Programmed training of visual discriminations: a comparison of techniques

The efficiency of three discrimination training procedures was compared for mentally handicapped children trained on a shape discrimination presented via a microcomputer. Each of three groups received one of the following training procedures; graded stimulus fading; graded prompt fading, a procedure approximating behavior modification prompting techniques; trial-and-error training. The stimulus fading and graded prompting procedures were equivalent in terms of the number of steps and criteria for fading. Results showed the two programmed techniques did not differ significantly, but both were significantly superior to trial-and-error learning in terms of number of children reaching criterion, number of trials to criterion, and number of errors. The training procedures were continued for subjects from each of the three groups for an additional two problems, followed by a trial-and-error test problem. Over the additional training problems the pattern of differences between the groups in number of errors remained significant, however differences between the groups in number of trials to criterion decreased. There were no significant differences in any of the measures on the trial-and-error test problem. Results are discussed in relation to the literature on errorless learning techniques, and the implications for applied training procedures are elucidated.