Observing that available standardized tests purporting to measure problem solving in mathematics do not validly reflect currently accepted definitions of problem solving, while more valid individualized measures are costly in terms of time, this researcher undertook (1) to construct a paper-and-pencil instrument which might better reflect problem-solving ability and (2) to validate this test by correlation with scores of subjects in taped and coded problem-solving interviews. Seventh-grade students with average and above-average mathematics achievement records were given two written tests in a group situation; later each was interviewed and given six problems to solve while talking aloud in a videotaping or audiotaping situation. Scores were assigned to performance in the individual session according to a previously developed coding system. Test statistics were computed; differences in the rank ordering of subjects by the two measures and correlations among scores were analyzed; differences in performance and coding ease between videotaped and audiotaped subjects were investigated. Results and instruments used are reported in Part 2 of this report. (SD)