Decision Making Among Adult Offenders and Non Offenders

To expand our understanding of the rational choice model of offender decision making the investigators plan to study the decision making processes of a sample of incarcerated offenders and a comparison sample of like-aged non-offenders from the community. They will select 500 imprisoned offenders (300 males and 200 females) who are currently residing in a pre-release center. The offenders will range in age from 19 to approximately 65, will include whites, African-Americans, and Hispanics. The investigators will recruit a comparison non-offender sample of 250 non-offenders (150 males and 100 females) from the general community. All subjects will respond to two hypothetical crime scenarios (petty theft and assault). Participants will be measured on several characteristics that have been shown in past research in psychology and economics to be related to decision making: risk seeking, risk avoidance, sensation seeking, risky decision making, perceptions of both formal and informal sanctions, and the presence of visceral factors or a craving for risky decisions like crime. In addition, a sample of 120 offenders (BO males and 40 females) will be interviewed in the community approximately 12 months after their release. In addition to their questionnaire data, the investigators will collect one year follow-up information from official sources from all those in the offender sample.

Descriptive comparisons between offenders and non-offenders, males and females, and younger adult versus older adult offenders will be conducted using standard regression-based methods and difference in means tests. When comparing outcomes across experimental conditions, the investigators will use one-way and two-way ANOVAs to test for differences across key conditions. To explore heterogeneity in these measures within the offending sample, the investigators will identify latent classes (i.e., typologies of individuals) that reveal different patterns of characteristics related to decision making across multiple domains. ca/ncf