The Relative Effects of General versus Descriptive Praise on a Card Sorting Task

Abstract

It has frequently been postulated that descriptive praise, which labels the behavior being praised, is superior to general praise, which delivers an accolade without specifying the behavior being praised. Research investigating this postulate is meager. The purpose of this study was to investigate whether in fact descriptive praise is superior to general praise. Fifty fifth- and sixth-grade students from the Lida Lee Tall Center in Towson, Maryland were randomly selected to serve as subjects. Twelve boys and eight girls were randomly assigned to each of two praise conditions (i.e. descriptive praise and general praise) and six boys and four girls were randomly assigned to a control condition. Subjects were seen individually and pretested to ensure they could perform the experimental task. The assigned task was to sort 108 cards by one of three possible sorting methods. The first 54 card sorts served as a baseline to determine the preferred sorting method for each subject. During the final 54 card sorts, subjects in the two praise conditions received either general praise (e.g. "Great") or descriptive praise (e.g. "Great. I like the way you are sorting by shape") on a FR3 schedule for sorting cards by a randomly selected sorting method. Baseline data were collected for the entire 108 card sorts in the control condition. Multivariate analyses of variance were carried out on the extent to which the three groups changed their sorting method from their baseline method and on the extent to which the two praise groups sorted by the method they were reinforced for. The results indicated that the descriptive praise group performed significantly better than both the general praise and control groups. No significant difference emerged between the general praise and control groups. The male and female subjects did not significantly differ in their response to the two praise conditions. These results support the position that descriptive praise is more effective than general praise. It was suggested that the labeling of the behavior being reinforced in descriptive praise increased the informative value of the reinforcer thereby giving subjects in this condition an advantage over the subjects receiving general praise who had to, in effect, guess what response on their part elicited the praise.