Abstract: Individually administered tests tend to contain more appropriate items and be better focused than their group counterparts. Group administered tests, however, tend to assure uniformity, reliability, and validity at a greatly reduced cost. Computer adaptive testing (CAT) provides a mix of the best of group testing and individual testing while still providing an accurate picture of the examinee's ability. Because examinee questions are quickly matched to the examinee's ability level, examinees are neither bored with easy questions nor frustrated with overly difficult materials. CAT versions of tests have been shown to yield results that are equivalent to paper-and-pencil tests, but the fact that the first models of CAT are ready for use does not mean that improvements are not possible. There is room for improvement in any testing program, and issues that need additional study in CAT are: (1) item pool development; (2) test equating; (3) test administration; (4) test interpretation; (5) reliability, precision, and validity in moving from paper and pencil tests to CAT; (6) future testing strategies and choices; (7) timing constraints; (8) dealing with omissions; (9) the creation of testlets; (10) hierarchical testlets; (11) equivalent tests; and (12) self-adapted testing. The application of CAT to many different testing environments is under way. The future of CAT is very exciting. It is recommended that the Maryland State Department of Education try a full-fledged implementation of a CAT in one or two counties in 1992.