Abstract
The fact that there is an illusory correlation between groups of small size and a biased impression toward members was revealed in this study. 106 items consisting of the description of likableness and unlikableness of group members were constructed in study 1. Based on this result, study 2 was administered which was to replicate Hamilton and Giffords' study in 1976, using Japanese subjects. The result of study 2 differed consistently from the Hamilton and Gifford study (1976) in the following important points: i.e., the number of unlikable members in small groups (consisting of 13 members) was consistently overestimated. This unique pattern suggested the existence of an association between small group sizes and negative bias toward group members. To explore this hypothesis further, two stimulus groups were presented in study 3. The two groups were identical in size and composition members, which were drawn from different sizes of population. The results strongly supported the hypothesis in that the proportion of unlikable members was overestimated, and the impression of the stimulus group, which was drawn from smaller population size was relatively unfavorable.
