Association for Public Policy Analysis & Management

We report on the results of a large-scale randomized controlled trial (RCT) from Chicago of a behavioral intervention that shows evidence for substantial impacts on criminal behavior for disadvantaged youth. The intervention is built on the idea that because deliberate decision-making and conscious cognition (what psychologists call “system 2” thinking) require effort, people rely heavily on automatic responses that are adaptive to commonly encountered situations (“system 1”). Our RCT includes elements of what psychologists call cognitive behavioral therapy (CBT) designed to reduce automaticity. We all develop and rely on automatic responses, but the variability in situations requiring them can lead to conflict for disadvantaged youth. While automatic retaliation to a perceived challenge from a peer outside of school can deter future victimization, the same response to a teacher inside school could lead to disciplinary action such as suspension or expulsion. The results of our experiment suggest that it is possible to generate sizable changes in outcomes by helping disadvantaged youth recognize contexts in which their automatic responses may not be useful. While our data on mechanisms of action have some limitations, we do not find any detectable evidence that these impacts are driven by a generic mentoring effect, or changes in perceived returns to schooling, emotional intelligence, or “grit.” We do find some evidence consistent with the idea that reduced automaticity helps explain program impacts.