CHANGES WERE STUDIED IN THE PROBLEM-SOLVING BEHAVIOR OF HIGH SCHOOL STUDENTS AS A FUNCTION OF A CAREFULLY DESIGNED TRAINING PROGRAM. TRAINING WAS DEFINED AS THE DEVELOPMENT OF STUDENT AWARENESS OF PROBLEM-SOLVING STRATEGIES USED. INSTRUMENTS WERE DEVELOPED AND REFINED TO MEASURE PROBLEM-SOLVING BEHAVIOR. SPECIFICALLY TESTED WAS THE FOLLOWING HYPOTHESIS--PROBLEM-SOLVING BEHAVIOR MAY BE EVALUATED BETTER BY STUDYING THE SEQUENCE OF QUESTIONS ASKED THAN BY EVALUATING PROBLEM SOLUTIONS. SOME OF THE PROBLEMS PRESENTED SETS OF QUESTIONS, OTHER PROBLEMS GENERATED STUDENT QUESTIONS. RECORDS WERE KEPT OF THE NUMBER AND ORDER OF QUESTIONS ASKED. APPROXIMATELY 50 HIGH SCHOOL STUDENTS WITH IQ RANGES ABOVE 120 (MEASURED BY THE HENMAN-NELSON TEST) WERE PAIRED AND RANDOMLY ASSIGNED TO EITHER A TRAINING OR CONTROL GROUP. TWO IDENTICAL SETS OF PROBLEMS WERE PRESENTED TO THE GROUPS. ONE SET WAS GIVEN IN A PRETEST ENVIRONMENT. PERFORMANCES ON THE SECOND SET OF PROBLEMS, GIVEN AFTER TRAINING, WERE COMPARED WITH PRETEST RESULTS. THE SCORING METHODS USED CONSIDERED (1) NUMBER OF QUESTIONS ASKED, (2) CHARACTERIZATION OF PERFORMANCE PROCESSES, (3) VERBAL REPORTS FROM TEST SUBJECTS, AND (4) NUMBER OF CORRECT SOLUTIONS. REPORTED FINDINGS INDICATED PERFORMANCE OF THE TRAINING GROUPS WAS SIGNIFICANTLY BETTER THAN THAT OF THE CONTROL GROUP. UTILIZATION AND ORDERING OF QUESTIONS WERE ESTABLISHED FOR LATER PUBLICATION (ED 003 006). (WN)