Identifying good cops early: Predicting recruit performance in the academy

Police departments have traditionally assessed their performance through crime-related activity measures that often have little to do with good police work and offer little hope for prediction of exemplary performance. This article suggests some progress can be made in predicting superior performance by considering an earlier stage in a police officer's career where performance is well-defined and measurable: the police academy. Using recruit performance data (n = 1,556) from a large metropolitan police department, the article uses linear and logistic regression, as well as Chi-square Automatic Indicator Detector (CHIAD), to identify predictors of superior performance in the academy. A number of interesting findings emerge with regard to factors that offer predictive value-reading level, age, gender, and race-and those that do not-college education, military experience, and residency. The article concludes with a discussion of the implications for recruitment, selection, and training, as well as for measuring and predicting performance on the street.