Trust behind bars: Measuring change in inmates' prosocial preferences

The paper presents the results of a Longitudinal Lab-in-the-Field Experiment implemented between September 2015 and July 2016 performed in two State Prisons in California (USA) to measure change in prosocial preferences. A subset of eligible inmates willing to undertake GRIP (Guiding Rage Into Power) program, were randomly assigned to it. The paper tests whether the participation to this program (used as a treatment in the experiments) affects prosocial preferences of participants, with specific reference to trust. The results of a Difference-in-Differences (DID) estimation procedure show that trust significantly increased in GRIP participants compared to the control group. This result is robust to alternative estimation techniques and to the inclusion of an endogenous behavioral measure of altruism.