCO) 0.347*** 0.347*** . . . 0.324*** 0.235* . . . . . . Ghana: Skills Towards Employability and Productivity (STEP)- Skills -0.156* -0.158* -0.076 -0.013 -0.039 -0.091 . -0.214** . -0.037 -0.024 -0.067 . Measurement Kenya: Skills Towards Employability and Productivity (STEP)- Skills 0.143** 0.145** 0.091 0.070 -0.004 0.124* . 0.168** . 0.106 0.109 -0.045 . Measurement Mozambique: Impact Assessment Integrated Growth Pole Project (IGPP) 0.053*** 0.051** 0.038* -0.004 . 0.025 -0.006 0.016 0.089*** . . 0.002 0.061*** Nigeria APPEALS 0.147*** 0.142*** 0.126*** . 0.096** 0.118*** . 0.157*** . . 0.119*** 0.104*** . Togo: Private Sector Development Project (PADSP)- Personal Initiative 0.173*** 0.175*** 0.065 0.018 0.074 0.140** 0.189*** 0.186*** 0.035 0.058 0.038 0.031 . Togo: Youth Employment and Skills Development (AIDE) 0.017 -0.035 0.083 0.088 . -0.124 0.075 . . . -0.011 . 0.132 Note: OLS regression specifications and study level analysis with the aggregate results(ALL) at the top of each section. The aggregated(All) regressions include study fixed effect. Panel A shows the gender difference in SE skills while panel B represents the heterogeneity of gender difference in SE skills by education. Panel C shows the correlation between SE skills and earnings for men and women. Woman is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Earnings is the inverse hyperbolic sine (IHS) transformation of the respondent's monthly earnings in US dollars. Note that only business profit rather than total earnings is reported for Nigeria. All of the regressions include education, age bins and marital status as controls. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Age bin represents dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, o otherwise. Married is added as a control and defined as a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 53 Appendix Appendix Table A1: Existing literature on Gender Differences in Socio-Emotional skills SE skills Associated Concept: Gender Difference (Source) Key considerations Positive Self- Self-Esteem: Gender difference Concept • F>M (Robins et al., 2002; Kling, Hyde, Showers, & Buswell, 1999; arises in Twenge & Campbell, 2001; Bleidorn et al., 2015; Gentile et al., 2009) adolescence and • F=M (Erol & Orth, 2011) persists; differences Self-Efficacy: Entrepreneurial FM (Alloy et al., 2000; Hankin & Abramson, 2001/2002; Nolen-Hoeksema 2001; Nolen-Hoeksema Tied to anxiety, 2012; Hyde, Mezulis, Abramson, 2008; McRae et al., 2008) depression, alcohol Biological Emotional Reactivity: F=M (McRae et al., 2008) use Cognitive Reappraisal: F=M ; Expressive Suppression: F>M (Gross & John, 2003) Physical aggression: FM (Cobb-Clark & Tan, 2010; Schmitt et al., 2008) Differentiates Conscientiousness (achievement striving facet): F=M (Costa et al., between steady 2001) effort and persisting Grit (Consistency of interest + Perseverance): until a challenge is • F>M (Christenson & Knezek, 2014) complete • F=M (Crede, 2017; Bazelais et al., 2016) Persistence in GRIT Perseverance subscale: F>M (Christenson & Knezek, 2014); FM (Steinberg et al., 2009) higher among men Personal Growth Initiative: F>M (Robitscheck & Cook, 1999) Other terms include Achievement/Goal Orientation 54 Problem Positive Problem Orientation : FM (Duckworth & Segilman, 2006; Silverman, 2003; Tied to risk taking, Gibson et al., 2010; Hyde, 2013; Nakhaie et al., 2000) obesity, substance Impulsivity: abuse, depression, • F=M (Feingold 1994) anxiety, parenting, • FM (Duckworth & Segilman, effort and task 2006) performance Inhibitory control & Attention: F>M (Hyde, 2013) especially among Sensation seeking: FM (Cross et al., 2011) Disruptive behavior (classroom): FM (Mestre et al., 2009; Toussaint Differs with gender & Webb, 2005; Obrien et al, 2013) of the interacting Facial expression processing: F>M (McClure, 2000) individual (Stuijfzand et al., 2016) Size of gender difference varies with focus on affective or cognitive empathy Expressiveness Assertive communication: Differs with age • FM (Park et al., 2016) themselves or Tentative, hedging, descriptive language: F>M (Leaper & Ayres, 2007 ; others, and whether Newman et al., 2008; Mulac et al., 2001) they expect returns Clarity in expression of emotions: F>M (Wagner, 1993) to assertive behavior (Amanatullah & Morris, 2010), and the gender of the interacting individual (Bowles, 2010). 55 Interpersonal Agreeableness: F>M (Costa et al., 2001; Cobb-Clark & Tan, 2010; Differences in trust Relatedness Feingold 1994; Mueller & Plug 2006; Schmitt et al., 2008; and trustworthiness Gunewardena et al., 2018) are moderated by Extraverson: social value • FM (Schmitt et al., 2008), F>M in 4/9 countries (Gunewardena et al., al., 2009), 2018) expectations of • Warmth facet of Extroversion: F>M (Costa et al., 2001) return, and • Gregariousness facet of Extroversion: F>M (Costa et al., 2001) unconditional Altruism: F>M (Croson & Gneezy, 2009) kindness (Ashraf et Forgiveness: F>M (Miller et al., 2008) al., 2006) and not Charitable giving: F>M (Willer et al., 2015) risk aversion Trust & Reciprocity: F=M, no diff, sometimes higher for women (Ashraf et al., 2006; Croson & Buchan, 1999; Buchan et al., 2008) Self-esteem is a Network size: F=M (Mengel, 2020) moderator of Socializing: FM (Cunningham et al., 2016) interactions are F>M if mixed-sex interactions (Balliet et al., 2011) with the same sex, F>M if being observed by peers (Charness & Rustichini, 2011) observed by the same sex, repeated, FM (Paola et al., 2018) Anthony & Horne, 2003; Charness & Rustichini, 2011; Vugt et al., 2007). However, collaboration improves with the presence of women in the group (Bear & Woolley, 2011; Sustein & Hastie, 2014) Note: We also include work on personality traits when evidence on socio-emotional skills is lacking. SE skills stands for socio-emotional skills. 56 Appendix Table A2: Sample Selection Criteria Ghana: Skills Mozambique: Togo: Private Congo: Skills Côte Côte D'Ivoire: Support Facebook: Ghana: Impact of Towards Kenya: Skills Towards Impact Côte Sector Benin: Youth development D'Ivoire: project for the Future of Outgrower Contracts Employability Employability and Assessment Nigeria: Togo: Youth Employment and Study D'Ivoire: Development Employment project for Factory agricultural Business on Smallholder and Productivity Productivity (STEP)- Integrated APPEALS Skills Development (AIDE) ProJeunes Project (PADSP)- employability Workers sector(PSAC) (FoB) Farmers(GADCO) (STEP)- Skills Skills Measurement Growth Pole Personal Initiative Measurement Project (IGPP) Sample size 3585 4023 1294 1126 1544 12372 1615 2987 3894 5431 6782 1500 4597 18-40 (men) Age 18-35 17-35 18-70 15-24 no restriction 30-60 No 15-64 15-64 No not restricted 18-40 18+ (women) Rural/Urban Both Urban Urban Urban Rural Both Rural Urban Urban Rural Both Urban Both, though most are urban three regions in the Kpong Irrigation Project 5 departements of 4 geographic areas south (Gbokle, La Me et (KIP) in the Greater Tsangano, the South: 4 towns: (Nairobi, Other Large Sud-Comoe) and seven Accra Region, in Shai Angónia, 5 states: Atlantique, Couffo, Dimbokro , 2 towns: Cities (over 100,000 Geography (e.g. Pointe Noire regions in the center Osudoku District and Macanga, Kaduna, Kano, All country, though heavy Mono, Oueme, Toumodi, Abidjan, 97 countries 10 regions households), Medium Greater Lome area particular states) and Brazzaville (Belier, Goh, Haut- Weta Irrigation Project Chifunde and Kogi, Lagos, concentration in Lome Plateau, 3 Djekanou, Bassam cities (60,000 to Sassandra, Iffou, (WIP) in the Volta Chiúta districts Cross River communes in each M'batto 100,000 households), Marahoue, Moronou et Region, in Ketu North of Tete province departement other urban areas) N’Zi) District Max schooling: completed junior Completed high school (9th Secondary All levels, Minimum CAP (technical primary; but did No Levels of Education grade) or short No restriction No No No No School concentration of training degree) through not complete restriction education vocational training, Complete low levels masters high school not currently in school Yes, to PSAC improved Applicants to a program? Yes, to the YE Yes No Yes No No No No No Yes Yes Yes rubber seedlings Supposed to be of Farmers who are the Agribusiness entrepreneurs unemployed; however, way of Business primary cultivators No formal (Farming, whose businesses defining unemployed is tricky. Employment type? Not a criteria Not a criteria Farmers owners and (PCs) of plots within the No No Farmers employment Processing, are not formally Applicants to an internship employees Kpong and Weta Marketing) registered program with the national irrigation projects employment agency Marital status All All All married / in a couple All No All All No No Both No restrictions Any differences between Yes, for sample Yes, Males male sample and female balance, all women respondents sample at point of fulfilling selection No, but larger only selected selection? (e.g. male criteria were added to sample of women No No No restriction Yes No No because they No No No respondents only the sample and then a for power were husbands selected because they stratified random of selected were husbands of selection was done to females. selected females) add male respondents. Sample was selected based on operating in all the demand of companies for sectors ofactivity youth with specific profiles (for except agriculture, example, if companies They should Interviewed in operation for at requested interns with a CAP in Out of school Farmers with less than have a for two least one year and mechanics but no company Other criteria? for at least a No two hectares of rubber No business page minutes by can prove it at time requested interns with a year cultivated pre-program in facebook program team of application, have Masters in psychology, we less than 50 would have youth with CAP in employees, not mechanics in the sample but registered formally no youth with masters in psychology. English and Language French French/English French French French English English English Portuguese English English French/English Spanish Country Benin Congo Angola,Benin,Botswana,Côte Côte d'Ivoire Côte d'Ivoire d'Ivoire,Cameroon,Ethiopia,Ghana,Kenya,Mozambique,Nigeria,Senegal,Tanzania,Uganda,South Côte d'Ivoire Ghana Ghana Africa,Zambia Kenya Mozambique Nigeria Togo Togo Date 2017 2015 2016 2020 2016 2018 2013 2013 2013 2016 2020 2017 2013 Source GIL GIL GIL GIL GIL WB GIL WB WB GIL GIL GIL GIL Note: WB stands for World Bank while GIL stands for the World Bank's Gender Innovation Lab. 57 Appendix Table A3: Socio-emotional skills categories - Definitions and example items Skill Category Definition Togo: Private Sector Development Project (PADSP)- Personal Kenya: Skills Towards Employability and Productivity (STEP)- Initiative Skills Measurement Positive Self Concept Intrapersonal The ability to identify and interpret one’s own thoughts and It would be easy for me to find another job Do you work very well and quickly? (Originally Self- behaviors and to evaluate one’s strengths and weakness and Awareness) knowing your preferences, values and biases. Emotional Regulation Intrapersonal The ability to maintain or change one’s own emotions by I get frequent mood swings (My mood changes quickly) Are you relaxed during stressful situations? controlling one’s thoughts and behavioral responses. Perseverance Intrapersonal The ability to sustain effort despite setbacks. I don’t lose sight of my goal, even if I make mistakes Do you finish whatever you begin? Personal Initiative Intrapersonal The ability to develop long-term goals, to seek opportunities I take the initiative immediately even when others don’t to improve one’s self and to be motivated to put these plans and goals into action. PSDM Intrapersonal The ability to approach a problem by gathering As soon as a problem arises, I look for an immediate Do you think about how the things you will do affect you in information, generating a number of solutions and solution the future? evaluating the consequences of these solutions before acting. Self Control Intrapersonal The ability to focus one’s attention, stay on task, break habits, I do my work without delay restrain impulses and keep good self-discipline. Empathy Interpersonal The ability to understand another’s viewpoint or thoughts and I sense the feelings of others Do you think about how the things you will do will affect have emotional concern for another’s situation or experience. others? Expressiveness Interpersonal The ability to explain ideas in a way that others will I tend to hold back Do you ask for help when you don’t understand something? understand and openly express one’s opinion. Interpersonal Interpersonal The ability to take actions intended to build trust and benefit I amuse people at parties Are you outgoing and sociable, for example, do you make Relatedness others, initiate and maintain relationships and be respectful, friends very easily? encouraging and caring towards others. Teamwork (Originally Interpersonal The ability to take other’s perspective, listen and Example from Mozambique: Impact Assessment Integrated Growth Pole Project (IGPP) : Collaboration) communicate in groups of two or more people, identify In a job, I always try to do my work alone situations involving group problem-solving and decision- making, and organizing and coordinate team members to create shared plans and goals. Note: Categorization was based in the definitions of socio-emotional skills above. Some items were categorized when they aligned with socio-emotional skills definitions but were not precisely included in the definition. For example, "Are you relaxed during stressful situations" was categorized under "Emotional regulation" even though an individual could still exhibit strong emotional regulation skills if they were not relaxed, but took steps to become relaxed. Similarly, "Positive self concept" mostly included questions on self esteem and generalized self efficacy. However, sometimes domain-specific self-efficacy questions were also included if they were key to the population of the study. The framework used for this study also included Emotional Awareness, Listening, Interpersonal Influence, and Negotiation. However, sufficient items were not found that fell into these categories. The framework also used Self-Awareness and Collaboration rather than Positive Self Concept and Teamwork, respectively. However, the items found in each study were more specifically focused on the latter concepts. Unlike Self-Awareness, Positive Self Concept incorporated items that measure self-esteem, which is sometimes considered a belief rather than a skill. 58 Table A4: Correlations between socio-emotional skills and positive earnings Panel A: Aggregate Skills Earnings All Intra Inter (1) (2) (3) Woman -0.315*** -0.315*** -0.391*** (0.042) (0.042) (0.042) SE skills 0.084*** 0.082*** 0.031* (0.013) (0.013) (0.016) Woman*SE skills -0.009 -0.006 0.014 (0.019) (0.019) (0.022) Education attainment 0.038*** 0.038*** 0.035*** (0.004) (0.004) (0.004) Woman*Education attainment -0.006 -0.007 -0.004 (0.004) (0.004) (0.004) Observations 23,387 23,383 18,394 R-squared 0.445 0.445 0.488 P-val SE skills+Woman*SE skills=0 0.000 0.000 0.004 P-val Woman + Woman*SE skills=0 0.000 0.000 0.000 Mean monthly earnings for Men 232.3 232.3 257.4 Mean monthly earnings for Women 160.2 159.6 167.2 Panel B: Intrapersonal Skills Earnings Positive Self Emotional Personal Perseverance PSDM Self Control Concept Regulation Initiative (1) (2) (3) (4) (5) (6) Woman -0.304*** -0.076 -0.353*** -0.441*** -0.335*** -0.470*** (0.043) (0.069) (0.041) (0.053) (0.043) (0.051) SE skills 0.084*** 0.035** 0.050*** 0.038* 0.091*** 0.036** (0.015) (0.017) (0.014) (0.021) (0.015) (0.017) Woman*SE skills -0.042* -0.009 0.035* 0.027 -0.025 -0.015 (0.021) (0.024) (0.019) (0.031) (0.021) (0.027) Education attainment 0.043*** 0.054*** 0.039*** 0.032*** 0.037*** 0.037*** (0.004) (0.006) (0.004) (0.006) (0.004) (0.006) Woman*Education attainment -0.010** -0.026*** -0.006 -0.002 -0.002 -0.005 (0.005) (0.006) (0.004) (0.006) (0.004) (0.006) Observations 17,553 11,040 22,110 9,706 16,839 10,038 R-squared 0.486 0.441 0.472 0.367 0.550 0.440 P-val SE skills+Woman*SE skills=0 0.007 0.118 0.000 0.005 0.000 0.311 P-val Woman + Woman*SE skills=0 0.000 0.243 0.000 0.000 0.000 0.000 Mean monthly earnings for Men 260.5 380 222.4 122.2 289.1 73.87 Mean monthly earnings for Women 190.7 231 154.6 74.92 175.7 50.71 Panel C: Interpersonal Skills Earnings Interpersonal Empathy Expressiveness Teamwork Relatedness (1) (2) (3) (4) Woman 0.006 -0.066 -0.241*** -0.676*** (0.102) (0.069) (0.050) (0.051) SE skills 0.041* 0.023 0.021 0.009 (0.022) (0.017) (0.016) (0.028) Woman*SE skills 0.028 -0.026 0.016 0.063 (0.036) (0.024) (0.022) (0.040) Education attainment 0.068*** 0.055*** 0.047*** 0.020*** (0.007) (0.006) (0.005) (0.007) Woman*Education attainment -0.031*** -0.031*** -0.013*** 0.013* (0.010) (0.006) (0.005) (0.007) Observations 5,382 13,470 13,320 7,975 R-squared 0.420 0.341 0.622 0.411 P-val SE skills+Woman*SE skills=0 0.016 0.847 0.015 0.013 P-val Woman + Woman*SE skills=0 0.759 0.213 0.000 0.000 Mean monthly earnings for Men 601.9 329.9 310.1 103.3 Mean monthly earnings for Women 452.2 221.6 195.6 52.72 Note: OLS regression specifications include study fixed effects. All studies have equal weights. Women is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Earnings is the inverse hyperbolic sine (IHS) transformation of the respondent's monthly earnings in US dollars, conditional on earnings being strictly positive. Note that only business profit rather than total earnings is reported for Nigeria. Education, age bins and marital status are added as controls. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Age bin represents dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, o otherwise. Married is added as a control and defined as a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 59 Appendix Table A5: Cohen's D Positive Self Emotional Personal Interpersonal All Intra Inter Perseverance PSDM Self Control Empathy Expressiveness Teamwork Sub-sample Concept Regulation Initiative Relatedness All 0.130 0.115 0.123 0.062 0.066 0.098 0.098 0.124 -0.008 0.076 0.105 0.087 0.141 Pooled sample Old 0.209 0.200 0.147 0.095 0.219 0.159 0.123 0.214 0.025 0.072 0.125 0.131 0.100 Young 0.106 0.088 0.114 0.047 0.034 0.079 0.093 0.100 -0.027 0.078 0.099 0.076 0.150 Didn't CP 0.132 0.135 0.087 0.029 0.132 0.130 0.180 0.128 -0.029 0.124 -0.122 0.086 0.015 CP 0.122 0.103 0.125 0.049 0.063 0.085 0.086 0.117 -0.013 0.072 0.109 0.087 0.158 Benin: Youth Employment All 0.345 0.296 0.267 0.088 . 0.292 0.195 0.217 0.116 . . . 0.267 Old 0.313 0.283 0.197 0.012 . 0.338 0.167 0.212 0.126 . . . 0.197 Young 0.357 0.297 0.307 0.132 . 0.254 0.211 0.215 0.108 . . . 0.307 Didn't CP 0.270 0.234 0.212 0.111 . 0.240 0.184 0.134 0.068 . . . 0.212 CP 0.396 0.322 0.333 0.043 . 0.307 0.168 0.293 0.151 . . . 0.333 Congo: Skills Development All 0.157 0.152 0.079 0.100 . 0.196 . . 0.010 . . . 0.079 Project for Employability Old 0.171 0.167 0.069 0.120 . 0.195 . . 0.017 . . . 0.069 Young 0.146 0.143 0.085 0.075 . 0.212 . . 0.002 . . . 0.085 Didn't CP -0.165 -0.059 . -0.983 . 1.418 . . 0.150 . . . . CP 0.157 0.153 0.083 0.103 . 0.194 . . 0.010 . . . 0.083 Côte d'Ivoire: Factory Workers All 0.371 0.383 0.107 0.258 0.179 0.267 . 0.241 . . 0.060 0.101 . Old 0.714 0.633 0.336 0.504 0.221 0.324 . 0.455 . . 0.191 0.304 . Young 0.239 0.278 0.023 0.154 0.154 0.235 . 0.149 . . 0.010 0.026 . Didn't CP 0.348 0.378 0.064 0.336 0.193 0.123 . 0.255 . . 0.135 -0.099 . CP 0.309 0.304 0.119 0.216 0.123 0.314 . 0.139 . . 0.048 0.147 . Côte d'Ivoire: ProJeunes All 0.176 0.185 0.137 . 0.102 0.081 0.171 0.070 0.202 0.089 0.123 0.109 0.123 Old 0.137 0.136 0.116 . 0.145 0.041 0.061 0.073 0.150 0.050 0.144 0.063 0.124 Young 0.212 0.231 0.154 . 0.068 0.112 0.260 0.069 0.260 0.122 0.107 0.148 0.121 Didn't CP 0.034 0.026 0.035 . -0.043 0.228 0.238 0.043 -0.254 0.015 0.168 0.120 -0.180 CP 0.169 0.181 0.128 . 0.117 0.058 0.164 0.084 0.202 0.088 0.114 0.085 0.130 Côte D'Ivoire: Support Project All 0.044 0.044 . -0.064 . . . 0.116 . . . . . for the Agricultural Sector Old . . . . . . . . . . . . . (PSAC) Young 0.044 0.044 . -0.064 . . . 0.116 . . . . . Didn't CP -0.063 -0.063 . -0.085 . . . -0.030 . . . . . CP 0.094 0.094 . -0.039 . . . 0.168 . . . . . Facebook: Future of Business All 0.034 0.012 0.072 . -0.070 0.042 0.052 -0.025 . . . 0.072 . (FoB) Old 0.247 0.220 0.487 . 0.219 0.256 0.587 0.118 . . . 0.487 . Young 0.032 0.010 0.067 . -0.074 0.040 0.047 -0.025 . . . 0.067 . Didn't CP 0.277 0.194 0.202 . 0.042 0.484 -0.063 0.159 . . . 0.202 . CP 0.028 0.008 0.069 . -0.072 0.035 0.050 -0.028 . . . 0.069 . Ghana: Impact of Outgrower All 0.138 0.138 . . . 0.092 0.145 . . . . . . Contracts on Smallholder Old -0.287 -0.287 . . . -0.543 0.053 . . . . . . Farmers (GADCO) Young 0.135 0.135 . . . 0.088 0.143 . . . . . . Didn't CP 0.124 0.124 . . . 0.155 0.028 . . . . . . CP 0.139 0.139 . . . 0.081 0.162 . . . . . . Ghana: Skills Towards All 0.253 0.270 0.145 0.074 0.290 0.115 . 0.314 . 0.126 0.073 0.107 . Employability and Productivity Old 0.210 0.237 0.093 0.151 0.191 0.027 . 0.281 . 0.068 0.049 0.068 . (STEP)- Skills Measurement Young 0.269 0.275 0.176 0.010 0.337 0.155 . 0.328 . 0.155 0.092 0.135 . Didn't CP 0.389 0.193 0.454 0.080 0.340 -0.042 . 0.215 . 0.547 -0.518 0.398 . CP 0.248 0.273 0.136 0.074 0.290 0.119 . 0.318 . 0.117 0.088 0.098 . Kenya: Skills Towards All 0.173 0.184 0.094 0.094 0.173 0.037 . 0.208 . 0.053 0.111 0.048 . Employability and Productivity Old 0.207 0.210 0.129 0.077 0.238 0.025 . 0.261 . 0.067 0.137 0.087 . (STEP)- Skills Measurement Young 0.149 0.166 0.070 0.102 0.133 0.041 . 0.173 . 0.041 0.094 0.021 . Didn't CP -0.045 0.046 -0.146 -0.053 0.072 0.131 . -0.001 . -0.060 -0.198 -0.084 . CP 0.190 0.195 0.114 0.108 0.181 0.026 . 0.225 . 0.064 0.136 0.058 . Mozambique: Impact All 0.110 0.097 0.091 0.013 . 0.124 0.132 0.104 -0.085 . . 0.102 0.019 Assessment Integrated Growth Old 0.048 0.042 0.041 -0.014 . 0.081 0.078 0.099 -0.112 . . 0.063 -0.017 Pole Project (IGPP) Young 0.122 0.107 0.100 0.020 . 0.129 0.138 0.107 -0.077 . . 0.108 0.029 Didn't CP 0.115 0.098 0.101 -0.001 . 0.128 0.153 0.120 -0.109 . . 0.125 0.006 CP 0.041 0.033 0.039 -0.018 . 0.052 0.059 0.033 -0.026 . . 0.045 0.007 Nigeria: APPEALS All 0.203 0.207 0.162 . 0.179 0.137 . 0.232 . . 0.157 0.132 . Old 0.280 0.262 0.257 . 0.253 0.194 . 0.248 . . 0.151 0.307 . Young 0.192 0.199 0.149 . 0.168 0.128 . 0.229 . . 0.157 0.108 . Didn't CP . . . . . CP 0.202 0.206 0.162 . 0.181 0.135 . 0.231 . . 0.156 0.132 . Togo: Private Sector All 0.124 0.155 -0.007 0.216 0.075 0.127 0.216 0.354 -0.170 0.056 -0.090 0.018 . Development Project (PADSP)- Old 0.210 -0.075 0.731 -0.331 0.063 0.154 0.021 0.108 -0.323 0.432 0.196 0.685 . Personal Initiative Young 0.122 0.160 -0.020 0.227 0.075 0.127 0.221 0.360 -0.168 0.047 -0.095 0.005 . Didn't CP -0.083 -0.034 -0.137 0.033 -0.134 0.011 0.020 0.312 -0.143 -0.037 -0.047 -0.208 . CP 0.060 0.081 -0.013 0.136 0.082 0.087 0.146 0.297 -0.205 0.019 -0.077 0.038 . Togo: Youth Employment and All 0.118 0.023 0.219 0.018 . -0.003 0.057 . . . 0.114 . 0.223 Skills Development (AIDE) Old -0.434 -0.595 0.022 -0.352 . -0.577 -0.309 . . . -0.025 . 0.060 Young 0.129 0.035 0.222 0.026 . 0.008 0.063 . . . 0.117 . 0.225 Didn't CP . . . . . . . . . . . . . CP 0.118 0.023 0.219 0.018 . -0.003 0.057 . . . 0.114 . 0.223 Note: Cohen's d is used to indicate the standardised difference between two means. Old is a dummy Variable equal to 1 if the respondent is 25 years old or older, 0 otherwise. Young is a dummy Variable equal to 1 if the respondent is 24 years old or young, 0 otherwise. CP stands for completed primary school and is define as a dummy variable equal to 1 if the respondent completed primary school, 0 otherwise. Didn't CP is defined as a dummy variable equal to 1 if the respondent didn't complete primary school, 0 otherwise. The blue color (light blue=small effect size and dark blue=medium effect size) indicates positive effectsizes while orage (light orange=small effect size and dark orange=medium or large effect sizes) represents negative values. 60 Table A6: Gender differences in levels of socio-emotional skills Coef on Women Coef on Women Coef on Women Coef on PSC (Atl 3) (PSC) (1) (2) (3) (4) All coef. -0.151*** -0.106*** -0.115*** 0.578*** se (0.013) (0.010) (0.012) (0.007) obs 41,873 40,761 25,551 25,551 R-squared 0.021 0.020 0.357 0.357 Intra coef. -0.143*** -0.123*** -0.102*** 0.640*** se (0.013) (0.010) (0.011) (0.006) obs 41,834 40,752 25,551 25,551 R-squared 0.021 0.023 0.423 0.423 Inter coef. -0.104*** -0.078*** -0.085*** 0.137*** se (0.016) (0.016) (0.019) (0.010) obs 33,658 18,868 20,433 20,433 R-squared 0.011 0.015 0.036 0.036 Positive Self coef. Concept -0.060*** -0.058*** se (0.015) (0.017) obs 25,551 19,863 R-squared 0.014 0.015 Emotional coef. Regulation -0.130*** -0.175*** -0.146*** 0.051*** se (0.018) (0.022) (0.024) (0.012) obs 22,573 12,758 8,433 8,433 R-squared 0.013 0.017 0.020 0.020 Perseverance coef. -0.099*** -0.142*** -0.079*** 0.276*** se (0.013) (0.018) (0.014) (0.008) obs 39,885 20,577 24,004 24,004 R-squared 0.012 0.015 0.094 0.094 Personal Initiative coef. -0.124*** -0.122*** -0.087*** 0.279*** se (0.018) (0.022) (0.019) (0.011) obs 22,052 12,675 13,047 13,047 R-squared 0.015 0.021 0.105 0.105 PSDM coef. -0.160*** -0.157*** -0.171*** 0.218*** se (0.015) (0.017) (0.017) (0.009) obs 31,959 26,185 18,220 18,220 R-squared 0.021 0.026 0.077 0.077 Self Control coef. 0.027 -0.009 0.060*** 0.102*** se (0.021) (0.024) (0.022) (0.010) obs 14,835 10,403 13,709 13,709 R-squared 0.015 0.025 0.016 0.016 Empathy coef. -0.051** -0.064 -0.035 0.135*** se (0.026) (0.065) (0.026) (0.014) obs 8,260 1,126 7,108 7,108 R-squared 0.014 0.023 0.033 0.033 Expressiveness coef. -0.070*** -0.104*** -0.034 0.109*** se (0.018) (0.019) (0.021) (0.011) obs 18,866 16,109 11,767 11,767 R-squared 0.007 0.011 0.021 0.021 Interpersonal coef. Relatedness -0.062*** -0.051** -0.028 0.154*** se (0.017) (0.020) (0.020) (0.010) obs 26,941 15,523 13,737 13,737 R-squared 0.010 0.015 0.037 0.037 Teamwork coef. -0.119*** -0.157*** -0.122*** 0.099*** se (0.027) (0.039) (0.029) (0.014) obs 13,115 4,450 11,989 11,989 R-squared 0.012 0.018 0.027 0.027 Note: OLS regression specifications include study fixed effects. All studies have equal weights. Women is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Age bins, marital status and education dummies are added as controls in all regressions. Coulmn 3 has PSC (Positive Self Concept) as additional control variable. Age bin represents dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, 0 otherwise. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Education dummies represent dummy variables equal to 1 if the respondent's highest educational attainment (completed) is 0, 1, ... or 14 where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed high school, 13=completed certificate or diploma and 14=completed university degree or above, 0 otherwise. Note that Nigeria & Facebook projects have only categorical variable for education and "completed primary" is coded as 9 and "completed secondary" is coded as 12. PSDM stands for Problem Solving and Decision Making. 'Atl 3' stands for atleast 3 and it represents socio-emotional skills captured by at least three items in an individual survey. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 61 Appendix Table A7: Gender differences in levels of socio-emotional skills - At least 3 Panel A: Aggregate Skills All Intra Inter (1) (2) (3) Women -0.051** -0.072*** 0.176*** (0.022) (0.023) (0.053) Education attainment 0.023*** 0.024*** 0.035*** (0.002) (0.002) (0.004) Women*Education attainment -0.006*** -0.005** -0.024*** (0.002) (0.002) (0.005) Observations 40,761 40,752 18,868 R-squared 0.018 0.021 0.013 P-val Women+Women*Edu Attain 0.005 0.000 0.002 P-val Edu Attain+Women*Edu Attain 0.000 0.000 0.000 Mean SE skills for Men 0.083 0.091 0.119 Mean SE skills for Women 0.000 0.000 0.000 Panel B: Intrapersonal Skills Positive Self Emotional Perseverance Personal Initiative PSDM Self Control Concept Regulation (1) (2) (3) (4) (5) (6) Women -0.020 0.028 -0.101*** -0.084** -0.096*** 0.156*** (0.030) (0.078) (0.031) (0.039) (0.035) (0.040) Education attainment 0.024*** 0.021*** 0.020*** 0.027*** 0.033*** 0.046*** (0.003) (0.006) (0.004) (0.004) (0.004) (0.005) Women*Education attainment -0.005 -0.019*** -0.005 -0.003 -0.006* -0.019*** (0.003) (0.007) (0.003) (0.004) (0.003) (0.005) Observations 19,863 12,758 20,577 12,675 26,185 10,403 R-squared 0.013 0.016 0.015 0.016 0.022 0.020 P-val Women+Women*Edu Attain 0.381 0.897 0.000 0.014 0.001 0.000 P-val Edu Attain+Women*Edu Attain 0.000 0.571 0.000 0.000 0.000 0.000 Mean SE skills for Men 0.0550 0.190 0.167 0.121 0.119 -0.010 Mean SE skills for Women -0.001 0.000 0.000 -0.001 0.000 0.000 Panel C: Interpersonal Skills Empathy Expressiveness Interpersonal Teamwork Relatedness (1) (2) (3) (4) Women 0.268 0.165** 0.172*** 0.144 (0.179) (0.077) (0.059) (0.159) Education attainment 0.042*** 0.037*** 0.037*** 0.039*** (0.014) (0.005) (0.005) (0.011) Women*Education attainment -0.031* -0.024*** -0.022*** -0.025** (0.016) (0.006) (0.005) (0.013) Observations 1,126 16,109 15,523 4,450 R-squared 0.014 0.010 0.011 0.015 P-val Women+Women*Edu Attain 0.149 0.045 0.006 0.422 P-val Edu Attain+Women*Edu Attain 0.253 0.004 0.000 0.108 Mean SE skills for Men 0.093 0.125 0.094 0.198 Mean SE skills for Women 0.000 0.000 0.000 0.001 Note: OLS regression specifications control for age and include study fixed effects. The sample is restricted to socio-emotional skills captured by at least three items in an individual survey. All studies have equal weights. Women is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Marital status and age bins are added as controls. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Age bins represent dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, o otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 62 Appendix Table A8 (At least 3) : Correlations between socio-emotional skills and earnings Panel A: Aggregate Skills Earnings All Intra Inter (1) (2) (3) Women -0.553*** -0.556*** -0.816*** (0.060) (0.060) (0.124) SE skills 0.108*** 0.123*** -0.055* (0.029) (0.026) (0.032) Women*SE skills 0.027 -0.013 0.089** (0.041) (0.038) (0.045) Education attainment 0.024*** 0.024*** 0.015 (0.006) (0.006) (0.010) Women*Education attainment -0.009 -0.009 -0.003 (0.006) (0.006) (0.011) Observations 33,896 33,887 18,155 R-squared 0.241 0.241 0.162 P-val SE skills+Women*SE skills=0 0.000 0.000 0.279 P-val Women + Women*SE skills=0 0.000 0.000 0.000 Mean Monthly Earnings for Men 169.9 169.8 257.8 Mean Monthly Earnings for Women 103.3 102.9 155.8 Panel B: Intrapersonal Skills Earnings Positive Self Emotional Perseverance Personal Initiative PSDM Self Control Concept Regulation (1) (2) (3) (4) (5) (6) Women -0.494*** -1.590*** -0.853*** -0.350*** -0.473*** -0.651*** (0.059) (0.218) (0.060) (0.085) (0.077) (0.086) SE skills 0.122*** 0.013 0.077*** 0.017 0.007 -0.003 (0.022) (0.040) (0.028) (0.035) (0.030) (0.039) Women*SE skills -0.087*** -0.013 0.045 0.051 0.070* -0.048 (0.032) (0.058) (0.036) (0.050) (0.041) (0.055) Education attainment 0.022*** -0.024 -0.022*** 0.018* 0.036*** -0.015 (0.007) (0.015) (0.008) (0.010) (0.009) (0.014) Women*Education attainment -0.011 0.055*** 0.023*** -0.026*** -0.029*** -0.012 (0.007) (0.019) (0.007) (0.010) (0.008) (0.012) Observations 19,464 12,439 20,571 12,282 19,707 10,397 R-squared 0.322 0.214 0.268 0.291 0.344 0.298 P-val SE skills+Women*SE skills=0 0.122 0.992 0.000 0.059 0.007 0.195 P-val Women + Women*SE skills=0 0.000 0.000 0.000 0.002 0.000 0.000 Mean Monthly Earnings for Men 99.15 315 53.53 90.41 214.2 40.38 Mean Monthly Earnings for Women 46.89 180.8 31.63 46.09 123.6 15.57 Panel C: Interpersonal Skills Earnings Empathy Expressiveness Interpersonal Teamwork Relatedness (1) (2) (3) (4) Women -1.219*** -1.479*** -0.854*** -0.968** (0.459) (0.209) (0.128) (0.387) SE skills -0.141 -0.029 -0.063** -0.090 (0.100) (0.034) (0.031) (0.061) Women*SE skills 0.209 0.103** 0.049 0.271*** (0.148) (0.051) (0.044) (0.096) Education attainment -0.054 -0.016 0.013 -0.034 (0.034) (0.014) (0.011) (0.026) Women*Education attainment 0.009 0.042** 0.004 -0.012 (0.042) (0.017) (0.012) (0.031) Observations 1,126 15,409 15,191 4,069 R-squared 0.061 0.185 0.182 0.086 P-val SE skills+Women*SE skills=0 0.530 0.048 0.650 0.014 P-val Women + Women*SE skills=0 0.041 0.000 0.000 0.084 Mean Monthly Earnings for Men 103.7 273.2 288.5 139 Mean Monthly Earnings for Women 43.02 171.3 161.4 74.16 Note: OLS regression specifications include study fixed effects. The sample is restricted to socio-emotional skills captured by at least three items in an individual survey. All studies have equal weights. Women is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Earnings is the inverse hyperbolic sine (IHS) transformation of the respondent's monthly earnings in US dollars. Note that only business profit rather than total earnings is reported for Nigeria. Education dummies, age bins and marital status are added as controls. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Age bin represents dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, o otherwise. Married is added as a control and defined as a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 63 Appendix Table A9: Gender differences in levels of socio-emotional skills - Positive Self-Concept Panel A: Aggregate Skills All- If Pos Self All- Control for Intra- If Pos Intra- Control Inter- If Pos Inter- Control Conc Pos Self Conc Self Conc for Pos Self Self Conc for Pos Self nonmissing nonmissing Conc nonmissing Conc (1) (2) (3) (4) (5) (6) Women -0.094*** -0.082*** -0.105*** -0.091*** -0.014 -0.010 (0.029) (0.023) (0.028) (0.022) (0.030) (0.030) Positive Self Concept 0.579*** 0.640*** 0.137*** (0.007) (0.006) (0.010) Education attainment 0.034*** 0.020*** 0.032*** 0.017*** 0.022*** 0.018*** (0.003) (0.002) (0.003) (0.002) (0.003) (0.003) Women*Education attainment -0.007** -0.004 -0.004 -0.001 -0.009** -0.008** (0.003) (0.002) (0.003) (0.002) (0.004) (0.003) Observations 25,551 25,551 25,551 25,551 20,433 20,433 R-squared 0.025 0.356 0.025 0.423 0.010 0.031 P-val Women+Women*Edu Attain 0.000 0.000 0.502 P-val Edu Attain+Women*Edu Attain 0.000 0.000 0.001 Mean SE skills for Men 0.161 0.161 0.143 0.143 0.135 0.135 Mean SE skills for Women 0.000 0.000 0.001 0.001 0.000 0.000 Panel B: Intrapersonal Skills Emotional Emotional Perseverance- Perseverance- Personal Personal PSDM- If Pos PSDM- Control Self Control- If Self Control- Regulation- If Regulation- If Pos Self Control for Initiative- If Initiative- Self Conc for Pos Self Pos Self Conc Control for Pos Self Conc Control for Conc Pos Self Conc Pos Self Conc Control for nonmissing Conc nonmissing Pos Self Conc nonmissing Pos Self Conc nonmissing nonmissing Pos Self Conc (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Women 0.048 0.052 -0.094*** -0.088*** -0.060* -0.060** -0.118*** -0.112*** 0.041 0.042 (0.060) (0.060) (0.029) (0.027) (0.031) (0.029) (0.028) (0.027) (0.034) (0.034) Positive Self Concept 0.050*** 0.278*** 0.282*** 0.218*** 0.102*** (0.012) (0.008) (0.011) (0.009) (0.010) Education attainment 0.025*** 0.024*** 0.024*** 0.017*** 0.037*** 0.027*** 0.028*** 0.023*** 0.006 0.003 (0.005) (0.005) (0.003) (0.003) (0.004) (0.004) (0.003) (0.003) (0.005) (0.005) Women*Education attainment -0.021*** -0.021*** -0.001 0.001 -0.003 -0.002 -0.009*** -0.008** 0.002 0.003 (0.006) (0.006) (0.003) (0.003) (0.004) (0.003) (0.003) (0.003) (0.005) (0.005) Observations 8,433 8,433 24,004 24,004 13,047 13,047 18,220 18,220 13,709 13,709 R-squared 0.017 0.020 0.016 0.092 0.021 0.101 0.026 0.074 0.005 0.015 P-val Women+Women*Edu Attain 0.569 0.000 0.021 0.000 0.141 P-val Edu Attain+Women*Edu Attain 0.461 0.000 0.000 0.000 0.171 Mean SE skills for Men 0.183 0.183 0.119 0.119 0.123 0.123 0.192 0.192 -0.026 -0.026 Mean SE skills for Women 0.003 0.003 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.000 Panel C: Interpersonal Skills Empathy- If Empathy- Expressiveness- Expressiveness- Interpers. Interpers. Teamwork- If Teamwork- Pos Self Conc Control for If Pos Self Control for Relatedness- If Relatedness- Pos Self Conc Control for nonmissing Pos Self Conc Conc Pos Self Conc Pos Self Conc Control for nonmissing Pos Self Conc nonmissing nonmissing Pos Self Conc (1) (2) (3) (4) (5) (6) (7) (8) Women 0.040 0.055 0.170*** 0.177*** 0.029 0.031 -0.058* -0.058* (0.078) (0.077) (0.061) (0.061) (0.037) (0.036) (0.033) (0.033) Positive Self Concept 0.133*** 0.110*** 0.154*** 0.097*** (0.014) (0.011) (0.010) (0.014) Education attainment 0.028*** 0.023*** 0.025*** 0.022*** 0.023*** 0.019*** 0.013** 0.010* (0.006) (0.006) (0.005) (0.005) (0.004) (0.004) (0.006) (0.006) Women*Education attainment -0.009 -0.010 -0.021*** -0.021*** -0.008* -0.007* -0.008* -0.008* (0.007) (0.007) (0.005) (0.005) (0.004) (0.004) (0.005) (0.005) Observations 7,108 7,108 11,767 11,767 13,737 13,737 11,989 11,989 R-squared 0.012 0.028 0.006 0.018 0.007 0.030 0.011 0.024 P-val Women+Women*Edu Attain 0.513 0.005 0.467 0.026 P-val Edu Attain+Women*Edu Attain 0.004 0.611 0.000 0.604 Mean SE skills for Men 0.074 0.074 0.079 0.079 0.076 0.076 0.139 0.139 Mean SE skills for Women 0.002 0.002 -0.001 -0.001 0.001 0.001 0.000 0.000 Note: OLS regression specifications include study fixed effects and control for positive self-concept. All studies have equal weights. Women is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Marital status and age bins are added as controls. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Age bins represent dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, 0 otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 64 Appendix Table A10 : Correlations between socio-emotional skills and earnings - Positive Self-Concept Panel A: Aggregate Skills All- 2SRI Intra- 2SRI Inter- 2SRI (1) (2) (3) Women -0.487*** -0.485*** -0.521*** (0.059) (0.059) (0.058) SE skills residual 0.049* 0.073*** -0.027 (0.027) (0.028) (0.030) Women*SE skills residual 0.005 -0.016 0.065* (0.038) (0.039) (0.039) Education attainment 0.047*** 0.047*** 0.045*** (0.006) (0.006) (0.007) Women*Education attainment -0.027*** -0.027*** -0.021*** (0.007) (0.007) (0.007) Observations 24,835 24,835 19,717 R-squared 0.283 0.283 0.319 P-val SE skills+Female*SE skills 0.044 0.041 0.117 P-val Female+Female*SE skills 0.000 0.000 0.000 Mean monthly earnings for Men 202.000 202.000 229.400 Mean monthly earnings for Women 120.300 120.300 126.200 Panel B: Intrapersonal Skills Emotional Regulation- Perseverance- Personal Initiative- Problem Solving & Self Control- 2SRI 2SRI 2SRI DecMaking- 2SRI 2SRI (1) (2) (3) (4) (5) Women -0.601*** -0.525*** -0.465*** -0.467*** -0.596*** (0.130) (0.055) (0.053) (0.061) (0.053) SE skills residual 0.038 0.072*** 0.009 0.020 0.045** (0.039) (0.024) (0.029) (0.030) (0.022) SE skills residual*Women -0.064 0.006 0.057 0.020 -0.006 (0.054) (0.033) (0.038) (0.040) (0.031) Education attainment 0.047*** 0.044*** 0.021*** 0.051*** -0.004 (0.011) (0.006) (0.008) (0.007) (0.007) Women*Education attainment -0.023* -0.028*** -0.030*** -0.027*** 0.006 (0.014) (0.007) (0.008) (0.008) (0.007) Observations 8,104 23,288 12,654 17,891 13,691 R-squared 0.194 0.307 0.425 0.332 0.465 P-val SE skills+Female*SE skills 0.472 0.001 0.008 0.115 0.073 P-val Female+Female*SE skills 0.000 0.000 0.000 0.000 0.000 Mean monthly earnings for Men 473.100 193.700 80.480 260.800 50.850 Mean monthly earnings for Women 259.100 116.300 37.470 134.300 30.090 Panel C: Interpersonal Skills Empathy- 2SRI Expressiveness- Interpers. Teamwork- 2SRI 2SRI Relatedness- 2SRI (1) (2) (3) (4) Women -0.609*** -0.589*** -0.356*** -0.601*** (0.168) (0.123) (0.068) (0.061) SE skills residual -0.030 -0.003 -0.096*** 0.027 (0.039) (0.032) (0.030) (0.045) SE skills residual*Women 0.051 0.014 0.058 0.067 (0.060) (0.047) (0.042) (0.053) Education attainment 0.051*** 0.048*** 0.067*** 0.008 (0.013) (0.010) (0.008) (0.011) Women*Education attainment -0.027 -0.028** -0.044*** -0.005 (0.017) (0.012) (0.009) (0.009) Observations 6,780 11,057 13,408 11,602 R-squared 0.278 0.167 0.394 0.403 P-val SE skills+Female*SE skills 0.636 0.755 0.195 0.001 P-val Female+Female*SE skills 0.002 0.000 0.000 0.000 Mean monthly earnings for Men 512.600 358.600 294.200 70.850 Mean monthly earnings for Women 312.700 230.300 154.400 27.170 Note: OLS regression specifications include study fixed effects and control for positive self-concept. All studies have equal weights. Earnings is the inverse hyperbolic sine (IHS) transformation of the respondent's monthly earnings in US dollars. Women is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Age is a continuous variable for the respondent's age. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Marital status and age bins are added as controls. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Age bins represent dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, 0 otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 65 Appendix Table A11: Gender differences in levels of socio-emotional skills-Heterogeneity by transition years in educational attainment Positive Self Emotional Personal Interpersona All Intra Inter Perseverance PSDM Self Control Empathy Expressiveness Teamwork Concept Regulation Initiative l Relatedness (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) Women -0.140*** -0.145*** -0.042 -0.064* -0.060 -0.122*** -0.139*** -0.134*** 0.066* -0.026 0.109 -0.004 -0.088** (0.029) (0.029) (0.032) (0.033) (0.060) (0.030) (0.032) (0.029) (0.038) (0.095) (0.076) (0.037) (0.043) Education attainment 0.028*** 0.028*** 0.014*** 0.008 0.019*** 0.022*** 0.021*** 0.021*** 0.012 0.014 0.001 0.018*** -0.010 (0.004) (0.004) (0.004) (0.006) (0.005) (0.004) (0.005) (0.004) (0.009) (0.012) (0.011) (0.004) (0.012) Women*Education attainment -0.002 0.001 -0.006 0.009 -0.000 0.002 -0.003 -0.003 0.006 0.000 -0.014 -0.010* 0.024* (0.005) (0.005) (0.005) (0.008) (0.007) (0.005) (0.006) (0.005) (0.012) (0.015) (0.013) (0.006) (0.014) Ever entered lower secondary -0.094*** -0.087*** -0.018 0.061 -0.084* -0.060** -0.116*** -0.050 0.065 -0.047 0.082 -0.081** 0.236*** (0.029) (0.029) (0.035) (0.045) (0.043) (0.030) (0.039) (0.034) (0.070) (0.071) (0.063) (0.041) (0.083) Ever entered senior secondary 0.090*** 0.086*** 0.045 0.124*** 0.077 0.046 0.128*** 0.084** 0.016 0.084 0.114** 0.087* 0.066 (0.028) (0.028) (0.041) (0.032) (0.047) (0.029) (0.043) (0.037) (0.047) (0.063) (0.054) (0.046) (0.075) Ever entered higher education -0.003 -0.042 0.084** 0.006 -0.067* -0.101*** -0.031 -0.006 0.252*** 0.159*** 0.067* 0.062 0.116** (0.027) (0.026) (0.033) (0.031) (0.039) (0.027) (0.038) (0.036) (0.068) (0.055) (0.037) (0.039) (0.055) Women*Ever entered lower secondary 0.038 0.031 -0.041 -0.065 -0.020 0.027 0.111* 0.012 -0.157* 0.026 -0.033 0.036 -0.335*** (0.043) (0.043) (0.051) (0.060) (0.060) (0.045) (0.059) (0.048) (0.092) (0.090) (0.081) (0.056) (0.110) Women*Ever entered senior secondary -0.041 -0.072* 0.086 -0.055 -0.125* -0.046 -0.106 -0.065 0.077 -0.032 0.016 0.068 0.025 (0.043) (0.043) (0.062) (0.049) (0.067) (0.044) (0.067) (0.056) (0.072) (0.083) (0.073) (0.066) (0.117) Women*Ever entered higher education -0.000 0.035 -0.081* -0.056 0.061 0.048 0.121** 0.070 -0.083 -0.162** -0.033 -0.061 -0.149* (0.037) (0.036) (0.049) (0.048) (0.052) (0.036) (0.057) (0.049) (0.101) (0.078) (0.049) (0.054) (0.087) Observations 41,873 41,834 33,658 25,551 22,573 39,885 22,052 31,959 14,835 8,260 18,866 26,941 13,115 R-squared 0.019 0.020 0.010 0.013 0.013 0.011 0.013 0.019 0.015 0.013 0.007 0.008 0.013 P-val Education+Female*Education 0.000 0.000 0.0760 0.002 0.002 0.000 0.001 0.000 0.0240 0.126 0.0620 0.102 0.119 P-val Lower Secondary+Lower Secondary*Female 0.115 0.122 0.153 0.941 0.030 0.384 0.916 0.324 0.138 0.730 0.365 0.296 0.198 P-val Senior Secondary+Senior Secondary*Female 0.149 0.673 0.005 0.059 0.329 0.998 0.679 0.654 0.092 0.343 0.008 0.001 0.322 P-val Higher+Higher*Female 0.927 0.823 0.949 0.199 0.884 0.098 0.068 0.131 0.037 0.959 0.371 0.990 0.621 Mean SE skills for Men 0.133 0.117 0.124 0.062 0.0670 0.099 0.100 0.126 -0.009 0.078 0.107 0.088 0.139 Mean SE skills for Women 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Note: OLS regression specifications control for age and include study fixed effects. All studies have equal weights. Women is a dummy variable equal to 1 if the respondent is a woman, 0 otherwise. Education stands for the highest educational attainment (completed) where 0=No education, 1=completed grade 1, 2=completed grade 2, … 12=completed highschool, 13=completed certificate or diploma and 14=completed university degree or above. Note that Nigeria and Facebook projects have only categorical variable for education and "completed primary is coded as 9" and "completed secondary" is coded as 12. Ever entered lower secondary, senior secondary and higher education, is a dummy variable equal to 1 if the respondent ever entered the respective grades, 0 otherwise. Age bins and marital status are added as controls. Married is a dummy variable equal to 1 if the respondent is married/cohabitating, 0 otherwise. Age bins are added as controls. Age bins represent dummy variables equal to 1 if the respondent's age belongs to the age cohort which ranges from 15 to 65 with a 5 year gap, o otherwise. SE skills stands for Socio-emotional skills. PSDM stands for Problem Solving and Decision Making. ***, **, and * indicate significance at the 1, 5, and 10 percent critical level. 66