During the formative evaluation stages of the instrument, three pilot tests were conducted with mentally retarded group home residents. Alpha range of internal consistency in these pilots was .87 to .93. Estimates of test-retest reliability obtained in the third pilot test were .89 and .90. Content validity was evaluated by expert appraisals by 10 professionals and 55 graduate students whose feedback was incorporated in revisions of the instrument. In the summative evaluation, the instrument was field tested in five locations in Ohio, Wisconsin, Maryland, New Mexico, and Nevada, including psychiatric clinics and forensic services. Construct validity was evaluated through a comparison of 93 mentally typical defendants, mentally retarded defendants whose competence was not in question, and mentally retarded defendants who had been declared competent or incompetent by a forensic evaluator. The instrument was found to distinguish incompetent defendants from others. Linear discriminant analysis indicated that the Vocabulary and Concepts section of the instrument was the best predictor of competence, and that the instrument had a 70 percent accuracy rate for classification of competence. Inter-observer reliability checks showed a 96 percent level of agreement. An internal consistency analysis yielded an alpha coefficient of .93. Results suggest that the instrument provides a valid and reliable measure of competence to stand trial. Appendixes, 23 tables, and 110 references. (Author abstract modified)