Data were gathered from the personnel files of 14 police departments, sheriff's offices, and a highway patrol serving populations of a wide range of sizes. Data were gathered on over 2,000 officers who had been separated from their agencies during the decade preceding the study's beginning and over 2,000 officers who were currently employed. From the group of separated officers, failures were defined as those whose services were no longer considered to be of value to their departments due to inefficiency on the job or because of nonjob behavior which was intolerable to the department. Officers who were terminated for cause by their departments tended to be younger at the time of appointment, to have a greater number of years of education, to have more marriages and shorter work histories, and to have more citations for vehicle code violations and other violations. The failures presented a pattern of greater mobility and uncontrolled impulsivity than did the officers who were retained. The finding on educational attainment suggested that police departments do not sufficiently meet the needs of their more educated officers. The other findings substantiated the hunches of many police officers. Findings indicated that screening devices for prospective law enforcement officers should be specifically tailored to the unique needs of the law enforcement profession rather than trying to determine which applicants are emotionally unstable. A person with no emotional pathology may be a higher risk than someone whose emotional makeup includes certain problem areas. Footnotes, a list of 22 references, and appendixes presenting data tables and related material are provided.