ACCURATE ASSESSMENT OF COMMUNITY OPINION AND REACTION DEPENDS UPON ADEQUATE SAMPLING TECHNIQUES. VOTER REGISTRATION LISTS PROVIDE ONE SOURCE FROM WHICH TO DRAW SAMPLES. RESEARCHERS MUST CONSIDER THE COSTS OF LARGE SAMPLES IN RELATION TO TWO QUESTIONS--(1) THE CLOSENESS WITH WHICH THE TOTAL POPULATION IS TO BE MATCHED BY THE SAMPLE, AND (2) THE DESIRED DEGREE OF CERTAINTY OF THIS MATCHING. IN ORDER TO DETERMINE DESIRABLE SAMPLE SIZE, A FORMULA AND TABLE OF "CERTAINTY FIGURES" HAS BEEN DEVELOPED, AND EXAMPLES OF THEIR USE ARE PRESENTED. IF A SAMPLE OF ANY SIZE IS TO BE VALID, RANDOM SELECTION IS ESSENTIAL. VALIDITY OF DATA ARE ALSO DEPENDENT ON THE CARE USED IN DEVISING A SURVEY INSTRUMENT AND UPON THE METHOD OF HANDLING THE PROBLEM OF NONRESPONSE. THIS ARTICLE IS PUBLISHED IN"JUNIOR COLLEGE JOURNAL," VOLUME 36, NUMBER 3, NOVEMBER 1965. (WO)