Hibbing et al. argue that the association between political attitudes and a wide range of psychological and physiological characteristics reflects elevated levels of negativity bias among political conservatives. We suggest that their account is both too general and too specific. Too general because many of the variables invoked by Hibbing et al. show clearer associations with social than with economic conservatism, sometimes showing no association at all with the latter. And too specific because attitudes outside of the political domain show clear connections to many of the variables cited by Hibbing et al.: Attitudes concerning religion and the structure of family and society show highly comparable results to those observed for social conservatism. Accordingly, negativity bias may be best construed as a predictor not of political opinions specifically but of general attitudes toward change across a range of domains. This conclusion is buttressed by a recent behavioral genetic study which found that, in a sample of twins reared apart, attitudes in the political, social, and religious spheres (the Traditional Moral Values Triad; TMVT; Bouchard Reference Bouchard, Voland and Schiefenhövel2009) were best construed as superficially different representations of a single underlying trait, reflecting traditionalism (Ludeke et al. Reference Ludeke, Johnson and Bouchard2013). The TMVT hypothesis is supported to the extent that the development of attitudes in all three domains can be accounted for by the same mechanisms.

The importance of distinguishing between attitudes toward change (reflected in social conservatism) and attitudes toward equality (reflected in economic conservatism) has been demonstrated elsewhere (e.g., Jost et al. Reference Jost, Federico and Napier2009). Here we highlight how the predictors cited by Hibbing et al. show importantly different connections to these two dimensions, and how findings for the social dimension typically apply to all traits in the TMVT.

The Big Five personality model provides some of the most frequently used predictors of attitude constructs. Although general political conservatism is associated with both low Openness/Intellect and high Conscientiousness (Sibley et al. Reference Sibley, Osborne and Duckitt2012), studies that distinguish between social and economic dimensions report that Openness/Intellect and Conscientiousness are primarily associated with social conservatism, whereas economic conservatism is primarily associated with low Agreeableness (Carney et al. Reference Carney, Jost, Gosling and Potter2008; Hirsh et al. Reference Hirsh, DeYoung, Xu and Peterson2010; Van Hiel & Mervielde Reference Van Hiel and Mervielde2004). This pattern is particularly clear in studies that used social attitude measures related to social and economic conservatism: meta-analysis indicated that Openness/Intellect and Conscientiousness are more strongly associated with Altemeyer's (Reference Altemeyer1996) Right-Wing Authoritarianism (assessing attitudes toward change) than with Pratto et al.s' (Reference Pratto, Sidanius, Stallworth and Malle1994) Social Dominance Orientation (SDO, assessing attitudes toward equality); SDO was primarily correlated with Agreeableness and more modestly with Openness/Intellect (Sibley & Duckitt Reference Sibley and Duckitt2008). Consistent with the TMVT model, meta-analysis demonstrated correlations of religious fundamentalism with Conscientiousness and Openness/Intellect similar to those associated with social conservatism (Saroglou Reference Saroglou2010).

In their values, social conservatives diverge from economic conservatives but converge with authoritarians and religious individuals. Duriez et al. (Reference Duriez, Luyten, Snauwaert and Hutsebaut2002) reported that, although both social and economic conservatism were associated with a tendency to favor the self-enhancing cluster of Schwartz's Values Scale (Reference Schwartz and Zanna1992), only social conservatism was associated with a preference for conservation values; this preference for conservation values is, however, common to both authoritarians and religious individuals (Feather Reference Feather, Maehr and Karabenick2005; Saroglou et al. Reference Saroglou, Delpierre and Dernelle2004).

Similarly, disgust sensitivity appears to be more correlated with social conservatism than economic conservatism and is also correlated with religiousness and authoritarianism (e.g., Haidt et al. Reference Haidt, McCauley and Rozin1994; Hodson & Costello Reference Hodson and Costello2007; Inbar et al. Reference Inbar, Pizarro, Iyer and Haidt2012b; Terrizzi et al. Reference Terrizzi, Shook and Ventis2010). Because evolutionary accounts of disgust have posited that the emotion evolved to prevent infection, countries with elevated infection risk (higher parasite loads) might be expected to exhibit corresponding attitude differences; Hibbing et al. note precisely this effect for conservative religious and social beliefs (Fincher & Thornhill Reference Fincher and Thornhill2012), and a recent study (Murray et al. Reference Murray, Schaller and Suedfeld2013) observed a comparable association between parasite load and cross-national differences in authoritarian attitudes. Even humor preferences are informative: both conservatism and religiousness appear to be negatively correlated with enjoyment of “sick” humor (which includes jokes with morbid, gruesome, or sadistic content; Saroglou & Anciaux Reference Saroglou and Anciaux2004, Wilson & Patterson Reference Wilson and Patterson1969). In Hibbing et al.'s framework, this might indicate that elevated sensitivity among conservatives and religious individuals to the aversive imagery in such jokes interferes with their ability to find humor in them.

Variables related to cognitive function follow the same pattern. Cognitive style measures such as Need for Closure and Need for Cognition correlate moderately with measures of social conservatism, authoritarianism, and religiousness, though they appear to be only modestly or even not at all associated with economic conservatism and SDO (Crowson Reference Crowson2009; Hunsberger et al. Reference Hunsberger, Alisat, Pancer and Pratt1996; Saroglou Reference Saroglou2002; Van Hiel et al. Reference van Hiel, Pandelaere and Duriez2004). Intelligence and education are negatively correlated with political conservatism, authoritarianism, and conventional religiousness (Lewis et al. Reference Lewis, Ritchie and Bates2011; Van Hiel et al. Reference Van Hiel, Onraet and De Pauw2010), whereas the relation of education and intelligence to economic conservatism and SDO appears to be smaller than that observed for the TMVT traits, and possibly even inverted (Heaven et al. Reference Heaven, Ciarrochi and Leeson2011; Johnson & Tamney Reference Johnson and Tamney2001; Kemmelmeier Reference Kemmelmeier2008).

Studies of the physiological and neurological correlates of attitude differences represent a recent but expanding addition to this literature. Consistent with the TMVT account, Hibbing et al. note that the same pattern of neural activity exhibited by conservatives during a task requiring response inhibition was observed among highly religious individuals (Amodio et al. Reference Amodio, Jost, Master and Yee2007; Inzlicht et al. Reference Inzlicht, McGregor, Hirsh and Nash2009). (Significantly, the results of these studies seem to be in conflict with the “negativity bias” hypothesis, as the pattern of neural activity exhibited by conservatives and the highly religious is typically interpreted as an indication of reduced sensitivity to negative information; Shackman et al. Reference Shackman, Salomons, Slagter, Fox, Winter and Davidson2011.)

Although Hibbing et al. have provided a novel and compelling integration of a broad literature, we suggest their proposed mechanism is better suited to account for differences in the trait identified by the TMVT than for political conservatism broadly construed. Future empirical work exploring the origins of political attitude differences could employ a broader range of outcome measures (differentiating between social and economic conservatism, and assessing religiousness, authoritarianism, and traditionalism) to assess this possibility.