Statistical analyses of the literature on the effects of exposure frequency on changes in evaluative meaning indicate that two previously unrecognized variables are significantly related to the outcome of these previous experiments, while a previously suggested variable is probably not related. The analyses suggest enhanced evaluative meaning is more likely to result from repeated exposure if a brief delay intervenes between exposure and rating than if no delay occurs, and more likely with paralogs, ideographs, and portraits than with abstract visual and auditory patterns. The nature of the exposure sequence (as discussed by A. A. Harrison and R. Crandall) has typically been confounded with rating delay and does not seem to have an effect independent of the delay factor. Two recent theories are used to account for the findings.

The mediating role of learning in the relationship between repeated exposure and affect was explored and supported in three experiments involving a total of 229 undergraduate participants. It was found that both learning and affect measures behaved in essentially the same way as a function of exposure duration (Experiments 1 and 3), serial position (Experiments 1 and 2), rating delay (Experiment 1) and stimulus properties (Experiment 1). These results suggest learning may be intrinsically rewarding and clarify one of the mechanisms involved in the relationship between exposure frequency and affect, extending Berlyne's two-factor theory of the effects of stimulus familiarity.

No abstract

For a time, many social psychologists were excited (if not pleased) with Kelman's (1967) suggestions for the use of role playing as an alternative to deception in research. While role playing usually has clear ethical advantages over deception studies, it lacks the control, realism, and explanatory power of the deception experiment; while role playing sometimes produces results apparently similar to the deception experiment (e.g., Stang, 1974a), the results are not always identical, and sometimes are quite different (e.g., Greenberg, 1967;Darroch and Steiner, 1970;Wicker and Bushweiler, 1970). In conformity research, successful deception has often produced different results than role playing (e.g., Horowitz and Rothschild, 1970;Willis and Willis, 1970), supplying partial information to Ss (Horowitz and Rothschild, 1970; but cf. Gallo, Smith and Mumford, 1973) or unsuc-* The author wishes to thank Monte

No abstract