THE USE OF SOCIAL INDICATORS TO EVALUATE THE IMPACT OF SPECIFIC SOCIAL PROGRAMS REQUIRES EXTENDED TIME-SERIES OF SOCIAL INDICATORS -- LOCAL STATISTICS ON UNITS THAT RECEIVE THE SERVICES OF A NEW PROGRAM AND ON COMPARABLE UNITS THAT DO NOT RECEIVE THE SERVICES. STATISTICS ON DEATHS, DAYS HOSPITALIZED, DAYS LOST FROM WORK, TRAFFIC ACCIDENTS, UNEMPLOYMENT, ACHIEVEMENT TEST SCORES, AND CRIME ILLUSTRATE THE KINDS OF INDICATORS THAT CAN BE USEFUL WHEN MADE AVAILABLE BY SCHOOLS, CENSUS TRACTS, POLICE DISTRICTS, CITIES, COUNTIES, OR STATES. THESE STSATISTICS SHOULD BE MADE FLEXIBLY RESEARCH-RETRIEVABLE BY LOCAL REGION, BY FREQUENT TIME INTERVAL, AND IN FINE TOPICAL BREAKDOWN. THE ABILITY TO REPORT AGGREGATE STATISTICS RETRIEVED FOR LISTS OF PERSONS WOULD ADD EVEN GREATER PRECISION IN PROGRAM EVALUATION AND COULD BE DONE WITH NO LOSS OF PRIVACY AND WITHOUT ANY RELEASE OF INDIVIDUAL DATA. THE USE OF UNIFORM INDIVIDUAL IDENTIFICATION NUMBERS COULD ADD EFFICIENCY TO EVALUATION WITHOUT POSING INCREASED RISKS TO PRIVACY. GRAPHS ILLUSTRATING THE APPLICATIONS OF TIME SERIES DATA IN PROGRAM EVALUATION ARE INCLUDED. A BIBLIOGRAPHY IS PROVIDED. (AUTHOR ABSTRACT MODIFIED--LKM)