Estimating an Evaluation Utilization Model Using Conjoint Measurement and Analysis

The conjoint approach to measurement and analysis is demonstrated in this article through a test of an evaluation utilizationprocess-model that includes two endogenous variables (predicted participation and predicted instrumental utilization). Conjoint measurement involves having respondents rate attribute profiles that are analogous to concepts based on cells in a factorial analysis of variance. Such multidimensional ratings result in ecologically valid measurements because respondents examine and react to wholes, rather than to single unidimensional items as in traditional survey research. Statistically, conjoint analysis is a "decompositional" technique in which respondents' overall reactions to profiles (hypothetical situations) are decomposed to determine how much importance is given to attributes (variables used in the profiles) and to levels of the attributes.