MEMBERS OF THE EVALUATION SECTION OF HARVARD PROJECT PHYSICS, DESCRIBING WHAT IS SAID TO BE THE FIRST ATTEMPT TO SELECT A NATIONAL RANDOM SAMPLE OF (HIGH SCHOOL PHYSICS) TEACHERS, LIST THE STEPS AS (1) PURCHASE OF A LIST OF PHYSICS TEACHERS FROM THE NATIONAL SCIENCE TEACHERS ASSOCIATION (MOST COMPLETE AVAILABLE), (2) SELECTION OF 136 NAMES BY A TABLE OF RANDOM NUMBERS DERIVED FROM ORDINAL NUMBERING OF THE 16,702 LISTINGS FOR THE CONTINENTAL U.S., (3) SENDING OF A DESCRIPTIVE INVITATION TO PARTICIPATE IN EXPERIMENTAL EVALUATION OF THE COURSE (INCLUDING A QUESTIONNAIRE TO INDICATE WILLINGNESS, REASON FOR REFUSAL, AND INFORMATION ON TEACHING LOAD, TEXT USED, ETC.), (4) COMPARISON OF ACCEPTORS WITH NON-ACCEPTORS (ACCEPTORS APPEARED MORE RECEPTIVE TO INNOVATION), (5) ASSIGNMENT BY RANDOM NUMBERS OF 46 TEACHERS TO THE EXPERIMENTAL GROUP AND 26 TO THE CONTROL GROUP (ATTRITION OVER THE FOLLOWING 4 MONTHS REDUCED THE EXPERIMENTAL GROUP TO 36 AND THE CONTROL GROUP TO 21). POSSIBLE BIAS IS DISCUSSED, LEADING TO CHARACTERIZATION OF THE POPULATION AS "ALL HIGH SCHOOL PHYSICS TEACHERS ON THE 1966 NSTA LIST WHO DID NOT MAKE SUMMER COMMITMENTS MORE THAN 3 MONTHS IN ADVANCE AND WHO WOULD NOT BACK OUT OF THE RESEARCH AGREEMENT." COST AND COMPARISON WITH A VOLUNTEER GROUP ARE ALSO DISCUSSED. (AF)