A STUDY WAS DESIGNED TO EVALUATE THE USEFULNESS OF AUTOMATED TEACHING PROCEDURES FOR HELPING MENTALLY RETARDED CHILDREN LEARN WORD RECOGNITION, READING, AND SPELLING. THE SUBJECTS FOR THE STUDY (66) WERE DRAWN FROM ESTABLISHED PUBLIC SCHOOL CLASSES FOR THE EDUCABLE MENTALLY RETARDED. SUBJECTS, RANGING IN AGE FROM 8-16 YEARS, WERE MATCHED AND ASSIGNED RANDOMLY TO EITHER AN AUTOMATED TEACHING GROUP OR TO A CONVENTIONAL CLASSROOM GROUP. THREE HYPOTHESES WERE TESTED--(1) THERE WILL BE NO DIFFERENCES BETWEEN AUTOMATED TEACHING AND STANDARD INSTRUCTION FOR TEACHING SELECTED TASKS, (2) EFFECTIVE RETENTION OF SKILLS, TAUGHT THROUGH AUTOMATED MEANS, WILL BE DEMONSTRATED BY POST-LEARNING TASKS, AND (3) NO SIGNIFICANT DIFFERENCES WILL BE FOUND BETWEEN AN AUTOMATED PROCEDURE UTILIZING A MULTIPLE-CHOICE METHOD AND ONE UTILIZING A TYPEWRITER KEYBOARD METHOD. HYPOTHESIS ONE WAS REJECTED IN LIGHT OF SIGNIFICANT SKILL IMPROVEMENT BY SUBJECTS USING BOTH AUTOMATED TEACHING PROCEDURES OVER CONVENTIONAL METHODS. HYPOTHESIS TWO WAS PARTIALLY SUPPORTED. SPELLING IMPROVEMENT WAS RETAINED OVER A RELATIVELY SHORT PERIOD. HYPOTHESIS THREE WAS PARTIALLY SUPPORTED. IN VIEW OF THE FINDINGS, FURTHER RESEARCH CONCERNING AUTOMATED TEACHING WITH THE RETARDED WAS RECOMMENDED. (JC)