Effects of ambivalence on attitudes

Antonio R Mladinic, Purdue University

Abstract

This research examined the effects of ambivalence on attitudes in particular, the effects on memory for arguments associated with the attitude, speed of response and polarization of response. In two studies, subjects were presented on a computer screen with a series of 6 arguments about a controversial issue: that of whether or not to inform people about terrorist threats to airline flights. Three of the arguments displayed supported the idea of informing people about terrorist threats, and three arguments were against doing so. After reading the arguments, ambivalence was induced in one condition, and non-ambivalence in the other. Then, subjects were presented several times with the attitude object and asked to evaluate it using unipolar scales. The critical attitude object was embedded in a list of other attitude objects including other social issues. The evaluation and response latencies were recorded. In the second phase of the experiment, subjects were asked to retrieve as many arguments as possible from those presented in the initial message. In addition, the second study also assessed a naturally occurring ambivalence (abortion on demand). One study hypothesis stated that ambivalent subjects would retrieve more arguments associated with the attitude than non-ambivalent subjects. Another hypothesis stated that ambivalence influenced the response latency of an attitude question; specifically, that non-ambivalent subjects would tend to evaluate an attitude faster than ambivalent subjects. It was also stated that ambivalent and non-ambivalent subjects should show shorter response latencies after repeated expression of the attitude. Finally, it was predicted that non-ambivalent subjects would tend to polarize after repeated expression of the attitude, but that ambivalent subjects would not. Findings of this study showed that, consistent with the hypothesis, subjects in the ambivalence condition recalled more arguments and responded faster toward an attitude inquiry than ambivalent subjects. The effect of repeated expression of the attitude over speed of response was also supported. However, contrary to expectations, the polarization effect was found for both ambivalent and non-ambivalent subjects. In addition, it was found that these results applied for the most part to the manipulated ambivalence and not to the naturally occurring ambivalence. The importance of the manipulation of ambivalence and the implications of the obtained results are discussed.

Degree

Ph.D.

Advisors

Eagly, Purdue University.

Subject Area

Social psychology

COinS