دوچرخه سوار خوشحال: بررسی ارتباط بین استبداد عمومی و بهزیستن ذهنی

Abstract Although authoritarianism can negatively impact others (e.g., by predicting prejudiced intergroup attitudes), implications for the self are mixed and require clarification. Extending previous research, we examined the association between generalized authoritarianism (GA, indicated by right-wing authoritarianism and social dominance orientation) and subjective well-being (SWB, indicated by positive affect, negative affect, and life satisfaction) by testing simultaneously the general-level association between GA and SWB as well as specific residual associations between GA and SWB components, independent of basic personality dimensions. We observed a significant general-level association between GA and SWB whereby heightened authoritarianism predicted greater SWB. No residual associations were found between specific GA and SWB components. Despite being “bad” for others, generalized authoritarianism may be “good” for the self

1. Introduction In a powerful metaphor, Adorno (1951) likened the authoritarian individual to a cyclist, bowing at the top and kicking at the bottom. Both parts of the cyclist, submissiveness and dominance respectively, are recognized in contemporary conceptualizations of authoritarianism ( Altemeyer, 1998 and Duriez and Van Hiel, 2002). However, whereas the submissive component is typically studied as right-wing authoritarianism (RWA), reflecting the preference for conventionality, submission to authority, and aggression toward norm violators, the dominance component is studied in terms of social dominance orientation (SDO), the preference for social hierarchical and unequal intergroup relations. Together, RWA and SDO account for approximately 50% of the variance in prejudicial attitudes ( Altemeyer, 1998). Such associations increase dramatically when tapping generalized authoritarianism (GA), reflected in the shared variance between RWA and SDO ( Hodson, MacInnis, & Busseri, under review). Beyond these negative interpersonal and intergroup implications, what are the personal implications for those endorsing greater GA? Might this be “bad” for others (e.g., Sibley & Duckitt, 2008), yet “good” for the self (Van Hiel & De Clercq, 2009)? Early theorizing strongly suggested the negative intrapsychic implications of authoritarianism (Adorno, Frenkel-Brunswik, Levinson, & Sanford, 1950). This was downplayed substantially by Allport (1954), however, and has little trace in contemporary theorizing until very recently, where the results are decidedly mixed. To inform this unresolved issue, we examined the association between GA and subjective well-being at both general and component-specific levels, while disentangling potential effects of basic personality dimensions from this association. In arguably the most influential conceptualization of personal well-being, Diener (1984) proposed that an individual’s overall evaluation of his/her life (typically referred to as life satisfaction, or LS), along with the relative frequency of positive versus negative emotional reactions to one’s experiences (i.e., positive versus negative affect, or PA versus NA, respectively), comprise the main components of ‘subjective well-being’ (SWB). Higher SWB is linked to multiple indicators of positive personal, interpersonal, and societal-level functioning (Eid & Larsen, 2008), with SWB likely representing what most laypersons consider “happiness” (Diener, 1984). Surprisingly, however, studies to date concerning the association between authoritarianism and SWB have proven inconclusive, with some noting a negative association (e.g., Peterson & Duncan, 2007), others no association (e.g., Butler, 2000), and others suggesting a positive association (e.g., Van Hiel & De Clercq, 2009). Further, in a recent meta-analysis, Onraet, Van Hiel, and Dhont (2013) observed no significant overall associations between components of authoritarianism (RWA or SDO) and components of SWB (LS, PA, or NA). Such results are surprising for two (opposing) reasons. First, because higher levels of SWB are generally associated with more positive personal, interpersonal, and societal outcomes, greater authoritarianism (commonly associated with negative interpersonal and societal outcomes, see Altemeyer, 1998 and Sibley and Duckitt, 2008) could be associated with lower SWB. Other evidence, however, suggests that components of SWB are related positively to sociopolitical conservatism (e.g., Choma et al., 2009 and Napier and Jost, 2008), which is typically associated positively with RWA and SDO (Altemeyer, 1998 and Jost et al., 2003). This supports the prediction that greater GA could be associated with greater SWB. At present, therefore, the nature of the association between GA and SWB is conceptually and empirically unclear. Critically, all previous studies examining this issue have focused on specific components of authoritarianism in relation to specific components of SWB. Each of these constructs, however, can be conceptualized as a general underlying (latent) tendency. For example, RWA and SDO are typically correlated positively ( Sibley & Duckitt, 2008), and emerging research demonstrates that drawing on their shared variance to estimate an underlying (latent) GA factor is an effective means through which to more fully capture the authoritarianism construct and its relations to other variables of interest (e.g., Hodson et al., under review and Leone et al., in press). Indeed, such an approach provides researchers with an opportunity to simultaneously study associations involving GA, and its specific components (RWA and SDO), with other theoretical constructs. SWB can similarly be conceptualized at the general level. Consistent with the moderate correlations typically observed among its indicators, a prominent approach to studying SWB is as a generalized underlying tendency and modeled as a latent factor manifested in greater LS, greater PA, and less NA (e.g., Arthaud-Day, Rode, Mooney, & Near, 2005). Such an approach provides an opportunity to study simultaneously associations involving SWB and its specific components (e.g., Busseri, Sadava, & DeCourville, 2007). Conceptually, this approach addresses the underlying commonality among SWB components, as well as the potential unique aspects of each component, rather than assuming that LS, PA, and NA function largely as independent indicators ( Busseri & Sadava, 2011). To date, there is no published evidence with all of the necessary components (RWA, SDO, LS, PA, and NA) assessed in the same study to evaluate the general and component-specific associations between GA and SWB. At present, therefore, it remains unknown whether a general-level association exists between GA and SWB, whether this association is positive or negative, and whether any specific components are also related (independent of the general-level factors). Informing such issues will provide valuable insights into the personal implications of GA, potentially complementing or contradicting the established interpersonal and intergroup consequences. In examining these questions, it is also useful to consider the potential influence of basic personality dimensions. Personality reliably predicts components of authoritarianism, with meta-analytic research noting pair-wise associations between (low) openness to experience and RWA, (high) conscientiousness and RWA, (low) agreeableness and SDO, and (low) openness to experience and SDO (Sibley & Duckitt, 2008). Whether personality dimensions predict underlying (latent) GA has yet to be investigated. Personality dimensions also reliably predict the components of SWB, with meta-analytic research noting pair-wise positive associations between extraversion and PA, agreeableness and PA, and neuroticism and NA, as well as a negative association between neuroticism and LS (Steel, Schmidt, & Shultz, 2008). Others have examined personality predictors of a general (latent) SWB factor, focusing only on extraversion and neuroticism, finding that both personality dimensions predict latent SWB (e.g., Vitterso & Nilsen, 2002). We evaluate the association between GA and SWB over and above the simultaneous predictive effects of six HEXACO personality dimensions (honesty/humility, emotionality, extraversion, agreeableness, conscientiousness, and openness to experience), consistent with evidence indicating that these dimensions collectively describe the basic structure of human personality (Ashton & Lee, 2009). This approach is valuable to determine whether associations between GA and SWB exist as a result of – or independent of – individual differences in personality and ultimately will inform whether and how variation in basic traits relate to global evaluations and emotional experiences of one’s life (Diener, 1996). In summary, in the present work we extend previous research by evaluating the association between GA and SWB through examining each construct as a general (latent) underlying tendency, as well as assessing associations between specific components of each construct unaccounted for by the general-level association. Our primary goal is to determine whether GA and SWB are associated at a general level, independent of basic dimensions of personality, as well as whether unique associations exist between specific components of GA and SWB components, independent of personality and the general-level GA-SWB association. Consistent with emerging research employing a GA approach (e.g., Hodson et al., under review and Leone et al., in press), we hypothesized that examining the relationship between GA and SWB at the general level would reveal a stronger association than has been identified to date. This approach could thus aid in clarifying the nature of the association between GA and SWB, while simultaneously informing whether associations exist between specific components of each construct independent of the general-level factors. A secondary goal was to evaluate how basic personality dimensions predict GA and SWB, and whether certain personality dimensions relate to specific components of GA or SWB beyond associations with the general-level constructs. It was predicted that greater openness to experience, a predictor of lower RWA and SDO (Sibley & Duckitt, 2008), would thus predict less GA. Further, it was predicted that greater extraversion and lower emotionality would predict greater latent SWB, consistent with previous work (Vitterso & Nilsen, 2002). It was also predicted that greater conscientiousness would predict lower RWA, that greater agreeableness would predict lower SDO (see Sibley & Duckitt, 2008), and that greater agreeableness would predict greater PA (see Steel et al., 2008). To address both research goals, we employed a novel latent-variable approach that allows us to determine simultaneously (1) whether GA is related to SWB independent of basic personality dimensions, (2) which specific components of GA are related to certain components of SWB independent of personality dimensions and the general-level GA-SWB association, (3) which personality dimensions uniquely predict GA and/or SWB, and (4) whether unique specific-level associations exist between specific personality dimensions and the components of GA or SWB beyond the hypothesized general-level relations.

3. Results Table 1 displays descriptive statistics and correlations among variables. As expected based on previous research (e.g., Onraet et al., 2013), pairwise associations between specific components of GA and SWB were generally small in magnitude. Table 1. Descriptive statistics and correlations among variables. Variable M SD 1 2 3 4 5 6 7 8 9 10 11 Generalized authoritarianism 1. RWA 3.01 .91 .86 .34∗ .19∗ −.01 .17∗ .02 .08 .06 .01 .03 −.31∗ 2. SDO 2.67 .89 .89 −.01 .01 .10 −.25∗ −.06 .02 −.06 −.10 −.29∗ Subjective well-being 3. PA 3.56 .50 .77 −.25∗ .32∗ .12 −.07 .53∗ .17∗ .22∗ .13∗ 4. NA 2.17 .64 .85 −.36∗ −.21∗ .39∗ −.41∗ −.30∗ −.14∗ −.08 5. LS 6.83 1.26 – .18∗ −.09 .35∗ .16∗ .12+ −.05 HEXACO 6. Honesty/humility 3.47 .57 .72 −.05 .03 .40∗ .22∗ .11+ 7. Emotionality 3.65 .62 .80 −.03 −.23∗ .21∗ −.09 8. Extraversion 3.51 .57 .79 .14∗ .14∗ .06 9. Agreeableness 3.16 .62 .79 .08 .02 10. Conscientiousness 3.66 .57 .77 .05 11. Openness 3.16 .71 .81 Notes: N = 237. ∗p < .05. RWA = right-wing authoritarianism; SDO = social dominance orientation, PA = positive affect, NA = negative affect, LS = life satisfaction. For variables measured using multi-item scales, Cronbach’s alpha appears on the diagonal. Table options To evaluate the general-level association between GA and SWB, structural equation modeling was employed. A latent GA factor was specified with loadings from standardized scores for RWA and SDO. Loadings were fixed at 1.0 to ensure equal loadings across indicators. A latent SWB factor was also specified, with loadings from standardized scores from PA, NA, and LS. Loadings were fixed at 1.0, −1.0, and 1.0, respectively, to ensure equal loadings across indicators. A predictive path was estimated from the latent GA factor to the latent SWB factor. The six HEXACO dimensions were specified as correlated, simultaneous predictors of the latent GA and latent SWB factors. Residual variance terms were specified for each latent factor. This baseline model did not demonstrate good fit (χ2 = 78.50, df = 25, p < .000, CFI = .87 RMSEA = .10, p < .001, SRMR = .06). Examining the standardized residual covariance matrix revealed two significant residual associations that were unaccounted for: A negative association between honesty/humility and SDO, and a positive association between emotionality and NA. The structural model was thus modified to include these two residual associations by specifying correlations between honesty/humility and the residual variance in SDO, and between emotionality and the residual variance in NA. The modified model provided good fit (χ2 = 47.30, df = 23, p = .002, CFI = .94, RMSEA = .07, p = .142, SRMR = .04), and no significant residual covariances remained. Results are displayed in Fig. 1. Results from structural equation modeling testing the association between latent ... Fig. 1. Results from structural equation modeling testing the association between latent GA and latent SWB factors, controlling for HEXACO dimensions. GA = generalized authoritarianism, SWB = subjective well-being, SDO = social dominance orientation, RWA = right-wing authoritarianism, PA = positive affect, NA = negative affect, LS = life satisfaction. Standardized values are shown. For presentation clarity, the residual variance terms in the latent factors and their indicators, as well as the covariances among HEXACO dimensions, are not displayed. Also not shown are correlations between honesty/humility and the residual variance in SDO (r = −.28, p < .05), and between emotionality and the residual variance in NA (r = .33, p < .05). ∗p < .05. Figure options Critically, the standardized path coefficient from the latent GA factor to the latent SWB factor was positive and significant (β = .32; see Fig. 1) 1.Of the six HEXACO dimensions, openness significantly predicted latent GA, such that greater openness predicted less GA. In addition, there was a significant negative residual correlation between honesty/humility and SDO. Honesty/humility, extraversion, agreeableness, conscientiousness, and openness were also significant predictors of SWB, such that higher levels of each personality dimension predicted higher latent SWB. Additionally, there was a significant positive residual correlation between emotionality and NA. Notably, when the partial correlation rather than predictive path between latent factors was examined (i.e., the association between the residual variances in latent GA and SWB factors, controlling for the personality dimensions) the correlation was .61 (p < .001). In the baseline model, the standardized path coefficient between latent factors was .33 (p < .01); the partial correlation was .69 (p < .01). When a ‘trimmed’ model was estimated in which the non-significant paths displayed in Fig. 1 were eliminated, results were consistent with those reported above: The standardized path from the latent GA factor to the latent SWB factor was .32, personality variables predicted latent GA and latent SWB as displayed in Fig. 1, and residual associations were similar to those observed in the non-trimmed model (i.e., the correlation between honesty/humility and the residual variance in SDO was −.24, p < .05; the correlation between emotionality and the residual variance in NA was .39, p < .05). The trimmed model also demonstrated good fit (χ2 = 52.23, df = 29, p = .005, CFI = .94, RMSEA = .06, p = .275