Abstract: | Two studies investigated methodological issues arising from the use of the Fishbein-Ajzen expectancy-value model of attitudes. Different expectancy-value models— based on manipulations of scoring procedure (bi-polar vs. uni-polar scoring of beliefs) and combinatorial rule (multiplication vs. addition of value and expectancy components) -were explicitly compared. Although they used different attitudinal items and samples, both studies obtained significantly lower correlations with overall attitude by using the multiplicative/bi-polar belief model than by using multiplicative/uni-polar belief and additive models. Linear transformations of the belief component had a dramatic effect on the multiplicative model, but had no impact on the additive model. The findings are discussed in terms of their practical and theoretical importance. |