Does “Political Bias” in the DIT or DIT-2 Threaten Validity in Studies of CPAs?

Abstract

The Defining Issues Test (DIT) has been popular among accounting researchers, and will surely find continued applications. The DIT-2 is still new to accounting research. Fisher and Sweeney (1998, 2002) and Sweeney and Fisher (1998, 1999) build upon other critical research to claim that the DIT P score is a preference measure strongly related to political beliefs. Experimental subjects, when asked to respond from an “extremely liberal” perspective or to “identify the statements designed to represent the highest levels of moral judgment,” tend to respond differently than under official DIT instructions. However, Rest et al. (1999) reject this approach to validity testing, claiming that appropriate tests should address the variance explained over and above other constructs. Based on random samples of CPAs in public accounting practice (741 taking the DIT in 1995 and 261 taking the DIT-2 in 2001, each of whom also responded to political preference scales) we measure the amount of variance shared by the “ethical” and “political” measures. In our samples, political position explains less than ten percent of the variance in DIT and DIT-2 P scores as well as the new N2 scores; these effect sizes are small, so that the scores may not be seriously threatened by confounding.