A survey was conducted of expert instructional design practitioners about their use of developmental testing and revision procedures. The degree of correspondence was examined between published recommended practice and actual practice, as reported by 90 respondents to a questionnaire. The surveyed sample were members of the National Society for Performance and Instruction. The SPSSX Multiple Response procedure was used to analyze the questionnaire data, since respondents were permitted to select more than one response per question. In developmental testing, which is a form of formative evaluation, qualitative or quantitative data, or both, are obtained for the revision of the instructional product. Results indicated that respondents endorsed a number of the recommended practices, such as using experimental subjects, using peer review, using both criterion-referenced tests and subjects' comments, and using subjects of varying abilities. Based on the response patterns found in the data analysis, the following representative profile of developmental testing and revision procedures is presented: (1) select a representative sample; (2) conduct small group sessions and perhaps also one-to-one sessions; (3) pretest the subjects, present the learning material, note their responses, posttest, and debrief the subjects; (4) analyze and interpret the data; (5) make revisions; and (6) test the revisions. (The questionnaire is appended). (GDC)