Skip to content
Findings from RCT Grants

A Randomized Controlled Trial of Rapid Employment and Development Initiative (READI), to Prevent Violent Crime

An interim analysis of this well-conducted RCT found no statistically significant effect on overall serious violent crime over a 20-month follow up period.

Grantee: University of Chicago Crime Lab. The full study report is linked here.

Description of the Intervention. This is an ongoing randomized controlled trial (RCT) of the Rapid Employment and Development Initiative (READI), a program that provides transitional jobs, cognitive behavioral therapy (CBT), and additional services over an 18 month period for adult men at high risk of being perpetrators and/​or victims of gun violence. As delivered in the context of the current study, it costs approximately $30,000 over the course of the program to recruit, serve, and pay wages to each READI participant.1

Study Design. Between August 2017 and March 2020, the study randomly assigned 2,456 adult men to have an offer of READI or continue having access to other services.2 At the time of study enrollment, 98% of sample members had previously been arrested; they averaged 17 prior arrests, and 35% had previously been shot. This interim report presents findings on the study’s primary outcome: a composite measure of involvement with serious violent crime, 20 months after random assignment. The intervention was generally well-implemented (e.g., 55% of treatment group members attended at least one day of programming; and those who participated in the program worked 75% of the total weeks available to them in their transitional jobs). The study’s final report, expected in 2023, will present results for the full sample over the 40 month period after random assignment. 

Impact on the Primary Outcome. Over the 20 months following random assignment, the study’s interim analysis found no statistically significant effect on the primary pre-registered outcome: an index” measure that combines (i) the number of arrests for Part I3 violent crimes and (ii) the number of serious violence victimizations reported to the Chicago Police Department (the treatment group saw a negligible, non-significant reduction in the combined index of serious violence of 0.03 standard deviations (p=0.21)).4 This outcome was measured using data from the Chicago Police Department. 

In their report, the study authors discuss possible explanations for READI’s observed lack of effects on the primary outcome (e.g., suggestive evidence that READI may have reduced some forms of violence and not others). The study is examining a range of policy-important secondary and exploratory outcomes as described in their study report.

Study Quality. Based on a careful review, we believe this study was well-conducted RCT and produced valid interim findings.5

  1. 1

    This is the approximate cost per treatment group member regardless of whether an individual participated in READI services (i.e., the overall cost of offering READI divided by the number of people in the study’s treatment group). The research team estimates this as the per-person cost over an average of 20 months per person, which includes the time needed to recruit study participants and the time to deliver the 18-month services.

  2. 2

    “Other services” refers to any other programming offered in the Chicago area over the same time period, including other programs aimed at reducing violent crime. This interim study does not report on which specific services were available or whether control group members participated in such services.

  3. 3

    Per the study’s pre-analysis plan: “Part I offenses are defined as Homicide, Manslaughter, Aggravated Assault, Aggravated Battery, Robbery, and Criminal Sexual Assault per the Uniform Crime Reporting (UCR) database.”

  4. 4

    This is the “intention-to-treat” effect – i.e., the impact of being offered access to the program.

  5. 5

    For example, the study had successful random assignment (as evidenced by highly similar treatment and control groups, as verified in correspondence with the study authors), valid analyses that were publicly pre-registered, and a primary outcome that was measured using high-quality administrative data.