The relationship between predictor variables and individual performance measures was determined by cross-tabulations and simple correlations. The stepwise least squares regression program was used to analyze variables. The results show that factors can be identified which predict performance for two distinct cohorts in the same police organization. The strongest predictors of police performance (such as higher education) are similar for both active cohorts. Education was a strong predictor of job performance, and the number of geographical transfers made by officers appears to be a strong predictor of later job performance. Preservice auto accidents and traffic citations were good predictors of future unsatisfactory performance. Members of the Michigan State Police were quite skillful at evaluating the applicants' characteristics and arriving at an overall appraisal. Recommendations for upgrading recruitment and selection criteria include increasing recruitment of minorities and college graduates; assigning older, college-educated new recruits to sensitive areas of the State; reviewing the background investigator's role in evaluating potential candidates; developing a computer-based information system to analyze police officer performance data and background characteristics; and developing new criteria for evaluating performance. Four appendixes include minimum entrance requirements, the police selection method, coding guidelines, and a correlation matrix. Footnotes, 43 tables, and over 100 references are included.