Preference and similarity: affective and cognitive judgement?

This study is a test of some implications of the distinction between cognitive and affective processes proposed by Zajonc (1980), applied to preference and similarity judgements. Preference and similarity data were obtained for male and female subjects judging male and female cinema actors. Preference judgements were assumed to be an example of primarily affective judgement, similarity was assumed to be mainly cognitive. It was furthermore assumed that men when judging women and women judging men should be more affectively involved than the opposite cases. It was predicted (a) that making preference judgements would be more confident, (b) that such judgements would be more stable, (c) that preference judgements would show a more simple structure than similarity judgements, and (d) that the difference between similarities and preferences with regard to the level of complexity indicated by the ratings should be especially large with large affective involvement. In support of Zajonc's model, preference judgements were given with more confidence by the subjects and they were more stable over time. There was a tendency for similarity spaces to be more complex for affectively involved subjects while the opposite occurred for preference ratings. Women seemed to react more negatively to the similarity task than men did.     

AN OBJECTIVE EVALUATION OF A BEHAVIOR MODELING TRAINING PROGRAM

Past research on the effects of behavior modeling training has rarely focussed on actual changes in job performance. This paper describes a study in which results of a behavior modeling training program for sales representatives were evaluated in relation to effects on the sales performance of the participants. Sales associates selling large appliances, radios, and television sets for a large chain retailer in seven stores in one metropolitan area participated in a behavior modeling training program designed to improve their sales effectiveness. Before and after comparisons were made in their sales records with similar sales associates in seven matched stores who participated in other kinds of sales training during the same period. Sales representatives who received the behavior modeling training increased their sales by an average of 7% during the ensuing six-month period, while their counterparts in the control group stores showed a 3% decrease in average sales. The extra effort entailed in obtaining the needed sales records seemed to be worthwhile in providing convincing evidence of the value of the training.     

Choice frequency and similarity

Abstract.— Within the framework of a Thurstonian approach to social choice frequencies, an assumption of a positive correlation between similarity and correlation predicts smaller standard deviations of utility differences for more similar choice objects. This prediction was supported in two studies of similarity and preference, one concerned with social science disciplines and the other with Swedish political parties. Implications for understanding social choice phenomena in politics and marketing are discussed.     

Common sense and psychological phenomena: A reply to Smedslund

Smedslund's argument that psycho-social reality is revealed by common sense is shown to be based on a confusion between logical inference and psychological induction. Common sense has no privileged access to psychological reality and is affected profoundly by reasoning errors. Its structure is loose, allowing for post hoc "explanations" of everything and clear-cut predictions of nothing.