Observational Equivalence in Explaining Attitude Change: Have White Racial Attitudes Genuinely Changed?

The forthcoming article “Observational Equivalence in Explaining Attitude Change: Have White Racial Attitudes Genuinely Changed?” by Andrew M. Engelhardt is summarized by the author below. 

Recent empirical work highlights White Americans’ shifting views of racial and ethnic minorities, and especially Black Americans. Further, White Democrats appear unique in the degree to which their attitudes have changed. That these attitudes contribute in important ways to Whites’ political thinking makes understanding these changes critical. 

In this article I call attention to how these observed trends face an important interpretational hurdle: observational equivalence. The same survey toplines that some scholars, myself included, have interpreted as evidence of genuine attitude change are consistent with three additional processes: social desirability, partisan expressive responding, and changing survey measures. This observational equivalence exists even in careful investigations of panel data. Understanding which of these four possibilities best explains observed trends matters because they differ in their substantive and methodological implications for studying White Americans’ racial attitudes. 

To overcome this observational equivalence I take advantage of a framework used to establish the validity of attitude measure comparisons across groups: measurement equivalence. Instead of exploring a specific racial attitude, the framework focuses on the link between the questions comprising a survey measure and the unobserved attitude they capture. Fortunately, the four attitude change explanations offer different testable implications regarding these links between survey item and latent attitude, allowing me to assess which explanation appears best able to account for observed trends. 

I use this framework and a validated approach for testing it to investigate the measurement equivalence of the racial resentment measure using American National Elections Studies data between 2000 and 2016. Racial resentment captures explanations for Black Americans’ social and economic status and features prominently in studies of the political import of White racial attitudes. Consequently, my analyses not only have immediate payoffs for understanding changing White racial attitudes, they also speak to the status of an important measure in public opinion research.  

Across several different tests I find little evidence that social desirability, expressive responding, or measurement change fully explain observed trends in White racial attitudes. While evidence manifests for these explanations, suggesting they may contribute some to observed trends, it is inconsistent and not substantively large enough to change these trends’ interpretations in meaningful ways.
These results have important substantive and methodological implications. Substantively, scholars should interpret changing racial attitudes as genuine. Methodologically, scholars can continue to validly gauge attitudes with explicit measures. Likewise, despite its formulation in a different context, the racial resentment measure still faithfully characterizes White partisans’ opinions. That’s not to say the measure cannot be improved; rather, its status as an important opinion measure need not necessarily change. 

About the Author: Andrew M. Engelhardt is Assistant Professor, Department of Political Science at University of North Carolina at Greensboro. His research “Observational Equivalence in Explaining Attitude Change: Have White Racial Attitudes Genuinely Changed?” is now available in Early View and will appear in a forthcoming issue of the American Journal of Political Science. 

Speak Your Mind



The American Journal of Political Science (AJPS) is the flagship journal of the Midwest Political Science Association and is published by Wiley.

%d bloggers like this: