FOR THE THIRD YEAR, THE FOLLOWING FOUR DIFFERENT METHODS OF TEACHING READING WERE CONTRASTED--(1) THE INITIAL TEACHING ALPHABET, (2) A PHONIC, FILMSTRIP, WHOLE-CLASS APPROACH, (3) A WHOLE-WORD, ECLECTIC BASAL READER METHOD, AND (4) THE PRECEDING APPROACH SUPPLEMENTED BY A PHONICS PROGRAM. SOME 400 FIRST-GRADE PUPILS WERE RANDOMLY ASSIGNED TO TREATMENTS UPON ENTRANCE INTO FIRST GRADE AND WERE FOLLOWED AS INTACT CLASSES INTO SECOND AND THIRD GRADES WITH THE SAME METHODS, EXCEPT THAT I/T/A PUPILS TRANSITIONED INTO THE MERRILL "TREASURY OF LITERATURE" PROGRAM IN SECOND GRADE. A SMALL REPLICATIVE STUDY WAS DONE IN THE LAST 2 YEARS OF THE PROJECT. CONSIDERABLE INSERVICE EDUCATION WAS PROVIDED ALL TEACHERS. TEACHING WAS OBSERVED FREQUENTLY BY A FULL-TIME FIELD DIRECTOR AND OTHER SUPERVISORY PERSONNEL TO CHECK ON TEACHER COMPETENCE AND ADHERENCE TO METHOD. TEACHERS COMPLETED ACTIVITY LOGS AS AN ADDITIONAL METHOD SAFEGUARD. SIGNIFICANT DIFFERENCES RESULTED FROM ANALYSIS OF VARIANCE AND COVARIANCE WHICH GENERALLY WERE IN FAVOR OF PROGRAMS (1) AND (2). THIS SUGGESTED THAT A COMBINATION OF (1) AND (2) WOULD BE WORTHY OF FUTURE USE AND ADDITIONAL STUDY. PROGRAM (2) PRODUCED THE BEST OVERALL RESULTS, PARTICULARLY FOR THE HIGH IQ THIRD. HOWEVER, THE PUPIL RETENTION RATE WITH THIS PROGRAM SUGGESTED THAT IT MIGHT BE A BETTER PROGRAM WITH ABILITY GROUPING. (AUTHORS)