The 58 male and 3 female recruits (average age 23 years, 3 months) were tested within the first week of their training. Situation tests, which placed recruits in simulated situations designed to measure certain skills, were presented individually in a standardized manner in the presence of two experimenters and an actor. The scripts of the six tests are provided. These tests attempted to measure approaches to maintaining personal rights and justice in social situations, as well as the ability to relate to a stranger requesting help and to verbalize and resolve situations while under mild stress; other tests were designed to demonstrate attitudes toward others and abilities in a socially recognizable position of authority. Subject performance was measured after 2 and 4.5 years on the job. At 2 years, the test which measured ability to relate to a stranger requesting help was the only situation test that significantly discriminated among the three groups of raters: sergeants, staff sergeants, and groups of raters: sergeants, staff sergeants, and senior constables. The test which involved speaking 3 minutes on 'what's wrong with the world' was very close to separating the groups. None of the variables, considered independently, were significantly related to job performance after 4.5 years on the job. In sum, the two tests which relied upon verbal expressiveness were the better predictors of job performance after 2 years on the job, suggesting that police selection officers should consider the verbal fluency and thought content of recruits as shown in these simple situation tests. Tabular data and 17 references are provided.