Attitude Change from an Implied Threat to Attitudinal Freedom.

College students were told they were to write an essay supporting 1 or the other side on each of S issues, and were led to believe they might be able to influence the decision about which side of an attiudinal issue they were to support. In all cases they were then told to support the side (either pro or con) they initially preferred, and postmeasures were then obtained before the actual writing. Those who were given the impression that their preference was taken into account in the decision regarding which side they would support on the 1st issue showed attitude change favoring the preferred position (i.e., moving toward greater extremity, while retaining their initial polarity). Others, who were given the impression the decision was made without regard to their preference, tended to show attitude change away from the preferred position they were told to support. Within this latter group, those who expected to be told on which side to write on all S issues showed significantly greater change away than did those who expected to be told which side to support on only the 1st of the 5 issues. The existing literature on social influence and attitude change has been primarily concerned with accounting for varying degrees of positive change. There has been some discussion of factors which might lead to resistance to persuasion and to boomerang effects (e.g., Cohen, 1962; Hovland, Janis, & Kelley, 1953; McGuire, 1964), but one can find surprisingly little systematic experimental investigation in this area, recent exceptions being studied by Berscheid (1966) and Abelson and Miller (1967). Indeed, in traditional attitude-change studies the subject's resistance to a persuasive communication is often interpreted as evidence of a weak manipulation, and a boomerang effect would generally be considered an experimental disaster. These interpretations are, of course, usually correct in that the investigators are primarily attempting to operationalize theories which predict differential amounts of positive influence, and they must therefore be able to show reliable amounts of positive attitude change. The purpose of the present study is to test a theory (Brehm, 1966) which predicts active resistance to persuasion and boomerang attitude change.