THE RELATIVE EFFECTIVENESS OF FIVE METHODS FOR IMPROVING THE SPELLING ABILITY OF PUPILS IN GRADES 2 AND 3 WAS DETERMINED. ASSIGNMENTS OF TEACHERS AND CLASSROOMS TO METHOD WERE BASED UPON TEACHER PREFERENCE FOR METHOD, ESTIMATED AVERAGE ACHIEVEMENT LEVEL OF PUPILS IN EACH CLASSROOM, AND RATINGS OF TEACHER QUALITY. AT THE BEGINNING OF THE STUDY A GROUP INTELLIGENCE TEST WAS GIVEN TO EVALUATE THE INTERACTION OF METHODS AND ABILITY LEVELS. TWO PREEXPERIMENT SPELLING TESTS, ONE BASED ON THE BASAL WORD LISTS FOR THE YEAR AND ONE BASED ON WORDS NOT TAUGHT DIRECTLY, WERE ALSO ADMINISTERED. AT THE CLOSE OF THE STUDY THREE FINAL CRITERION TESTS WERE EMPLOYED, THE TWO PREEXPERIMENT TESTS WERE READMINISTERED, AND A SENSITIVITY-TO-ERROR TEST IN A PROOFREADING SITUATION WAS GIVEN. THE FIVE METHODS USED WERE ENTITLED TEST-STUDY-TEST, WORD PERCEPTION WITH TEST, WORD PERCEPTION WITHOUT TEST, CORRECTION AND PROOFREADING, AND WORKBOOK. DESCRIPTION OF THE DATA COLLECTED AND THE ANALYSES OF DATA ARE INCLUDED. NOTHING IN THE RESULTS SUGGESTS A CLEAR CUT SUPERIORITY OF ONE METHOD OVER ALL THE OTHERS.