The present work consists of three papers which focus on contemporary prejudice. The first paper outlines the differences between contemporary prejudice and prior, more blatant conceptualisations of prejudice expression. In the German-speaking part no (published) valide scale of prejudice is present which explicitly measures contemporary prejudice. Therefore, the first paper focuses on the development of a German version of the Racial Argument Scale (RAS, Saucier & Miller, 2003). The scale allows for an indirect measurement of contemporary prejudice. The German version RAS-G was validated with prejudice scales, scales to measure stereotypes and discriminating behavioural intentions. Further, the attitude dimensions Right-Wing-Authoritarianism (RWA, Altemeyer, 1981, 1988, 1996, 1998) and Social Dominance Orientation (SDO, Pratto, Sidanius, Stallworth & Malle, 1994) were included in the validation, because both underlie prejudice (e.g. Altemeyer, 1998; Duriez & Van Hiel, 2002). N = 195 people with different education levels completed a questionnaire. Results supported the convergent validity of the RAS-G. Answers were not biased by social desirability.
The goal of the second study was a reduction of prejudice towards Turks based on cognitive dissonance theory (Festinger, 1957). If participants chose to perform a "pro-Turkish" behaviour voluntarily, dissonance should be aroused. By adding additional consonant cognitions dissonance should be reduced. This is supposed to lead to an attitude change (reduction of prejudice). In order to measure these effects items of the RAS-G as well as the other scales to measure prejudice, stereotypes and discriminating behavioural intentions used in the first paper were aggregated. According to a screening only participants (N = 202) with moderate prejudice level took part in the study. Participants generated arguments for the integration of Turks in Germany and spoke these into a microphone. Because of the screening, this act is supposed to be a counterattitudinal behaviour which should arouse dissonance (at least under certain circumstances). In an experiment we manipulated choice with which the counterattitudinal behaviour is performed (high vs. low). Further, personal threat of the topic was manipulated (high vs. low). In a control group participants generated arguments for a neutral topic (greener cities). High choice and high personal threat of the topic should lead to the arousal of higher dissonance. Dissonance reduction should result in an attitude change (reduction of prejudice) on the explicitly argued position as well as on the generalised measure of prejudice.
However, Cialdini et al. (Cialdini et al., 1995) showed that only people with a high preference for consistency revealed the typical dissonance effects. Hence, we expected a reduction of prejudice as a result of the dissonance induction only for people with high preference for consistency. Our assumptions were supported: People with high PFC had lower generalized prejudice as well as lower prejudice on the explicit topic compared to the control group if they were in the high (vs. low) choice condition or high (vs. low) personal threat condition. The observed effect was still obvious in a posttest about 4 weeks after the dissonance induction.
The personality variable preference for consistency (PFC) was further examined in the third paper. We questioned in how far PFC was associated with prejudice and its underlying attitudes authoritarianism and social dominance orientation. Research showed that authoritarian people tend to simple black-and-white thinking and the disregard of ambivalent, inconsistent information (e.g., Frenkel-Brunswik, 1949; Miller & Rokeach, 1968). Prejudice also is a means to cognitively simplify the world. SDO is defined as the promotion of hierarchies between groups (Sidanius, Pratto & Levin, 1996). The conceptualization of SDO does not show any overlap to the concept of PFC. Therefore, we expected an association between PFC, prejudice and authoritarianism, but not with SDO. The reanalysis of the data from the first study with structural equation models supported our assumptions. Part of the variance of authoritarianism was mediated by PFC. These results are a conceptual expansion of the construct preference for consistency.