A COMPARISON OF FADING, ANTICIPATION, AND OVERLEARNING PROCEDURES ON RETENTION OF COMPUTATION FACTS IN HANDICAPPED CHILDREN (ERRORLESS)

This study examines whether a training history without errors would increase retention in handicapped children. It also examined the effects of overlearning on errorless training. Twenty handicapped children were taught arithmetic facts with an anticipation drill (flashcards) and by a fading method. Overlearning was added to each method. A second sample of ten nonhandicapped children was added, to measure performance of handicapped children against a nonhandicapped standard. Retention was defined by performance on posttests given the day after training in each of four conditions: Fading, Anticipation, Fading with overlearning, Anticipation with overlearning. Pretesting was conducted to determine entry level and the number of facts taught per session. Children were then taught facts in treatment phases, randomly ordered, until a criterion of three correct tests was reached. Overlearning was a repetition of the number of trials to criterion. A maintenance phase of daily worksheets followed training in each condition. Results indicated that children retained facts equally well with the methods and that there was no interaction of the methods with overlearning on immediate posttests. Lack of significant differences between methods lends support to conclusions of researchers that errors in training do not affect retention once a mastery criterion is reached. Within the handicapped sample informal assessment indicated that subgroups showed contrasting learning styles. Ten children appeared to use fading better than anticipation in terms of error rates; of these ten, five children used fading in an errorless manner with scores approximating those of the nonhandicapped children. Seven children appeared to make fewer errors with anticipation than fading, and three children showed no difference between methods. This analysis suggests the possibility that individual differences in response to treatments accounted for different results within the sample. Results of subgroup analysis suggest the importance of future research to examine learning styles in interaction with the methods