Panel Paper: Testing Heterogeneity in the Effects of Summer Jobs

Friday, November 7, 2014 : 8:50 AM
Enchantment Ballroom B (Hyatt)

*Names in bold indicate Presenter

Jonathan Davis, University of Chicago and Sara Heller, University of Pennsylvania
Heller (2013) provides first-of-its kind experimental evidence on the efficacy of a summer youth employment program, Chicago’s 2012 One Summer Plus program, in reducing crime. Applicants to this program were recruited from 13 public high schools, so all youth in the study were still connected to school. Given the enormous reduction in violence seen for that population, the City hoped a similar strategy could have an even larger impact by targeting youth who are at the very highest risk of violence involvement. However, little is known about the generalizability of the initial results to other, more disconnected populations.  

This project presents a follow-up randomized controlled trial measuring the efficacy of a second implementation of Chicago’s One Summer Plus program, including new evidence on treatment heterogeneity. Youth were recruited from two applicant pools: males from the juvenile justice system who are among the most criminally-involved youth in the city, and males living in one of Chicago’s 30 most violent neighborhoods who applied to Chicago’s broader summer programming. Having these two applicant groups allows us to assess whether program effects are different for youth at greater risk of being disconnected from school and involved in crime (group 1) than they are for more connected youth (group 2). These results will shed important light on the robustness and generalizability of the One Summer Plus approach.

In total, 2,130 eligible youth from group 1 and 3,098 eligible youth from group 2 applied for 1,000 available treatment slots. We randomly assigned applicants to a control group or to be offered a 15-hour per week minimum wage part time job and 10 hours per week of paid social emotional learning for six weeks. Preliminary results suggest that there may be important treatment heterogeneity in terms of the program’s effect on arrests and victimizations, which would have important policy implications for the targeting of summer jobs programs.