Apr 04, 2016


One Summer Plus is a summer job program aimed at preventing violence.  Participants in the intervention were significantly less likely to be arrested for violent crimes in the following year although no additional impacts were found for those who received an additional socio-emotional learning curriculum.


Target population: Teenagers living in violent neighborhoods

One Summer Plus is a summer job program.  The program links Chicago students (8th grade to 12th grade) who attend high schools in some of the most violent neighborhoods in the city with summer, minimum wage ($8.25/hour) jobs.  In addition some of the students receive a socio-emotional learning (SEL) component where they are given skills based in cognitive behavioral therapy to understand and manage their thoughts, emotions and behaviors.  Students in the jobs only program work 25 hours per week and students in the jobs+SEL program work 15 hours per week and receive 10 hours per week of SEL.  Both experimental groups are paid.  The desired goals of the program are to reduce violence with the hypothesis that young people might stay busy over the summer.  The One Summer program is implemented over 8 weeks in the summer by the Department of Children’s and Family Services (DCFS).


Heller, SB. (2014). Summer jobs reduce violence among disadvantaged youth. Science, 346, 6214, 1219-1223.

Evaluated population: In this evaluation, 1,634 8th to 12th graders were randomized into one of three groups.  These young people came from 13 different high schools in violent neighborhoods were the majority of 5,000 previously identified “most at-risk” kids in Chicago came from; though, all kids in these schools were eligible to participate. There were 904 young people in the control group and 730 in the two different treatment groups (364 in the jobs only group and 366 in the jobs+SEL group).  The average age in the two groups was just under 17 years, most of the young people had just finished 10th grade, 95 percent of them were Black, and 92 percent qualified for free or reduced lunch.  Twenty percent of the participants reported ever being arrested for violent, property, drug, or other crimes and a similar percentage reported being the victim of crimes themselves.  Participants also reported being absent from school nearly 20 percent of days the previous year and had a C average GPA.  The participants all came from neighborhoods with high violence (approximately 2.1 violent crimes per 100,000) and low-income (median household income of $34-35,000).

Approach: Participants in this study were randomized at the individual level in gender-school blocks.  This was done in order to allow the program to manage the number of young people that each implementing organization managed (as some schools had more applicants than others but agencies worked with specific schools) and to over-accept male applicants (who were less likely to apply). There were no statistically significant differences in the demographic characteristics of the participants across treatment and control groups in this study at baseline.  Data were collected by matching participant data to administrative data.  For example, arrest results come from the police department, neighborhood results come from matching the respondents’ addresses to their census tracts, and schooling data come from the Chicago Public Schools.  The randomization occurred at month zero, the program ended at month three, and trends are studied for 13 more months.  This study explores three sets of questions.  First, was there a reduction in arrests for participants in either of the intervention arms relative to the control?  Second, were there differences between the two control arms?  Third, are there explanations for the results that support theory about why behavior might have changed?  Outcomes of interest include total arrests (for violent, property, drug, and other crimes), violent crime arrests over 16 months after the study, and days present in school. Ninety percent of the participants completed at least 7 weeks of the program (the 8th week was optional).

Results: There were significant differences for treatment compared with control youth in terms of the number of violent arrests but not in terms of other types of arrests.  Over the 16-month follow-up period, violent crime arrests in the two treatment groups combined decreased by a statistically significant 43 percent.  However, there were no differences between the two different treatment groups; the mean number of violent crime arrests in the two groups was statistically similar.

In order to test whether the impact of the program can be explained by the participants being busy – or as some authors have described: an “incapacitation” hypothesis – trends were explored.  However, this hypothesis is not supported because there is a steady and significant decrease in violence over the 13 months studied after the program ended.  Thus, change in violence cannot be explained simply by the participants being busy and distracted and thus having less time for crime. Finally, a second theoretical question was asked: were the young people who had been exposed to jobs were more likely to invest in school subsequently?  No evidence was found though as there was no significant impact on the number of days in school or academic outcomes in the school year after the program.



Heller, SB. (2014). Summer jobs reduce violence among disadvantaged youth. Science, 346, 6214, 1219-1223.


Contact Information

Sarah Heller conducted the research while at the University of Chicago Crime Lab

Crime Lab

KEYWORDS: Adolescents, job skills, employment, arrest, violent crime, victimization, school attendance, Black, urban, community-based, summer program

Program information last updated on 3/25/16.