Noncognitive Effects on Attitude Formation and Change: Fact or Artifact?

Abstract
Throughout the 1960s and 1970s, there was widespread acceptance of belief‐based models of attitude formation and change. Beginning in the 1980s, a number of theories, models, and approaches began to argue for nonbelief‐based determinants and to reject the notion of a purely cognitive, expectancy‐value or multiattribute basis for attitude. In this article, we empirically demonstrate that many findings that appear to support this latter view may be nothing more than methodological artifacts resulting from the use of inappropriate (i.e., theoretically incorrect, noncorrespondent, or invalid) attitudinal predictors and/or criteria.