Study
Heller (2014a) examined the effects of One Summer Plus–Jobs Only on a cohort of 1,634 students (grades 8–12) enrolled in Chicago, Ill. schools. Youth were identified as being at high risk for violence involvement through data from multiple city agencies on factors such as gang involvement and truancy. Thirteen schools with the highest number of high-risk youth were identified for the study. All enrolled students and rising ninth graders at these 13 schools were invited to enroll, and 75 percent of those agreed to participate.
Randomization occurred within school and gender blocks to account for different numbers of applications within schools and higher application rates of females, compared with males. Youth were randomly assigned to a treatment group (n = 730) or a control group (n = 904). Within the treatment group, youth were further randomly assigned to receive either 1) One Summer Plus–Jobs Only (n = 350), a condition in which they were provided 25 hours of minimum-wage employment per week for 7 weeks (with the option for an additional 8th week) plus a job mentor; or 2) One Summer Plus–Jobs + Social and Emotional Learning (SEL) (n = 350), in which youth were provided with 15 hours of minimum wage employment each week for 7 to 8 weeks, plus a job mentor, and spent 10 hours each week completing an SEL course. Thirty youth who were randomized into the treatment group declined to participate. Youth in the control group were free to pursue outside services as they wished. The CrimeSolutions review of this study focused on the comparison between the One Summer Plus–Jobs Only treatment group and the control group.
In the control group, 96 percent of the youth were Black, and 2.9 percent were Hispanic. Ninety-two percent received free/reduced priced lunch. Of this group, the average age was 16.2 years, and they had missed 18 percent (roughly 6 weeks) of the pre-program school year. Nineteen percent had been arrested at baseline. The number of arrests on average was 0.13 for violent crimes, 0.09 for property crime arrests, 0.05 for drug arrests, and 0.15 for other arrests. Twenty-one percent of youth had been victimized, and the average number of victimizations was 0.29.
The treatment group (Jobs Only and Jobs + SEL) also consisted of mostly Black youth (94 percent) with a small proportion of Hispanic youth (3.8 percent). Ninety-two percent received free or reduced priced lunch. Of this group, the average age was 16.8 years, and they had missed 18 percent (roughly 6 weeks) of the pre-program school year. Twenty-two percent had been arrested at baseline. The number of previous arrests on average was 0.18 for violent crimes, 0.09 for property crime arrests, 0.08 for drug arrests, and 0.19 for other arrests. Twenty-four percent of the youth had been victimized. The average number of victimizations was 0.33.
There were no statistically significant differences between the treatment and control groups in baseline characteristics, including age, grade, job, race, receipt of free/reduced lunch; number of grades (A–F); percent of days absent; school attendance for fall 2011; number of disciplinary incidents; in- and out-of-school suspensions; neighborhood characteristics such as median income and unemployment; number of each arrest type; and prior victimizations. Between each separate treatment arm (Jobs Only or Jobs + SEL) and the control group, the only statistically significant difference was that youth in the Jobs Only group received more Cs on their report cards for the fall semester prior to the program.
Violent crime arrests (assault, homicide, sexual offenses, robbery, threats, kidnapping, and aggravated arson), property crime arrests (larceny, burglary, non-aggravated arson, and motor vehicle theft), drug arrests (possession and dealing), and other arrests (e.g., trespassing, violence against animals, and parole violations) were measured using juvenile and adult Chicago Police Department arrest records covering dates up to September 15, 2013 (13 months after the end of the program). Demographic information was obtained from Chicago Public Schools records, and neighborhood demographics were obtained by matching youth address data to the 2010 American Community Survey. Program effects were assessed using regression analyses that controlled for baseline characteristics and blocking variables. No subgroup analyses were conducted.
