A Puzzle of Racial Attitudes: A Measurement Analysis of Racial Attitudes and Policy Indicators

• In Event: 1284 - Section on Racial and Ethnic Minorities Refereed Roundtable Session
  In Refereed Roundtable 60min: Table 14. Meanings and Measures of Racism

Abstract

Since the 1940s, traditional racial prejudice among white Americans has dramatically declined (Schuman, Steeh, Bobo, and Krysan 1997; Bobo, Charles, Krysan, and Simmons 2012). However, in the 1970s, researchers uncovered a new and distinctive dimension of racial prejudice, terming it “symbolic racism” (Sears and Kinder 1971; McConahay and Hough 1976; Kinder and Sears 1981) and “modern racism” (McConahay 1982), and later referring to it as “racial resentment” (Kinder and Sanders 1996). Most scholarly contestations of the conceptualization and measurement of racial resentment have been resolved (Sears and Henry 2005). Recently, however, Carmines, Sniderman, and Easter (2011) raised a new issue. Comparing various models estimated with data from three separate survey years of the ANES, they claimed that the best explanatory model was one in which racial resentment items and racial policy items constituted a single dimension (labelled racial policy attitudes), and traditional prejudice (stereotypes) constituted a separate dimension.
The conclusion that racial resentment is a measure of racial policy attitudes seriously undermines theories arguing that racial prejudice among white Americans is a primary determinant of their continued opposition to racial policies. The purpose of the present paper is to replicate the confirmatory factor analyses designed by Carmines et al. (2011), and test an alternative model using three additional waves of the ANES along with two additional samples that are not part of the ANES program. We find that an alternative model that specifies racial resentment as distinct from racial policy attitudes was a better fit to the data, and additionally we find preliminary evidence that conceptializations of racial policy attitudes as a single dimension using traditional indicators may be inappropriate.

Authors

• J. Micah Roos, Virginia Polytechnic Institute and State University

• Michael Hughes, Virginia Polytechnic Institute and State University

• Ashley Veronica Reichelmann, Virginia Polytechnic Institute and State University