The Effect of Using Self-Report Measures in Tests of Attitude—Behavior Relationships

This study examined the accuracy of self-reports and their effect on tests of attitude-behavior relationships. Accuracy was estimated by comparing records of nongame tax-fund donations to self-reports gathered in a mail survey. Theoretical effects of self-report were examined by using path analysis to test a modified Fishbein and Ajzen model. An analysis using actual behavior was compared to an analysis using self-report, and both measures were included in a third path analysis. Self-reports were reasonably accurate, but they produced results different from actual behavior in attitude—behavior tests; self-report had effects on attitude and behavioral intention measures independent of actual behavior. We concluded that self-reports and past behavior should be measured, tested, and modeled separately in examining attitude—behavior relationships.