Social diversity at school, school performances, and social skills – DIVERSCHOOL

Our project is motivated by the increasing socio-economic inequalities in the French educational system. As existing policies have failed to reduce inequalities, it is necessary to develop new tools to tackle the issue of high social inequalities at school. This research project evaluates the impacts of a large desegregation program that has started in September 2016 in France. The program involves 22 sites and 78 middles schools (56 public and 22 private). The sites were approached to implement a variety of student mixing strategies, including: i) redrawing school catchment areas; ii) merging school districts into a single multi-school district in which students are assigned through a school choice algorithm; iii) pooling students by cohort and then assigning them alternatively to one middle school or the other; iv) closing a middle school and assigning its students to the neighboring schools, or building a new middle school; v) introducing specific optional courses in a disadvantaged school, such as bilingual courses, to attract more socially advantaged students. The project aims to analyze the effects of this large desegregation program on segregation, as well as on a wide range of student outcome – both cognitive and non-cognitive. It relies on a difference-in-difference strategy to assess the effects of the desegregation actions: we compare cohorts benefitting from the program with older cohorts not benefitting, in both participating and non-participating middle schools. In order to build a counterfactual to the treated schools, we matched our 22 treatment sites with 22 control sites using administrative data from the Statistical department of the Ministry of Education (DEPP). The middle schools in the control site were chosen to be as similar as possible as those in the treatment site using a minimum distance criterion for social composition and final exam pass rate. Hence, our sample includes 156 middle schools from 27 French départements. We will first assess whether the program effectively increases social diversity at school. Preliminary results show that the average desegregation action led to only a modest decrease in segregation: we observe a slight increase in the proportion of high SES students in schools attended by low SES, and no change in exposure of high SES students to low SES. However, the small average effect hides great variations across sites. In some sites, the change in social composition is huge while in other sites it is almost non-existent. This project thus proposes to deepen the analysis and explore the heterogeneity in action efficiency by action category, in particular depending on the degree of constraints imposed by the action on school choice. Second, we will measure the impact of social diversity on average academic performances and social inequality. To this end, we will exploit the variety of actions that have been undertaken and focus on 10 sites (79 middles schools) where the actions have a great potential efficiency. In these sites, we can compare students who experienced a great decrease in segregation with students who are similar except that they did not benefit from a radical change in segregation. We will use econometrics to estimate the impact of social diversity on academic performance levels and inequality between low and high SES students. Third, using the same methodology, the project will assess the impact of social diversity on social and behavioral skills, in particular views on social differences and tolerance, perceptions of merit, social fatalism and optimism, academic self-esteem, and the composition of interpersonal networks. The project will thus provide a unique contribution by assessing quantitatively the effects of social diversity on cooperation, social integration, and social cohesiveness.