DISCRIMINANT ANALYSIS LIES SOMEWHERE BETWEEN MULTIPLE REGRESSION AND CANONICAL CORRELATION IN ITS USEFULNESS FOR ANALYZING SOCIAL SCIENCE DATA. SINCE IT WAS FIRST INTRODUCED IN 1936, IT HAS BEEN USED MAINLY IN THE FIELD OF EDUCATION TO CLASSIFY PUPILS ON THE BASIS OF TEST SCORES. IN THE 1970'S, IT WAS DISCOVERED BY OTHER SOCIAL SCIENCE DISCIPLINES AND HAS SINCE BEEN USED WIDELY TO DIFFERENTIATE INDIVIDUALS OR OBJECTS ON THE BASIS OF CERTAIN KNOWN CHARACTERISTICS. THE MATHEMATICAL BASIS OF THE METHOD IS DISCUSSED; AND THE TECHNIQUE IS THERE ILLUSTRATED WITH A STEP-BY-STEP ANALYSIS OF A STUDY OF CITIZEN COUNCIL MEMBERS' ATTITUDES. THE STUDY TRACED CHANGES IN ATTITUDES OF 99 MEMBERS OF A LOCAL SOCIAL SERVICES COUNCIL. THE STUDY RELATED BOTH ATTITUDES AND BACKGROUND CHARACTERISTICS TO CITIZEN PARTICIPATION AND ATTEMPTED TO PREDICT CORRELATIONS BETWEEN ATTITUDES AND SOCIAL STATUS. THE CLASSIFICATION ANALYSIS CORRECTLY LOCATED 71 PERCENT OF THE SERVICE PROVIDERS IN THE CORRECT GROUP, WITH 50 PERCENT OF THE CONSUMERS AND 49 PERCENT OF THE COMMUNITY REPRESENTATIVES CORRECTLY ASSIGNED. THIS IS A 33.4 PERCENT IMPROVEMENT OVER CHANCE. THE STUDY ITSELF EXPECTED TO CONSUMER TO HAVE LOWER SOCIOECONOMIC STATUS AND A LOWER LEVEL OF CIVIL PARTICIPATION. INSTEAD, ATTITUDE, NOT BACKGROUND, PROVED TO BE THE MOST SIGNIFICANT PREDICTOR OF COUNCIL PARTICIPATION. THE ARTICLE CONTAINS NOTES, REFERENCES, DATA TABLES, AND A LIST OF DISCRIMINANT COMPUTER PROGRAMS. (GLR)