After spending 45 s studying the names of 4 members of a hypothetical group, subjects showed both implicit liking and implicit identification with the group. These effects of studying names were much larger than the mere exposure (R. B. Zajonc, 1968) effects of either 6 (Experiment 2) or 10 (Experiment 3) extra exposures to each name. This implicit partisanship effect differs from the minimal group effect (H. Tajfel, 1970) because its procedure involves no membership in the target group. It

more »

... differs from the mere exposure effect because the target stimuli are presented once as members of a group rather than multiple times as unrelated individuals. A plausible (but not established) interpretation is that the attitude and identification effects are consequences of mere categorization. 1 A standard form of research report tells of theoretical inspiration followed by research efforts that culminate in the report of predicted results. This article tells of unpredicted results followed by research efforts that culminate in an unanticipated theoretical inspiration. This reversed sequence may occur more often than is apparent, possibly because researchers do not readily admit to such glaring lack of foresight. The reason for reporting this work in its actual sequence may be less a compulsion to tell the truth than an indication of the authors' inability to construct, within the standard research-story form, any plausible scenario that could produce the procedure of the central independent variable. 2 The IAT turned out to be successful for this purpose, even though, as will be seen, not in the expected fashion. Recently, Ashburn-Nardo, Voils, and Monteith (2001) have independently demonstrated the usefulness of the IAT as a dependent measure for the minimal group procedure.