FOUR HYPOTHESES WERE CONSIDERED DURING THIS EVALUATION STUDY, NAMELY THAT THE TEACHING OF ARITHMETIC IN GRADES 4, 5, AND 6 BY THE MADISON PROJECT WILL--(1) PRODUCE AS GREAT AN INCREASE IN ACHIEVEMENT AS THE METHOD COMMONLY PRACTICED, (2) PRODUCE A GREATER INCREASE IN ALGEBRAIC APTITUDE, (3) PRODUCE A GREATER INTEREST IN MATHEMATICS, AND (4) INFLUENCE CHANGES IN THE SCORES OF STUDENTS ON A TEST OF ABSTRACT THINKING EQUAL TO CHANGES EXPERIENCED BY STUDENTS RECEIVING INSTRUCTION AS COMMONLY PRACTICED. TWENTY-SIX TEACHERS IN SYRACUSE, NEW YORK, WERE SELECTED FOR PARTICIPATION AND WERE RANDOMLY DIVIDED INTO EXPERIMENTAL AND CONTROL GROUPS. STUDENTS IN ARITHMETIC CLASSES IN SYRACUSE PUBLIC SCHOOLS PARTICIPATED FOR ONE SCHOOL YEAR. INSTRUMENTS WHICH MEASURED INTELLECTUAL LEVEL, ALGEBRAIC APTITUDE, ARITHMETICAL ACHIEVEMENT, AND INTEREST IN MATHEMATICS WERE ADMINISTERED AT THE BEGINNING AND END OF THE SCHOOL YEAR. COMPARISON OF THESE PRE- AND POST-TESTING DATA REVEALED THAT STUDENTS IN THE EXPERIMENTAL GROUPS INCREASED THEIR SKILL AND UNDERSTANDING OF COMPLEX MATHEMATICAL PRESENTATIONS, AND AT THE SAME TIME APPARENTLY DID NOT LOSE ANY ABILITY TO BE SUCCESSFUL IN CONVENTIONAL ARITHMETIC ACHIEVEMENT AS MEASURED BY TESTS. INCONSISTENCIES WERE NOTED FOR BOTH EXPERIMENTAL AND CONTROL GROUPS IN THE ANALYSIS OF STUDENT ATTITUDES TOWARD ARITHMETIC AS MEASURED BY THE QUESTIONNAIRE. (WB)