Attitude change as a function of threat to attitudinal freedom and extent of agreement with a communicator

Subjects who previously expressed either partial or full agreement with an attitudinal position were exposed to a low- or high-threat essay advocating the position. Anonymous post-communication measures of opinion indicated a reactance effect (negative attitude change) only for subjects who (1) read the hrgh-threat essay, and (2) previously expressed complete agreement with the communicator. Theoretical implications are discussed.