Abstract
Nonreward has been found to facilitate subsequent responding in a number of experiments with children whereas failure has consistently resulted in a decrement in performance. Two variables confounding comparisons of nonreward and failure studies are (a) the use of constant interresponse intervals of 5 sec or longer in nonreward studies as compared with the use of 0-sec. intervals in most failure studies and (b) the use of other-blame instructions in nonreward studies and self-blame instructions in failure studies. In the present experiment 18 second- and third-grade children were assigned to each of six conditions formed by the factorial combination of constant interresponse intervals of 0, 4, and 8 sec, and self-blame vs other-blame instructional sets. On each trial the subjects performed a light-switching task followed by a lever-pulling response. They were failed prior to completion of the light-switching task on one-half of the 20 trials. Lever-pulling responses following failure were slower than responses following success in the 0-sec. condition but not in the 4- or 8-sec. condition, and response speeds in the 0-sec. condition on failure trials were slower than speeds on 4- and 8-sec. condition failure trials. Decrements in response speeds following failure were found in both the self-blame and other-blame conditions although the decrement was greater in the former than the latter condition. In addition self-blame subjects made more perseverative responses on the light switching game than other-blame subjects. The results were discussed in terms of the effects of interresponse intervals and instructional set on the extent to which responses produced by failure would occur and interfere with the lever response.