Social perception : an experimental study
In extensive review of the literature gathered during the past fifteen years under the title of "Social Perception" or "Empathy" revealed little consensus of opinion as to the exact definition of the concept and the nature of mental abilities or processes underlying it. On the basis of this review, a tentative attempt has been made to re-define Social Perception as the process of predicting the attitudes of other persons. The importance of such a process or ability for interpersonal communication has been discussed and its links with the ordinary processes of thinking and perceiving have been demonstrated. It has been argued that Social Perception in this operational sense is an inferential process depending on two main classes of evidence s Evidence provided by the person whose responses are to be predicted and evidence derived from the context of prediction, i.e., from the predictee's larger class-memberships, the kind of response that is to be predicted I and the connotative properties of the medium of expression employed. A consideration of the second type of evidence has given rise to the hypothesis that accuracy of prediction should be largely dependent on the prevalence and cultural value of the response categoly to be predicted. The multiplicity of possible sources of inference in each case has given rise to some serious doubts as to the possibility of finding a general predictive ability except when the common features of the media of prediction or the responses to be predicted predetermine such a generality. Two experiments have been carried out. The first was concerned with the ability of a group of 240 school children between the ages of 13 and 14 years to predict the positive and negative affective and evaluative responses of their classmates towards themselves. It was found that while children acted significantly better than comparable robots in predicting the positive and socially desirable responses of their classmates, their achievement in predicting the negative and socially undesirable attitudes of others was not significantly different from that of a chance model. Similarly, while accuracy scores in predicting either positive or negative responses on different criteria correlated significantly, there was only a slight negative correlation between the two kinds of accuracy. Nor was there any significant correlation between accuracy of predicting others' affective and evaluative responses. In the second experiment, a sample of 106 training college students were asked to predict the responses of the majority of men and women of their cultural sub-groups and those of two specific individuals on a battery of tests covering the areas of social attitudes, values, and personal dispositions. The amount of accuracy was found to vary positively with the degree of prevalence and objectivity and, negatively, with the variance of the response variable concerned. Predictions of own-sex others' responses were highest in terms of accuracy and those of a relatively unknown other's responses lowest. -3- Predictions of a well-known individual's responses and those of the opposite-sex others-in-general occupied the second and third places in terms of accuracy There was little sign of generality in predicting the responses of the same category of others on various dimensions or in predicting the responses of several others on the same dimension or medium of response. Of correlations between the sums of accuracy scores over the four types of prediotees only that for the two individual others was significant at . 05 level. A consideration of the inconsistent correlations obtaining between different types of accuracy and certain personality measures cast farther doubt on the generality of the underlying capacity. Intelligence, for instance, was found to correlate positively and significantly with accuracy in predicting the well-known other's responses but negatively -albeit nonsignificantly- with accuracy in predicting the responses of others-in-general The results, thus, appear to bear out the tentative hypotheses derived from our conceptual analysis.