for retarded Ss, though it was -.54 for the normal Ss. A detailed analysis of errors indicated that stimulus perseveration errors disappeared early in the training of the retarded Ss, and position errors occurred more frequently in retarded than in normal Ss.
House ( 1960) found that when four objects were presented in different combinations on subsequent discrimination problems, the least difficult change was created by retaining the stimulus which had previously been correct and inserting a new negative stimulus on the subsequent problem. Next in difficulty was an old negative stimulus with a new positive stimulus. The introduction of two new stimuli was of intermediate difficulty. The most difficult changes were, in order, changing the previously negative stimulus to the positive stimulus while introducing a new negative stimulus, and changing the previously positive stimulus to the negative stimulus while introducing a new positive stimulus. It was considered that the deficit shown by retarded Ss in learning-set problems cannot easily be attributed to a failure to suppress negative transfer from previous problems.
The studies that have been reviewed provide representative demonstrations of experimental investigations of learning by retarded Ss. The work in the area is heterogeneous and fragmentary, and it is not possible to reach general conclusions. Only a few simple questions can be answered with surety. Retarded Ss are able to learn discrimination problems of the types presented, and they show evidence of developing learning set; duller Ss have more difficulty than brighter Ss in solving problems; and conditions can be arranged so that learning is facilitated. The studies suggest that experimental investigation of the learning processes of retarded Ss is feasible and that gradually more specific, less obvious, and more useful data will be obtained.
Before it can be concluded that a certain level or type of performance is a result of the intellectual deficit of retarded Ss, consideration must be given to other characteristics of retarded Ss associated with their atypical social experiences and environmental histories. The failure to consider such characteristics restricts the types of interpretations and the generalizations that can be made from experimental work.
Usually, comparisons have been made between the performance of institutionalized retarded and noninstitutionalized normal Ss. It is impossible from these studies to determine the relative effects of institutionalization and retardation on the differences in learning found between the two groups of individuals. The difficulty that retarded Ss have been found to have, especially during the early phases of learning, may result as much from fear and anxiety over being examined as from a basic learning difficulty or an inability to attend to stimuli.
Incentives varying from beads to candy have been chosen arbitrarily for use in the studies. Actually, in some studies different incentives have been used for the normal and the retarded Ss. It is likely that a particular type of incentive may have quite a different effect in motivating retarded Ss and normal Ss. Praise for correct response, for example, may have different