Association for Public Policy Analysis & Management

Many developing countries provide cash to low-income families to encourage children to attend school. Some of these initiatives are called “scholarships” and others “cash transfers”, but they operate under the same theory of change. Low-income parents may not send their children to school if they perceive that the costs of schooling are too high, its benefits are too low (or take too long to materialize), or they lack access to credit (Banerjee et al. 2013). Thus, these programs aim to cover the costs of and raise the (immediate) returns to schooling, while relaxing credit constraints by offering their beneficiaries cash to enroll and stay in school (Fiszbein et al. 2009).

Scholarships and cash transfers are among the most rigorously evaluated educational interventions in developing countries (Ganimian and Murnane 2016). These initiatives have increased student participation in school, but they have rarely increased student achievement.

One reason why scholarships and cash transfers have had a limited impact on student achievement is that the beneficiaries of these programs may lack the requisite skills to succeed in school (Borghans et al. 2008; Farrington et al. 2012; Gabrieli et al. 2015). If the lack of these “character”, “socio-emotional”, or “soft” skills is a binding constraint for children, they could benefit from programs that combined scholarships or cash transfers with support to develop such skills.

This paper reports the second-year results of a three-year randomized evaluation of a program that provides seventh graders in the Province of Buenos Aires, Argentina with a scholarship and non-academic mentoring. To our knowledge, this is the first study to rigorously assess the effect of combining financial incentives with non-academic mentoring on school performance in a developing country.

We administered two surveys prior to randomization: (a) a student survey; and (b) a household survey. We followed students for two years and collected data on their: (a) program participation; (b) socio-emotional skills (self-beliefs about academics, organization and planning skills, motivation, grit, and self-control) through self-reports and performance tasks; (c) “school navigation” skills (frequency of negative school habits and proactive “preventive” and “corrective” behaviors concerning homework, exams, absenteeism, and free school periods); (d) school performance; and (e) academic skills.

We find that the program had the intended effects, but mostly not through its hypothesized mechanisms. On average, it reduced student absenteeism and grade failure, but it did not reduce drop out rates. It also improved school performance and school navigation skills, but it had limited impact on socio-emotional skills, and no impact on academic skills.

One way of interpreting these results is that instead of changing students’ general socio-emotional skills, such as grit and self-control, the program helped students succeed in school by affecting specific socio-emotional skills, such as doing homework in advance and reviewing failed assignments. This interpretation is consistent with the heterogeneous effects that we observe, since the program was most beneficial for students who are least likely to have developed these habits, such as students who had previously dropped out of school and students from low-income families.