THE RESULTS OF PROGRAMED INCORRECT SPELLING PRACTICE ON ACHIEVEMENT AND RETENTION WERE STUDIED. SIXTY FOURTH-GRADE SUBJECTS WERE PAIRED ON THE BASES OF SEX AND EQUAL SPELLING ACHIEVEMENT. ONE MEMBER OF EACH PAIR WAS THEN RANDOMLY ASSIGNED TO EXPERIMENTAL GROUPS A OR B. AN ADDITIONAL 20 FOURTH-GRADE STUDENTS WERE ASSIGNED TO EXPERIMENTAL GROUPS C AND D IN A SIMILAR MANNER. A PRETEST CONTAINING 200 DIFFICULT SIXTH-GRADE WORDS WAS ADMINISTERED TO ALL SUBJECTS. SUBJECTS WERE REQUIRED TO SPELL WORDS WHETHER THEY COULD OR COULD NOT SPELL THEM CORRECTLY. FORTY WORDS MISSPELLED BY ALL MEMBERS OF GROUPS A AND B WERE SELECTED FOR PROGRAMED PRACTICE. FOURTEEN WORDS MISSPELLED BY ALL MEMBERS OF GROUPS C AND D WERE SIMILARLY SELECTED. IN GROUPS A AND B EACH SUBJECT PRACTICED CORRECT AND INCORRECT SPELLING OF 20 WORDS INCORRECTLY SPELLED ON THE PRETEST. CORRECT AND INCORRECT PROGRAMED SPELLING PRACTICE WAS ALTERNATED SO THAT SETS OF WORDS PRACTICED CORRECTLY AND INCORRECTLY VARIED BETWEEN GROUPS A AND B. GROUPS C AND D DID NOT RECEIVE PROGRAMED PRACTICE BUT WROTE EXPERIMENTAL WORDS (CORRECT OR ERROR FORM) 15 TIMES. MEAN POST-TEST ACHIEVEMENT SCORES FOR NEGATIVE AND POSITIVE PRACTICE WERE COMPARED FOR ALL METHODS. SIMILAR COMPARISONS WERE MADE WITH RETENTION SCORES. RESULTS FAILED TO INDICATE SUPERIORITY OF EITHER NEGATIVE OR POSITIVE SPELLING PRACTICE. (WN)