Curiosity Reward in Children

Abstract

Recently, considerable interest has centered on curiosity as an incentive, e.g., Butler (1953) with monkeys and Mirtrnan and Terrell (1964) with grade-school Ss found curiosity to be effective in discrimination learning. Brackbill and Jack (1958) mentioned the need for a methodological cookbook for the effectiveness of rewards at different ages. In addition, material rewards and satisfaction of curiosity are so frequently used as incentives for children's learning that comparison is desirable. The present study compared candy, curiosity, and curiosity-plus-manipulation as incentives for grade-school children. Sixty Ss were used in a 3 x 2 X 2 factorial design: 3 reward conditions, 2 sexes, and 2 grades (kindergarcen and first-grade). The task and procedure were similar to those used by Brackbill and Jack (1958) : two 2-lb. and one 1-lb. coffee cans confronted Ss, with each can painted a different shade of red. The child was told that on each trial E would place an object under one of the cans, and that there was a way to find the correct can every single time. For the candy groups, the object was an M&M which S kept whenever he was correct. For the curiosity groups, the object was a different toy on each trial, and after a correct response S looked at the toy for 5 sec. The curiosity-plus-manipulation Ss could pick up the toy and play with it for 5 sec. There were 64 toys (all different and novel, at least to E) presented in a fixed sequence to all Ss. Trials were at 15-sec. intervals. Between trials, S put his head on a pillow to minimize audible or visual cues. The correct response was a simple position alternation sequence, with color, size, and the presence of the small can as irrelevant cues. If S did not reach criterion in 125 trials, a score of 125 was assigned. Analyses of variance for number of trials to eliminate the small can from the response repertoire and for trials to criterion of 10 successive correct responses indicated that kindergarten Ss required more trials than first-grade Ss (P < .05). Neither reward condition, sex, nor any of the interactions approached significance. Probably the task used here was sufficiently interesting and of sufficiently short duration that performance remained at a high level regardless of the reward condition. In a more monotonous task with less built-in interest, type of reward might well be a significant variable.