Two experiments examined the processes by which positive mood influences attitude change under high and low message elaboration conditions. To examine elaboration, Experiment 1 included individuals who differed in their need for cognition, and Experiment 2 manipulated the relevance of the message. In each study, Ss were exposed to a persuasive communication following a positive or neutral mood induction. In both studies, positive mood produced more positive attitudes toward the advocacy, but positive mood influenced the positivity of Ss' thoughts only under high-elaboration conditions. Path analyses showed that positive mood had a direct effect on attitudes in the low-elaboration conditions but influenced attitudes indirectly by modifying the positivity of thoughts in the high-elaboration conditions. Does a positive mood facilitate persuasion? Many of the initial studies addressing this question suggested that manipulations inducing a positive mood either just before or during a persuasive communication increased the likelihood of attitude change (e.g., Biggers & Pryor, 1982;Dribben & Brabender, 1979;Srull, 1983). These "main effect" findings were so consistent that in reviewing the work on positive affect and persuasion, McGuire (1985) concluded that Persuasive impact is greater if the person is in a happy, benevolent mood when the message comes, noshing on peanuts and soda (Janis, Kaye, and Kirschner, 1965), watching a good program (Krugman, 1983), and with pleasant musical background (Galizio and Hendrick, 1972), an appropriately scented other (Baron, 1983), a smile on one's face (Laird, 1974), nodding one's head (Wells and Petty, 1980), or relaxed in posture (Petty, Wells, Heesacker, Brock, and Cacioppo, 1983). (p. 285) Since 1985, however, the picture has become more complicated. In particular, recent studies have found that positive

The role of counterfactuals in judgments of affective reactions to outcomes was examined. Subjects read about individuals who experienced gains or losses as a result of either deciding to take action and make a change or deciding not to take any new action. In addition, the salience of the counterfactual alternative was manipulated. Past results were replicated in the case of negative outcomes: Individuals who lost money on the basis of action were judged as feeling worse than those who lost money on the basis of inaction. This occurred under both high and low salience of the counterfactual. With positive consequences, however, exaggerated affect for outcomes associated with action rather than inaction occurred only when the counterfactual alternative was made highly salient. Implications for the construction and use of counterfactuals are discussed, and a process model is developed on the basis of the data and the proposed conceptualization.

No abstract