Answers to the following questions were sought in this study: (1) Does the interview technique provide information that cannot be obtained from the usual normative approach?; (2) Does the interview technique provide information leading to the revision of mathematics test questions? Are the revised questions better than the original version in specified ways?; and (3) Does the interview technique provide information about the extent to which the student has been exposed in a mathematics course to the topic, concept, or skill that is central to the correct solution of a particular question? The College Board Mathematics Level II pretest was administered to 10 students in a senior mathematics course, to 30 juniors in a mathematics honors course, and to 35 juniors and seniors in mathematics courses. Fifteen selected students were then interviewed as to their methods of problem solution. The original pretest was then administered to the pretest population, and a pretest consisting of 16 questions identical to those in the original pretest and 9 questions that were revisions was also administered. Results of the study show that the answer to question (1) above was "yes"; in answer to question (2), nine questions in the pretest (items 2, 10, 14, 18, 20, 21, 23, and 24) were revised; and the answer to question (3) was a definite "yes." Appendixes provide the Item Interview Record and Test Interview Record, Procedures Used in Carrying Out the Study, Score Distributions for Total Group and Interview Group, Item Analysis for Total Group, Item Analysis for Interview and Pretest Groups and Interview Results for Each Item and Graphs. (DB)