The Stanford Research Institute (SRI) Classroom Observation Instrument was modified to study individual teaching configurations (teaching methods and styles) in adapting a criterion-referenced mathematics curriculum in grades 2, 4, and 6. Ten configurations characterized 72% of the 168 teachers who were interviewed. Although all teachers taught and tested students on program objectives and recorded achievement immediately, the configurations differed by choice of instructional material, grouping technique, and use of test results and remediation. Fifty eight of these teachers were observed in the classroom to identify differences between classrooms and between the 10 dominant configurations. Although the program allowed teachers to be flexible, most organized their classrooms aroung large group instruction and paper-pencil activity. Significant configuration differences were found for 25% of the Classroom Checklist variables, and 12% of the Five Minute Observation variables. The impact of the configurations on student achievement could not be determined because test data was not available, although research indicates that achievement is depressed if class management takes too much time. (CP)