Institutions selected for the study were classified as either medium or maximum security. The Gray-Votaw-Rogers and the General Educational Development tests were the most often administered educational tests. Other achievement tests that may be appropriate include the Adult Basic Learning Exam, the Peabody Individual Achievement Test, the College Qualifications Test, and the Gates MacGinitie Reading Tests. Many more residents were given educational or academic achievement tests than intelligence tests. An appropriate intelligence test would be the Raven Progressive Matrices, especially good for a population with potentially different cultural experiences and low educational achievement. The measures of personality most common to all institutions in the study were the Minnesota Multiphasic Personality Inventory (MMPI) and the Rotter Incomplete Sentence Blank. However, the use of the MMPI should be regulated carefully according to the reading levels of inmates, and other tests that would be useful include the 16 Personality Factor questionnaire and the Edwards Personal Preference Schedule or the Mooney Problem Checklist. The use of the Bender Gestalt seems inappropriate unless brain damage is suspected. Five of the 11 institutions gave no vocational tests other than the General Aptitude Test Battery. As clients with low frustration tolerance often express negative feelings about this test, the Differential Aptitude Test Battery might be used, along with other specialized tests (the Bennett Mechanical Comprehension Test), or interest tests such as the Minnesota Vocational Interest Inventory. Accommodations also must be made in testing special populations. Clearly, future research must address the purposes of such testing, the administration of the tests, and the appropriateness of the tests to the population. Tables and 10 references are provided.