The purpose of this study was to test Byrne and Clore's hypothesis that at high levels of “effectance motivation” (produced by exposure to a confusing and non-predictable stimulus situation) the similarity-liking effect is lessened because individuals become less sensitive to environmental variation. 240 Ss were randomly assigned to conditions in which (1) the attitude responses of a target person were either variable or constant, (2) the target's attitudes were similar to S‘s own on 17, 50, or 83% of the topics, and (3) Ss were exposed either to a control movie or to a confusing and unpredictable video tape prior to evaluating the target. Results indicated that target variability and similarity significantly affected both perception and liking of the target and that arousal produced by the video tape dampened the similarity-liking effect. Contrary to Byrne and Clore's hypothesis, however, arousal did not reduce sensitivity to environmental variation. It was concluded that these data require revision of the effectance motivation explanation.