People and the Place: Social Dominance Orientation Is Reciprocally Associated With Hierarchy-Enhancing Occupations Over Time

Abstract

Social dominance orientation (SDO) is a widely researched construct that indexes a preference for hierarchical intergroup relations. However, it remains unclear whether this preference (a) motivates people to seek out occupations that enhance hierarchical relations between groups (i.e., occupational assortment), (b) develops as a result of working within hierarchy-enhancing occupations (i.e., occupational socialization), or (c) both. One reason for this gap is that the large-scale longitudinal data required to simultaneously model assortment and socialization processes are scarce. In this study, we analyzed data from two waves of longitudinal data (spaced either 1, 3, or 5 years apart) from a nationwide sample of adults (N = 3,452–4,412) who were already working in either hierarchy-enhancing occupations (e.g., law enforcement) or hierarchy-attenuating occupations (e.g., social work). Results showed that SDO predicted an increased probability of working in a hierarchy-enhancing occupation 3 and 5 years later. Working in a hierarchy-enhancing occupation was also positively associated with SDO after 1 and 5 years. These patterns generally suggest that occupations both shape, and are shaped by, intergroup beliefs.

Keywords

social dominance orientation intergroup relations person-occupation fit socialization longitudinal analysis

The past decade has seen numerous social movements highlight how institutional norms reinforce hierarchies between groups of people. For example, Occupy Wall Street underscored practices in the financial sector that inflated the wealth of the “One Percent,” Black Lives Matter drew attention to racial discrimination in the criminal justice system, and #MeToo raised awareness of gendered exploitation in the workplace (Buckley, 2018; Cobb, 2016; DeLuca, Lawson, & Sun, 2012). Research substantiates their concerns: Police are 3.5 times more likely to use physical force against Black Americans than White Americans (Goff, Lloyd, Geller, Raphael, & Glasser, 2016), and 1 in 4 women report workplace sexual harassment (Equal Employment Opportunity Commission, 2016). Moreover, while millions worldwide face economic hardship due to the ongoing coronavirus pandemic, a recent report found that billionaires had increased their wealth by more than 25% during this same period (UBS/PwC, 2020). In light of these trends, management scholars have called for empirical examination of the institutional factors that may contribute to inequities between social groups (Amis, Munir, Lawrence, Hirsch, & McGahan, 2018; Bapuji, Ertug, & Shaw, 2019; Bapuji, Patel, Ertug, & Allen, 2020).

An extensive line of research shows people’s psychological dispositions often match the characteristics of their occupation (Holland, 1959; for review, see van Vianen, 2018). In recent years, scholars have expanded our understanding of such congruence to incorporate sociopolitical beliefs, including employees’ preference for—and their occupation’s general impact on—societal hierarchies (Haley & Sidanius, 2005). However, such findings raise an intriguing conundrum with respect to whether this alignment occurs because people are sorted into occupations based on their sociopolitical beliefs (Schneider, 1987), whether occupational norms socialize people’s sociopolitical preferences (Cooper-Thomas, Van Vianen, & Anderson, 2004), or through a combination of both processes (e.g., De Cooman et al., 2009).

Researchers using a social dominance theory framework (Sidanius & Pratto, 1999) have examined these questions for almost three decades, but the evidence remains equivocal (Haley & Sidanius, 2005). This literature assesses the degree to which social dominance orientation (SDO; Pratto, Sidanius, Stallworth, & Malle, 1994)—a measure of one’s preference for hierarchical arrangements in society—predicts people’s membership in occupations that foster (i.e., hierarchy-enhancing [HE]) or mitigate (i.e., hierarchy-attenuating [HA]) intergroup hierarchies. Whereas some studies suggest that people seek out, and are selected into, occupations that match their (dis)inclination for intergroup inequality (Pratto & Espinoza, 2001; Pratto, Stallworth, Sidanius, & Siers, 1997; Sidanius, van Laar, Levin, & Sinclair, 2003), others indicate that the occupational environment may be more critical in shaping the attitudes of its members (Dambrun, Kamiejski, Haddadi, & Duarte, 2009; Muheljic & Drace, 2018; Nicol, Charbonneau, & Boies, 2007).

However, these contradictory findings may be due to the disproportionate use of student samples to test these important associations. Aside from their usual limitations (see Bergman & Jean, 2016), student samples are particularly problematic for studying preferences for intergroup hierarchies because universities are generally viewed as HA environments (Sidanius et al., 2003; Sinclair, Sidanius, & Levin, 1998). Exposure to higher education tends to increase egalitarian beliefs, sometimes irrespective of whether students are in HE or HA microenvironments, such as their academic major (e.g., Pascarella, Edison, Nora, Hagedorn, & Terenzini, 1996; Peterson & Lane, 2001). Because the salience of the macroenvironment may neutralize some of the HE tendencies of specific academic majors, there is a clear need to test these associations beyond a university setting (Bergman & Jean, 2016; Weston, Ritchie, Rohrer, & Przybylski, 2019).

A second limitation is a paucity of large-scale longitudinal data, outside of student samples, required to simultaneously model the effects of (a) occupations on individuals and (b) individuals on occupational choice (e.g., Nicol et al., 2007; Sidanius et al., 2003). Although experimental designs are the only way to establish causality definitively, random assignment of occupation type and/or SDO is impossible without proxy manipulations that lack external validity (Rohrer, 2018). Moreover, small-scale studies are often underpowered, yielding parameter estimates that obscure the true effect size (Götz, Gosling, & Rentfrow, in press; van Assen, van Aert, & Wicherts, 2015). Therefore, we test the reciprocal associations between SDO and occupation type using longitudinal data from a large, nationwide sample of adults employed in a range of HE and HA occupations to understand whether SDO is a precursor to—or a product of—the type of occupation people work in. But first, we briefly outline social dominance theory and the critical distinctions between HE and HA occupations.

Social Dominance Theory

Social dominance theory argues that people tend to form intergroup hierarchies in which dominant groups (e.g., people who are wealthy) exert social and economic power over other groups (e.g., people in poverty; Sidanius & Pratto, 1999). These hierarchies produce disparities in the distribution of power and resources between social groups whereby dominant groups accumulate disproportionate advantages (e.g., prestige, wealth, housing) relative to other groups (Sidanius, Cotterill, Sheehy-Skeffington, Kteily, & Carvacho, 2017). As allocators of such resources, institutions, such as occupations, play an integral role in preserving intergroup hierarchies because the aggregate effect of their norms has a greater and more persistent impact on social inequality than does any one person in isolation (Haley & Sidanius, 2005).

However, institutions can have differential impacts on intergroup hierarchy. Social dominance theory defines HE occupations as those that promote dominant groups’ interests by disproportionately allocating them positive status and resources. Although dependent on the cultural milieu, examples of HE occupations include those related to big business, finance, the military, and the criminal justice system (Sidanius, Liu, Shaw, & Pratto, 1994). Conversely, HA occupations aim to redress intergroup hierarchy by providing disadvantaged social groups access to valuable resources. Some examples of HA occupations include social workers, educators, and those who care for people with illness and disability (Haley & Sidanius, 2005). Although these examples reflect general trends, individual members of HE and HA occupations are diverse and can have high, moderate, or low preferences for intergroup inequality. Specifically, individuals can strengthen (e.g., voting against universal health care) or mitigate (e.g., voting for wealth redistribution) social hierarchies irrespective of their occupational subgroup membership. Rather, the distinction between HE and HA occupations reflects their net effects on social inequality, which can be both pervasive and persistent in the aggregate (Pratto, Sidanius, & Levin, 2006).

In conjunction with institutions, social dominance theory proffers an individual-level factor involved in maintaining social inequality, namely, SDO—an individual-difference measure of one’s preference for different groups in society to be organized hierarchically (Pratto et al., 1994). Accordingly, research shows that SDO consistently predicts generalized prejudice (Osborne, Satherley, Little, & Sibley, in press), including sexism (Sibley, Wilson, & Duckitt, 2007), xenophobia (Thomsen et al., 2010), and low empathy for outgroups (Hudson, Cikara, & Sidanius, 2019). Crucially, SDO is also linked to adverse workplace outcomes, including discrimination in hiring and evaluations (Simmons, Watkins, & Umphress, 2015; Umphress, Simmons, Boswell, & Triana, 2008), unethical decision making (Son Hing, Bobocel, Zanna, & McBride, 2007), and a desire for workplace homogeneity based on race and gender (Umphress, Smith-Crowe, Brief, Dietz, & Watkins, 2007) as well as abusive supervision (Khan, Moss, Quratulain, & Hameed, 2018). Put simply, SDO is a powerful predictor of both negative societal and workplace outcomes.

While social dominance theory provides a multilevel account of inequality, it acknowledges that the relationship between institutions and individuals is reciprocal. As such, people’s sociopolitical beliefs should partly reflect the type of occupation to which they belong. Consistent with this thesis, the hierarchy-relevant nature of various types of institutions (such as occupations) correlates with its members’ SDO (see Pratto et al., 2006). For example, Sidanius et al. (1994) demonstrated that police officers (an HE occupation) in Los Angeles had higher SDO than did public defenders (an HA occupation). Likewise, research conducted in France has shown that psychology students (an HA major) have comparatively lower SDO than those in law (an HE major; Dambrun, Guimond, & Duarte, 2002). Accordingly, social dominance theory outlines five nonmutually exclusive mechanisms to explain the emergence and maintenance of these instances of sociopolitical fit (Van Laar, Sidanius, Rabinowitz, & Sinclair, 1999). Because these mechanisms relate to two broad theoretical domains, we collapse them into the following two processes: occupational assortment and occupational socialization.

Occupational Assortment

One way for sociopolitical fit to develop over time is through a recurrent process in which people gravitate toward, and are selected into, occupations that share their characteristics and beliefs and ultimately leave those with values that are incompatible with their own (Schneider, 1987; Van Laar et al., 1999). We refer to these three mechanisms as occupational assortment because they reflect processes whereby people end up in occupations that mirror their beliefs rather than having occupations shape their beliefs. Thus, according to social dominance theory, a person’s SDO should reflect their assortment into HE versus HA occupations.

Prior research shows that SDO positively predicts preference for, and selection into, HE occupations, including law enforcement and business, while negatively predicting attraction to HA occupations, such as human rights advocacy (Pratto & Espinoza, 2001; Pratto et al., 1997; Sidanius et al., 1996; cf. Gatto, Dambrun, Kerbrat, & De Oliveira, 2010). The strongest evidence for assortment came from a five-wave longitudinal study among university students that found that participants with antiegalitarian beliefs were more likely to select HE academic majors and future vocations than those with more egalitarian beliefs (Sidanius et al., 2003). However, this study found no evidence that students’ SDO changed as a function of their academic major. These findings suggest that SDO should predict assortment into HE occupations over time.

Occupational Socialization

Although SDO reflects a relatively stable preference for intergroup hierarchy, it is also malleable to environmental input (Pratto et al., 2006). However, much of the literature has examined how SDO predicts outcomes, without nearly as much research devoted to how situational factors predict SDO. This asymmetry is regrettable, as SDO is theorized as both “cause and effect” (Sidanius et al., 2017: 167)—a variable that influences social inequality through its impact on intergroup hierarchies yet is also affected by context (Duckitt, 2001; Guimond, Dambrun, Michinov, & Duarte, 2003; Kteily, Sidanius, & Levin, 2011).

One of the key situational factors that may influence SDO over time is the institutional norms of an occupation. Occupational socialization refers to the process in which people adopt the norms and values that govern their occupation (Gray & Kish-Gephart, 2013; Van Maanen & Schein, 1977). Although occupations remain an underexamined institution, their affiliations are increasingly overtaking organizational membership as one of the more stable aspects in people’s working lives (see Anteby, Chan, & DiBenigno 2016; Bechky, 2011). In the contemporary labor market, people are becoming more likely to work in multiple organizations within the same occupation. As such, some research shows that workers identify more strongly with their occupation than with their industry (Carnevale, Rose, & Cheah, 2011). Thus, occupations can foster strong norms that guide their members’ behavior. For example, Cohn, Fehr, and Maréchal (2014) have shown that the salience of occupational identity among bank workers increased their dishonesty, which they attributed to the prevailing norms of the participants’ profession within the banking industry.

Similarly, social dominance theory posits that the norms of HE occupations can socialize more antiegalitarian beliefs among its members over time. Indeed, prior research shows that antiegalitarian beliefs tend to be higher for those in HE majors (law) than for those in HA majors (psychology) and that these differences are mediated by perceived social norms (Dambrun et al., 2002; Muheljic & Drace, 2018). Further work by Dambrun et al. (2009) comparing 1st- and 3rd-year undergraduates showed that exposure to HA academic majors lowered participants’ SDO by decreasing their belief in genetic determinism. Finally, Nicol et al. (2007) observed that applicants to the military were unexpectedly lower on SDO than civilians but that military students’ SDO increased during their 4-year training program. Together, these findings suggest that working in an HE occupation is associated with an increase in SDO over time.

The Present Study

In this study, we test the reciprocal associations between SDO and the types of occupations in which people work (i.e., HE or HA occupations). Although prior research provides evidence for both occupational assortment and occupational socialization, results have been inconsistent. Specifically, studies typically use student samples and find evidence for one process but not the other (Nicol et al., 2007; Sidanius et al., 2003). It bears noting, however, that social dominance theory does not conceptualize assortment and socialization as competing processes (Van Laar et al., 1999). Nevertheless, the mixed findings pose interesting theoretical and practical implications about the directionality of the relationship between SDO and occupation type. This is particularly important given that (a) SDO predicts a host of detrimental societal and workplace outcomes and (b) occupational members can allocate valuable resources within their communities (Cohn et al., 2014; Khan et al., 2018; Sibley et al., 2007; Unzueta, Knowles, & Ho, 2012). Therefore, these associations should be examined beyond lab designs and student samples and tested among a broader range of HE and HA occupations as they occur in the population (Bergman & Jean, 2016).

Another consideration is the timing of occupational assortment and occupational socialization, given they may not have temporal symmetry (e.g., Dormann & Griffin, 2015). Indeed, occupational assortment may be a slower process, occurring for reasons that include—but extend beyond—considerations of sociopolitical fit with any given occupation (e.g., Carnevale & Hatak, 2020; Cooper-Thomas & Wright, 2013). For example, people sometimes remain in a job despite value incongruence with their occupation because quitting can be costly, especially if individuals have economic and familial responsibilities to consider (Vogel, Rodell, & Lynch, 2016). In contrast, occupational socialization may occur fairly soon after exposure to a work environment, with some research showing that this can happen within months (Cooper-Thomas et al., 2004). In sum, occupational assortment may take longer intervals, whereas occupational socialization may occur more rapidly.

Using repeated measurements that are too close or too far apart may yield evidence of one process but not the other—a problem evident in prior research. To address this, we test the reciprocal associations between SDO and occupation type using variable time lags of 1, 3, and 5 years. Because changing occupations can incur costs for individuals, we expect SDO to be associated with an increased probability of switching to an HE occupation after longer time lags. In contrast, because socialization occurs more quickly than assortment, working in an HE occupation should positively associate with SDO irrespective of the time lag. To test these hypotheses, we examine data from participants already working in a broad range of HE and HA occupations from a nationwide, longitudinal study using three different temporal intervals. We also control for several demographics and personality traits, which allows us to model the effects of individual differences and intergroup attitudes separately. Our hypotheses for these varying time lags are as follows:

Hypothesis 1: Higher SDO is associated with an increased probability of working in an HE occupation after three years (Hypothesis 1a) and five years (Hypothesis 1b).

Hypothesis 2: Working in an HE occupation will associate with higher SDO than working in an HA occupation after 1 year (Hypothesis 2a), 3 years (Hypothesis 2b), and 5 years (Hypothesis 2c).

Method

Sampling Procedure

This study uses data from Times 5, 6, 8, and 10 of the New Zealand Attitudes and Values Study (NZAVS), a longitudinal panel study that surveys respondents annually on a range of sociopolitical attitudes, personality, and health outcomes. In 2009 (Time 1), invitations to participate in a 20-year panel study were sent to a random sample of adults registered to vote, which is compulsory in New Zealand. This resulted in 6,518 participants, a response rate of 16.6%. One nonrandom booster sample was recruited in 2011 (Time 3) to address sample attrition, and four subsequent random booster samples (in 2012, 2013, 2016, and 2018) have since been conducted to increase and diversify the sample (response rates ranging from 9.2% to 10.6%).

We chose Time 5 as the initial point for our analyses because this measurement occasion yielded a sample large enough to examine HE and HA occupational subgroups afforded by the overall sample frame. We chose Time 6 as the second point of analysis because it is the shortest lag available in the data set to examine socialization processes, as socialization effects can emerge fairly rapidly (e.g., Cooper-Thomas et al., 2004). Given that people’s occupations are highly stable (Statistics New Zealand, 2018), Times 8 and 10 were included because they provide longer time lags needed to identify the potential reciprocal associations between occupational socialization and occupational assortment. More detailed sampling procedures and attrition rates for the NZAVS are documented elsewhere (Satherley et al., 2015; Sibley, 2020), and a complete dictionary of all variables included in the NZAVS is available at http://www.nzavs.auckland.ac.nz.

Participants

We extracted data from the 5,301 participants (M_age = 47.70, SD = 11.91) who were working in either an HE or an HA occupation (see Measures) at our initial point of analysis (i.e., Time 5), which excludes all booster samples from subsequent years. We restricted our analyses to those who responded to the waves in each analysis when testing different temporal lags. This resulted in final sample sizes of 4,412 (Times 5 and 6), 3,849 (Times 5 and 8), and 3,452 (Times 5 and 10). In terms of sample demographics, participants identified as New Zealand European/Pākehā (80.2%), Māori (12.9%), Asian (4.1%), and Pasifika (2.8%), comprising the four largest ethnic groups in New Zealand.

With respect to employment, the majority worked in HA occupations (62.8%). Of the women in the sample (i.e., 72.1% of the total sample), most worked in HA occupations (73.1%), whereas the majority of men in the sample worked in HE occupations (63.8%). However, the gender asymmetry in occupational type was more pronounced among ethnic-majority men compared with ethnic-minority men (i.e., whereas 66.2% of the New Zealand Europeans/Pākehā men in the sample worked in HE occupations, only 51.6% of the Māori, Asian, and Pasifika men in the sample did so). Unsurprisingly, the median income for people in HE occupations was markedly higher ($71,000) than for those in HA occupations ($50,000). Finally, there was a very low base rate of movement in occupation type. Specifically, of those working in HA occupations at Time 5, 20 moved into HE occupations by Time 6, 25 by Time 8, and 45 by Time 10, and only 19 of those working in HE occupations at Time 5 moved into HA occupations by Time 6, 22 by Time 8, and 25 by Time 10.

Measures

Details for the measures included in this study are provided next. In addition to our focal variables, we included a number of control variables that were theoretically relevant to SDO and occupation type. These included right-wing authoritarianism (RWA), age, gender, ethnic-majority status, income, and Big Five personality traits. Although these covariates are not the focus of these analyses, we provide a brief theoretical justification for their inclusion below and also reran the analyses without them as recommended in the literature (Bernerth & Aguinis, 2016). As shown in Table S1 in the online supplementary file, our focal pattern of results replicated when excluding all control variables.

Social dominance orientation was measured using six items from Sidanius and Pratto’s (2001) 16-item SDO₆ scale, including, “It is OK if some groups have more of a chance in life than others” and “We should have increased social equality” (reverse-coded). We randomly selected three pro-trait and three con-trait items from the original scale in order to avoid acquiescence bias (1 = strongly disagree to 7 = strongly agree; SDO_T5, α = .70; SDO_T6, α = .73; SDO_T8, α = .79; SDO_T10, α = .78).

Occupation type was determined by asking participants to report their current occupation. These open-ended responses were then classified according to minor group categories (~100 codes) from the Australian and New Zealand Standard Classification of Occupations (ANZSCO). The ANZSCO classification system excludes those who are unemployed and does not distinguish between full-time and part-time workers. Two independent coders (average agreement = 91.5%) then further categorized occupations as either HE or HA based on definitions provided by social dominance theory and consistent with prior research in this area (e.g., Pratto et al., 1997; Sidanius et al., 2003; Van Laar et al., 1999). A third independent coder resolved any discrepancies. Consistent with previous studies, we excluded any occupation that could not be disambiguated into HE and HA categories (Sidanius et al., 2003).

Occupations were classified as HE when they served the interests of dominant social groups (e.g., high socioeconomic status, socially powerful, ethnic majorities) and included business executives/managers, sales/marketing professionals, financial/insurance workers, real estate agents, legal professionals, defense force/police personnel, and prison/security officers. Conversely, HA occupations served the interests of socially oppressed groups (e.g., socially vulnerable, ethnic minorities, women, children, and people with illnesses/disabilities) and included nurses, health professionals, social/welfare workers, child care workers, education professionals, and volunteers (see Table S2 online). Occupation type was dummy coded with HE occupations reflecting our focal outcome (0 = HA, 1 = HE). To validate the distinction between occupational categories, an independent-samples t test demonstrated that people in HE occupations (M = 2.47, SD = 0.88) had significantly higher levels of SDO than those in HA occupations (M = 2.12, SD = 0.81), p < .001.

RWA was measured with six items from Altemeyer’s (1996) 30-item scale, including “Our country will be destroyed someday if we do not smash the perversions eating away at our moral fibre and traditional beliefs” and “People should pay less attention to the Bible and other old traditional forms of religious guidance, and instead develop their own personal standards of what is moral and immoral” (reverse-scored). We randomly selected three pro-trait and three con-trait items from the original scale to avoid acquiescence bias (1 = strongly disagree to 7 = strongly agree; α = .70). RWA was included as a control variable because of its consistent associations with SDO (Duckitt & Sibley, 2017) and because research suggests the relationship between assortment and SDO can be explained by RWA (Gatto et al., 2010).

Big Five personality was assessed using the Mini International Personality Item Pool (Donnellan, Oswald, Baird, & Lewis, 2006). Because of the complexity of our analyses, we modeled all Big Five personality traits as latent variables with single indicators using the mean of four items rated on a scale from 1 (very inaccurate) to 7 (very accurate). To account for measurement error, each indicator’s residual variance was fixed to (1 – Cronbach’s alpha) times the variance of the observed scale score (Bollen, 1989; Hayduk, 1987). Extraversion was assessed with items including “I am the life of the party” and “I keep in the background” (reverse-scored; α = .75). Agreeableness was assessed by items such as “I sympathize with others’ feelings” and “I am not really interested in others” (reverse-scored; α = .67). Conscientiousness was assessed using items such as “I like order” and “I often forget to put things back in their proper place” (reverse-scored; α = .67). Neuroticism was assessed with items including “I get upset easily” and “I am relaxed most of the time” (reverse-scored; α = .70). Finally, openness to experience was assessed with items such as “I have a vivid imagination” and “I have difficulty understanding abstract ideas” (reverse-scored; α = .70).

We controlled for Big Five personality traits because of their associations with SDO and occupational interests. Specifically, Duckitt (2001) argued that SDO develops in part from personality traits such as low agreeableness. Indeed, meta-analytic data show that SDO is predicted by low agreeableness and, to a lesser extent, openness to experience (Sibley & Duckitt, 2008). These findings are generally echoed in our study, with the correlation matrix in Table 1 showing that SDO is negatively correlated with agreeableness and openness but is also positively associated with conscientiousness. Likewise, Holland (1959) argued that vocational preferences are an expression of personality. For example, a meta-analysis of the personality traits and vocational choice demonstrates that social vocational interests are correlated with agreeableness and extraversion and that the latter trait is also associated with enterprising vocational interests (Larson et al., 2002). Therefore, personality traits may be linked with HE and HA occupations, given these occupational subgroup distinctions largely consist of enterprising and conventional types of work (i.e., HE) and caring professions (i.e., HA). Indeed, Table 1 shows that HE occupations are positively correlated with extraversion but negatively correlated with agreeableness and neuroticism. Accordingly, we include Big Five personality traits to partition out individual traits from intergroup beliefs in our analyses.

Table 1

Descriptive Statistics and Bivariate Correlations for the Studied Variables

Variable	M	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
1. Age_T5	47.70	11.91
2. Gender_T5	0.28	0.45	.09**
3. New Zealand European/Pākehā_T5	0.80	0.40	.11**	.03
4. Income_T5	$65,908	$55,876	.11**	.32**	.09**
5. RWA_T5	3.33	1.11	.10**	−.07**	−.18**	−.10**
6. Extraversion_T5	4.11	1.15	−.08**	−.01	−.02	.06**	−.05**
7. Agreeableness_T5	5.48	0.90	.01	−.25**	.07**	−.11**	−.02	.16**
8. Conscientiousness _T5	5.14	1.01	.05**	−.05**	−.05**	.04*	.12**	.03*	.09**
9. Neuroticism_T5	3.42	1.08	−.16**	−.15**	−.00	−.12**	.00	−.14**	−.03**	−.14**
10. Openness_T5	4.96	1.09	−.07**	.06**	.01	.05**	−.24**	.24**	.24**	−.04**	−.07**
11. HE occupation_T5	0.37	0.48	−.04**	.34**	.06**	.33**	−.06**	.07**	−.18**	.01	−.07**	.00
12. HE occupation_T6	0.34	0.47	−.05**	.35**	.04*	.35**	−.05**	.06**	−.20**	.01	−.06**	−.02	.98**
13. HE occupation_T8	0.33	0.47	−.03	.36**	.07**	.36**	−.06**	.07**	−.19**	.00	−.05**	−.03	.97**	.97**
14. HE occupation_T10	.033	0.47	−.03	.36**	.07**	.35**	−.06**	.04*	−.18**	−.01	−.03	−.02	.94**	.95**	.98**
15. SDO_T5	2.25	0.85	.07**	.16**	−.04*	.11**	.18**	−.00	−.28**	.00	−.01	−.17**	.20**	.21**	.23**	.23**
16. SDO_T6	2.25	0.86	.10**	.18**	−.04*	.12**	.23**	−.01	−.27**	.04*	−.01	−.18**	.19**	.20**	.22**	.19**	.70**
17. SDO_T8	2.28	0.93	.12**	.19**	−.06**	.10**	.26**	−.03	−.27**	.06**	−.01	−.18**	.20**	.21**	.21**	.21**	.67**	.74**
18. SDO_T10	2.23	0.92	.12**	.23**	−.07**	.09**	.23**	−.05**	−.29**	.03	−.02	−.17**	.22**	.21**	.22**	.21**	.64**	.71**	.78**
n			5298	5298	5077	4645	5299	5200	5199	5199	5198	5199	5301	3685	2992	2522	5299	4411	3848	3452
α							.70	.75	.67	.67	.70	.70					.70	.73	.79	.78

Note: Gender, 0 = women, 1 = men; New Zealand European/Pākehā, 0 = ethnic minority, 1 = New Zealand European/Pākehā; HE occupation, 0 = hierarchy attenuating, 1 = hierarchy enhancing. HE = hierarchy enhancing; SDO = social dominance orientation; RWA = right-wing authoritarianism; T = time.

p < .05.

p < .001.

Demographic control variables

Social dominance theory posits that certain groups (e.g., those who are older, ethnic majorities, and men) will be higher on SDO, holding all else equal (Sidanius & Pratto, 2003). As such, we also included gender (0 = women, 1 = men), ethnicity (0 = ethnic minority, 1 = New Zealand Europeans/Pākehā), and age as control variables in our analyses. Given that personal income is markedly higher for HE occupations in our sample (see Participants), it may also foster switches in occupation (i.e., the desire to work in an occupation that allows more opportunity for material resources). Conversely, higher income (and, therefore, being in a higher socioeconomic group) may be a unique predictor of SDO in the same way that other demographics, such as gender, age, and ethnicity, are theorized to be (Sidanius & Pratto, 1999). As such, we also included personal income as a control variable in our analyses.

Results

Preliminary Analyses

Table 1 shows descriptive statistics and bivariate correlations for all studied variables. All preliminary analyses described next were conducted in Mplus Version 8.3 (Muthén & Muthén, 1998-2017) using full information maximum likelihood with robust estimation. Given one of our focal variables, SDO, was composed of both positively and negatively worded items, we first ran a series of eight confirmatory factor analyses outlined by Marsh, Scalas, and Nagengast (2010) on SDO at Times 5, 6, 8, and 10, which accounted for the possibility of bias introduced from wording effects (DiStefano & Motl, 2006; Marsh et al., 2010). This set of analyses considered a range of models testing either (a) correlated uniqueness or (b) latent methods factors for positively and negatively worded items separately or simultaneously. We restricted these initial measurement models to those who responded at each wave (i.e., Times 5, 6, 8, and 10). We evaluated the fit of each model by comparing their relative parsimony as well as a number of fit indices, including the comparative fit index (CFI; ≥.95; L. Hu & Bentler, 1999), the root mean square error of approximation (RMSEA; ≤.08; McDonald & Ho, 2002), and the standardized root mean square residual (SRMR; ≤.08; L. Hu & Bentler, 1999). We also report the overall chi-square (χ²) test statistic, although the large samples used in these analyses render it uninformative due to its sensitivity to trivial deviations in model misspecification (Wang & Wang, 2012).

In Model 1, SDO was treated as a unidimensional construct (i.e., no methods factors), whereas Model 2 treated SDO as two different trait factors, as there is evidence for distinct dominance and egalitarian facets in recent scales (e.g., SDO₇; Ho et al., 2015) that were developed after the NZAVS began measuring SDO. Models 3 to 5 treated SDO as one trait factor but correlated the residuals among both positive and negative items, respectively (Model 3); only positive items (Model 4); or only negative items (Model 5). Models 6 to 8 treated SDO as one trait factor but included both positive and negative latent methods factors, respectively (Model 6); only a positive latent method factor (Model 7); or only a negative latent method factor (Model 8). In Models 6 to 8, congeneric methods factors were allowed to covary (i.e., positive method factor at Time 5 with positive method factor at Time 6 and so forth), but covariances between methods and trait factors were fixed at zero, as were the covariances between positive and negative methods factors. In all the models, we allowed the residuals for congeneric items to correlate across time. Model 6 produced nonpositive definite results, which yielded unreliable parameter estimates. As such, we conducted two neighboring models: one that constrained the covariances between the negative latent methods factors to equality (Model 6a) and one that fixed the covariances between positive and negative methods factors, respectively, to zero (Model 6b).

As shown in Table 2, the fit indices indicated that Model 8—which is commensurate with the correlated trait–correlated methods minus 1 (CT-C[M-1]; Eid, 2000) model—provided the most parsimonious and best-fitting model to these data. In this model, the positively worded items are used as the referent trait, with the negatively worded method factor partialed out. A benefit of the CT-C(M-1) model is that the trait, method, and measurement errors are uncorrelated, which allows the separation of the nonreference method variances and covariances (Geiser et al., 2008). Therefore, we used Model 8 for the remainder of our analyses.

Table 2

Measurement Models for Social Dominance Orientation (SDO) Times 5, 6, 8, and 10

Model	χ²	df	p	CFI	RMSEA	90% CI RMSEA	SRMR
Model 1	1347.37	210	<.001	.949	.043	[.041, .045]	.078
Model 2	327.98	188	<.001	.994	.016	[.013, .019]	.021
Model 3	948.66	186	<.001	.966	.037	[.035, .040]	.055
Model 4	1193.26	198	<.001	.955	.041	[.039, .044]	.076
Model 5	1078.21	198	<.001	.960	.039	[.037, .041]	.058
Model 6a	286.54	179	<.001	.995	.015	[.011, .017]	.015
Model 6b	948.61	186	<.001	.966	.037	[.035, .040]	.055
Model 7	423.74	192	<.001	.990	.020	[.018, .023]	.040
Model 8	321.55	192	<.001	.994	.015	[.012, .018]	.021

Note: N = 2,926. All measurement models were conducted using full information maximum likelihood with robust estimation. Model 1 = SDO as a unidimensional construct; Model 2 = SDO as two trait factors; Model 3 = SDO as one trait factor with correlated residuals for positive and negative items; Model 4 = SDO as one trait factor with correlated residuals for positive items only; Model 5 = SDO as one trait factor with correlated residuals for negative items only; Model 6a = SDO as one trait factor with positive and negative latent methods factors but constraining the covariances between the negative latent methods factors to equality (due to nonconvergence); Model 6b = one trait factor with positive and negative latent methods factors but fixing the covariances between positive/negative latent methods factors to zero (due to nonconvergence); Model 7 = SDO as one trait factor with positive methods factor only; Model 8 = SDO as one trait factor with negative methods factor only. Besides Model 6b, we allowed congeneric methods factors to covary. In all models, we allowed the residuals for congeneric items across time to correlate. CFI = comparative fit index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual.

Measurement Invariance

To investigate the consistency of our assessments over time, we tested for two types of measurement invariance: configural invariance, which indicates that items map onto traits in a similar pattern over time, and metric invariance, which refers to a similar relationship between the latent factor and items across time (Vandenberg & Lance, 2000). Although scalar invariance is important for examining latent mean differences, only configural and metric invariance are needed to test cross-lagged effects (C. Hu & Cheung, 2008; Little, Preacher, Selig, & Card, 2007). only configural and metric invariance are needed to test cross-lagged effects (C. Hu & Cheung, 2008; Little, Preacher, Selig, & Card, 2007). Nevertheless, we include results from testing scalar invariance (i.e., constraining item intercepts to equality over time) in Table S3 in the online supplementary file. When testing configural invariance, we estimated a measurement model in which the same items loaded onto the same latent trait factor at each of the time points used in our analyses. In the metric invariance model, we constrained the factor loadings for congeneric items across time to equality. In both models, we allowed the residuals for the same item across time to covary (Little et al., 2007).

Our results show that the configural invariance model fit these data well, χ²(df) = 321.55(192), p < .001, CFI = .994, RMSEA = .015, 90% CI [.012, .018], SRMR = .021. Given our large sample size (Wang & Wang, 2012), we evaluated each model using Cheung and Rensvold’s (2002) criterion that differences smaller than –.01 in the CFI across sequential models reflect that model fit has not depreciated significantly. Crucially, constraining the factor loadings to equality in the metric invariance model, χ²(df) = 390.39(207), p < .001, CFI = .992, RMSEA = .017, 90% CI [.015, .020], SRMR = .027, did not lead to an appreciable decrement in model fit, ∆χ²_(∆df) = 68.84₍₁₅₎, ∆CFI = –.002, ∆RMSEA = .002, ∆SRMR = .006. Together, these tests show evidence of configural and metric invariance for SDO over time.

Hypotheses Testing

To examine the associations between HE occupations and SDO, we ran a cross-lagged panel analysis—a statistical technique commonly used to investigate the temporal ordering of variables with longitudinal data (see Selig & Little, 2012). Specifically, two waves of data at different time intervals (i.e., 1, 3, and 5 years apart) were used to test a cross-lagged panel model containing both autoregressive and cross-lagged paths. The autoregressive paths were included to examine and account for the stability of each construct over time, whereas the cross-lagged paths were included to examine the bidirectional effects of occupation type and SDO over the same period. To examine the unique effects of occupation type on SDO (and, conversely, SDO on occupation type), we also controlled for the effects of several variables in our analyses. These included RWA, which was modeled as a latent variable that partialed out the negative method factor given the pro-trait and con-trait items (as we did with SDO). However, given the complexity of modeling reciprocal effects between noncommensurate variables (i.e., occupation type was a binary measure, whereas SDO was a continuous measure), the Big Five personality traits were modeled as latent variables with reliability-corrected single indicators. The rest of our control variables, including gender, age, ethnic majority status, and income, were single-item measures and treated as observed variables.

As previously mentioned, we tested our hypotheses using three different time lags (1 year, 3 years, and 5 years), as shown in Figure 1. This is because temporal intervals are likely to play an important factor in predicting occupation type over time (Dormann & Griffin, 2015). Given socialization occurs quickly, we used the shortest interval available in the NZAVS (i.e., 1 year), whereas we used 3-year and 5-year lags to test potential reciprocal effects between occupational socialization and occupational assortment. Thus, we chose the most appropriate time lags based on data availability and the temporal evidence of socialization and assortment effects in the literature.

Figure 1

Cross-Lagged Panel Analysis of the Associations Between Social Dominance Orientation and Hierarchy-Enhancing Occupations at Different Time Lags

We tested our hypotheses using a series of Bayesian structural equation models (BSEMs) employing a Gibbs sampler (Casella & George, 1992) using default settings in Mplus Version 8.3, which includes noninformative (i.e., diffuse) priors (Muthén & Muthén, 1998-2017). We used a BSEM given its computational advantages over other estimation techniques, such as maximum likelihood (ML) and weighted least squares adjusting for means and variances (WLSMV), when modeling categorical outcomes with latent continuous variables (Asparouhov & Muthén, 2010; Muthén, 2010; Muthén, Muthén, & Asparouhov, 2015). Notably, the estimates from BSEM are similar to—but more robust than—those from frequentist approaches that use null-hypothesis testing (i.e., ML; Yuan & MacKinnon, 2009). However, Bayes estimates differ from ML-based estimates by using probits instead of logits to estimate binary outcomes. Similar to ML, Bayes draws from all available data to generate its estimates (Muthén et al., 2015), which is a more optimal way of handling missing data than WLSMV, which uses only pairs of variables to address missing data (Asparouhov & Muthén, 2010). Nevertheless, in large samples, the estimates obtained by Bayes when using diffuse priors should be comparable to ML-based estimates (see Muthén, 2010).

The Bayesian approach produces 95% credible intervals (95% CI), which reflect the upper and lower 2.5 percentiles in a posterior distribution (Muthén, 2010). In addition to 95% CIs, these analyses use one-sided p values based on the posterior distribution to test the direction of an effect, whereby the p value of a positive estimate reflects the proportion of the posterior distribution below zero and the p value of a negative estimate reflects the proportion of the posterior distribution above zero (Marsman & Wagenmakers, 2017; Muthén, 2010). For statistical inferences, we inspected whether the 95% CIs included zero, in conjunction with one-sided p values to understand the probability of the direction of the effects observed (see Zyphur & Oswald, 2015). Finally, BSEMs produce posterior predictive p values to indicate overall model fit, with excellent fit indicated by values close to .50 (Asparouhov & Muthén, 2010). However, recent evidence suggests that these estimates may be negatively biased by large sample sizes (Hoofs, van de Schoot, Jansen, & Kant, 2018). Therefore, although we report these values (see notes underneath Table 3), we focus on the parameter estimates for the paths reported next.

Table 3

Structural Equation Models for Different Time Lags

Variable	Hierarchy-Enhancing (HE) Occupation					Social Dominance Orientation (SDO)
Variable	b	Post. SD	β	95% CI	p	b	Post. SD	β	95% CI	p
One-year interval (Time 5 to Time 6), N = 4,412
HE occupation_T5	4.69	0.16	0.88	[4.37, 4.98]	<.001	0.05	0.03	0.03	[0.01, 0.11]	.018
SDO_T5	0.20	0.12	0.06	[−0.08, 0.41]	.078	0.88	0.03	0.84	[0.82, 0.93]	<.001
Age_T5	−0.01	0.01	−0.04	[−0.02, 0.00]	.043	0.00	0.00	0.02	[0.00, 0.00]	.043
Gender_T5	0.40	0.17	0.07	[0.06, 0.73]	.009	0.07	0.03	0.04	[0.00, 0.13]	.025
New Zealand European/Pākehā_T5	−0.21	0.18	−0.03	[−0.54, 0.11]	.113	0.04	0.03	0.02	[−0.03, 0.10]	.145
RWA_T5	−0.18	0.13	−0.05	[−0.44, 0.05]	.063	0.14	0.03	0.12	[0.09, 0.20]	<.001
Extraversion_T5	0.11	0.09	0.04	[−0.08, 0.28]	.114	0.00	0.02	0.00	[−0.03, 0.03]	.435
Conscientiousness_T5	0.11	0.12	0.03	[−0.11, 0.34]	.189	0.05	0.02	0.05	[0.01, 0.08]	.007
Agreeableness_T5	0.16	0.18	0.04	[−0.17, 0.52]	.176	−0.03	0.03	−0.03	[−0.09, 0.02]	.119
Neuroticism_T5	0.08	0.10	0.03	[−0.10, 0.27]	.191	0.03	0.02	0.03	[−0.01, 0.06]	.091
Openness_T5	−0.16	0.13	−0.06	[−0.39, 0.17]	.133	−0.01	0.02	−0.01	[−0.05, 0.03]	.294
Income_T5	0.03	0.14	0.01	[−0.25, 0.32]	.424	0.02	0.03	0.02	[−0.03, 0.08]	.027
Three-year interval (Time 5 to Time 8), N = 3,849
HE occupation_T5	4.31	0.15	0.82	[4.02, 4.63]	<.001	0.05	0.03	0.03	[−0.01, 0.11]	.061
SDO_T5	0.32	0.14	0.10	[0.05, 0.63]	.010	0.92	0.03	0.78	[0.86, 0.99]	<.001
Age_T5	−0.01	0.01	−0.02	[−0.02, 0.01]	.141	0.00	0.00	0.02	[−0.02, 0.00]	.087
Gender_T5	0.46	0.17	0.08	[0.16, 0.81]	<.001	0.07	0.04	0.04	[0.00, 0.15]	.024
New Zealand European/Pākehā_T5	0.06	0.17	0.01	[−0.28, 0.40]	.368	−0.03	0.04	−0.01	[−0.11, 0.04]	.196
RWA_T5	−0.23	0.12	−0.07	[−0.47, 0.01]	.033	0.20	0.03	0.16	[0.14, 0.26]	<.001
Extraversion_T5	0.16	0.09	0.06	[−0.01, 0.35]	.029	−0.00	0.02	−0.00	[−0.04, 0.03]	.477
Agreeableness_T5	0.34	0.16	0.10	[0.05, 0.71]	.010	−0.05	0.04	−0.04	[−0.12, 0.02]	.076
Openness_T5	−0.21	0.13	−0.07	[−0.43, 0.09]	.073	0.02	0.02	0.02	[−0.03, 0.06]	.217
Conscientiousness_T5	−0.04	0.10	−0.01	[−0.26, 0.16]	.339	0.06	0.02	0.06	[0.02, 0.10]	.004
Neuroticism_T5	0.04	0.09	0.01	[−0.13, 0.22]	.352	0.03	0.02	0.03	[−0.01, 0.07]	.059
Income_T5	0.39	0.16	0.09	[0.04, 0.66]	.015	0.00	0.03	0.00	[−0.05, 0.06]	.435
Five-year interval (Time 5 to Time 10), N = 3,452
HE occupation_T5	3.81	0.15	0.82	[3.55, 4.11]	<.001	0.07	0.03	0.04	[0.01, 0.13]	.016
SDO_T5	0.34	0.13	0.11	[0.10, 0.63]	.001	0.83	0.03	0.74	[0.76, 0.89]	<.001
Age_T5	−0.01	0.01	−0.04	[−0.02, −0.00]	.014	0.00	0.00	0.03	[0.00, 0.01]	.011
Gender_T5	0.19	0.15	0.04	[−0.09, 0.47]	.094	0.19	0.04	0.10	[0.12, 0.26]	<.001
New Zealand European/Pākehā_T5	0.08	0.16	0.01	[−0.23, 0.38]	.315	−0.04	0.04	−0.02	[−0.12, 0.04]	.140
RWA_T5	−0.31	0.14	−0.10	[−0.59, −0.08]	.007	0.19	0.03	0.16	[0.13, 0.25]	<.001
Extraversion_T5	0.01	0.08	0.00	[−0.13, 0.17]	.450	−0.01	0.02	−0.01	[−0.05, 0.03]	.309
Agreeableness_T5	0.15	0.16	0.05	[−0.15, 0.48]	.155	−0.10	0.04	−0.09	[−0.17, −0.03]	.003
Openness_T5	0.04	0.10	0.02	[−0.16, 0.24]	.333	0.03	0.02	0.04	[−0.01, 0.08]	.084
Conscientiousness_T5	−0.01	0.09	−0.00	[−0.19, 0.17]	.453	0.04	0.02	0.04	[0.00, 0.08]	.024
Neuroticism_T5	−0.02	0.08	−0.01	[−0.17, 0.14]	.404	0.01	0.02	0.01	[−0.03, 0.04]	.399
Income_T5	0.18	0.15	0.04	[−0.10, 0.49]	.104	−0.09	0.03	−0.06	[−0.14, −0.03]	.002

Note: Estimated using Bayesian structural equation. Significant effects shown in bold. b = unstandardized coefficients; Post. SD = posterior standard deviation; β = standardized coefficient; 95% CI = unstandardized 95% credible interval; p = one-sided p value. One-year interval model: posterior predictive p values (ppp) < .001; 95% CI for observed and replicated χ² = [3874.37, 4024.04]. Three-year interval model: ppp < .001; 95% CI for observed and replicated χ² = [3402.22, 3560.60]. Five-year interval model: ppp < .001; 95% CI for observed and replicated χ² = [2927.81, 3104.06]. T5 = Time 5; RWA = right-wing authoritarianism.

One-Year Interval (Time 5 to Time 6)

We first examined the possibility that SDO would be associated with an increased probability of working in an HE occupation a year later, although we are not testing a hypothesis for this relationship at this time interval. Table 3 shows that occupation type was highly stable after 1 year (b = 4.69, posterior SD = 0.16, β = 0.88, 95% CI [4.37, 4.98], p < .001).Adjusting for the stability of occupation type, men had a higher probability of working in an HE occupation at Time 6 than did women (b = 0.40, posterior SD = 0.17,β = 0.07, 95% CI [0.06, 0.73], p = .009). However, neither age, ethnicity, RWA, extraversion, agreeableness, openness to experience, conscientiousness, neuroticism, income, nor SDO (b = 0.20, posterior SD = 0.12, β = 0.06, 95% CI [−0.08, 0.41], p = .078) had statistically significant associations with working in an HE occupation at Time 6 (i.e., 1 year later).

We also tested our hypothesis that occupation type would predict SDO after 1 year (Hypothesis 2a). First, our model demonstrated that SDO was highly stable over 1 year (b = 0.88, posterior SD = 0.03, β = 0.84, 95% CI [0.82, 0.93], p < .001). After controlling for this autoregressive pathway, RWA (b = 0.14, posterior SD = 0.03, β = 0.12, 95% CI [0.09, 0.20], p < .001) and conscientiousness (b = 0.05, posterior SD = 0.02, β = 0.05, 95% CI [0.01, 0.08], p = .007) were significantly positively associated with SDO 1 year later. However, neither age, gender, ethnicity, extraversion, agreeableness, openness to experience, neuroticism, nor income had statistically significant associations with SDO 1 year later. Crucially, our results showed that after adjusting for the control variables and the stability of SDO, working in an HE occupation had a small, though still statistically significant, positive cross-lagged effect on SDO (b = 0.05, posterior SD = 0.03, β = 0.03, 95% CI [0.01, 0.11], p = .018). Thus, consistent with Hypothesis 2a, these results reveal an occupational socialization effect whereby working in an HE occupation is associated with higher SDO after 1 year.

Three-Year Interval (Time 5 to Time 8)

We next tested our hypotheses using a wider time interval (i.e., 3 years) and predicted that SDO would be associated with an increased probability of working in an HE occupation after 3 years (Hypothesis 1a) and that working in an HE occupation would be associated with higher SDO over the same time period (Hypothesis 2b). Consistent with the previous model, occupation type was highly stable over 3 years (b = 4.31, posterior SD = 0.15, β = 0.82, 95% CI [4.02, 4.63], p < .001). In terms of the control variables, income (b = 0.39, posterior SD = 0.16, β = 0.09, 95% CI [0.04, 0.66], p = .015), being a man (b = 0.46, posterior SD = 0.17, β = 0.08, 95% CI [0.16, 0.81], p < .001), and agreeableness (b = 0.34, posterior SD = 0.16, β = 0.10, 95% CI [0.05, 0.71], p = .010) were significantly associated with a higher probability of working in an HE occupation at Time 8. After adjusting for our control variables and the stability of occupation type, SDO at Time 5 had a significant positive cross-lagged effect on the probability of working in an HE occupation at Time 8 (b = 0.32, posterior SD = 0.14, β = 0.10, 95% CI [0.05, 0.63], p = .010). Hence, consistent with Hypothesis 1a, these results indicate that SDO is associated with a higher probability of working in an HE occupation after 3 years.

We also tested Hypothesis 2b by examining the reverse pathway. Once again, SDO was highly stable after 3 years (b = 0.92, posterior SD = 0.03, β = 0.78, 95% CI [0.86, 0.99],p < .001). However, only RWA (b = 0.20, posterior SD = 0.03, β = 0.16, 95% CI [0.14, 0.26], p < .001), being a man (b = 0.07, posterior SD = 0.04, β = 0.04, 95% CI [0.00, 0.15], p = .024), and conscientiousness (b = 0.06, posterior SD = 0.02, β = 0.06, 95% CI [0.02, 0.10], p = .004) were positively associated with SDO at Time 8. Thus, contrary to Hypothesis 2b, working in an HE occupation did not have a statistically significant cross-lagged effect on SDO after 3 years (b = 0.05, posterior SD = 0.03, β = 0.03, 95% CI [−0.01, 0.11], p = .061).

Five-Year Interval (Time 5 to Time 10)

Finally, we tested Hypotheses 1b and 2c using a 5-year interval, again predicting reciprocal effects between HE occupations and SDO. Consistent with the previous results, occupation type was stable after 5 years (b = 3.81, posterior SD = 0.15, β = 0.82, 95% CI [3.55, 4.11], p < .001). Both age (b = −0.01, posterior SD = 0.01, β = −0.04, 95% CI [−0.02, −0.00], p = .014) and RWA (b = −0.31, posterior SD = 0.14, β = −0.10, 95% CI [−0.59, −0.08], p = .007) were significantly associated with a decreased probability of working in an HE occupation at Time 10, whereas none of the other control variables were significantly associated with occupation type in the same period. Most importantly, SDO at Time 5 had a significant positive cross-lagged effect on occupation type at Time 10 (b = 0.34, posterior SD = 0.13, β = 0.11, 95% CI [0.10, 0.63], p = .001). Hence, consistent with Hypothesis 1b, these results indicate that SDO is associated with an increased probability of working in an HE occupation after 5 years.

We also tested Hypothesis 2c, that working in an HE occupation predicted higher SDO after 5 years. Once again, SDO was highly stable over this period (b = 0.83, posterior SD = 0.03, β = 0.74, 95% CI [0.76, 0.89], p < .001). RWA (b = 0.19, posterior SD = 0.03, β = 0.16, 95% CI [0.13, 0.25], p < .001), being a man (b = 0.19, posterior SD = 0.04, β = 0.10, 95% CI [0.12, 0.26], p < .001), age (b = 0.00, posterior SD = 0.00, β = 0.03, 95% CI [0.00, 0.01], p = .011), agreeableness (b = −0.10, posterior SD = 0.04, β = −0.09, 95% CI [−0.17, −0.03], p = .003), conscientiousness (b = 0.04, posterior SD = 0.02, β = 0.04, 95% CI [0.00, 0.08], p = .024), and income (b = −0.09, posterior SD = 0.03, β = −0.06, 95% CI [−0.14, −0.03], p = .002) were significantly associated with SDO 5 years later. None of the other control variables were significantly associated with SDO at Time 10. After adjusting for these control variables and the stability of SDO, our results showed that working in an HE occupation had a significant positive cross-lagged effect on SDO 5 years later (b = 0.07, posterior SD = 0.03, β = 0.04, 95% CI [0.01, 0.13], p = .016), which was consistent with Hypothesis 2c.

Taken together, our results show that (a) working in an HE occupation was associated with higher SDO after 1 year (consistent with Hypothesis 2a), (b) SDO was associated with a higher probability of working in an HE occupation after 3 years (consistent with Hypothesis 1a), and (c) there were reciprocal associations between SDO and HE occupations over a 5-year period (consistent with Hypotheses 1b and 2c). We did not, however, find support for Hypothesis H2b, as working in an HE occupation was not significantly associated with SDO three years later.

Discussion

In this study, we examined the longitudinal associations between SDO and occupation type using data from a nationwide sample of adults who were members of HE or HA occupations in New Zealand. Consistent with social dominance theory, our results suggest that sociopolitical fit occurs through reciprocal relationships between occupational assortment and occupational socialization. However, these two processes depend on specific time intervals. As expected, our findings indicate that occupational socialization may develop faster than occupational assortment, as working in an HE occupation was associated with higher SDO after only 1 year compared with working in an HA occupation. As expected, we found no support for higher SDO resulting in an increased chance of working in an HE occupation after 1 year. On the other hand, although we anticipated bidirectional associations after a 3-year interval, we found evidence only for occupational assortment during this period. That is, SDO predicted a higher probability of working in an HE occupation, but working in an HE occupation was not associated with higher SDO after 3 years. Finally, we found evidence of a reciprocal association between SDO and occupation type over a longer period of 5 years. Table S4 in the online supplementary file shows the predicted probability of switching occupation types. More specifically, irrespective of gender, the probability of switching from an HA occupation to an HE occupation 3 or 5 years later roughly doubled with each standard deviation increase in SDO, and these effects were approximately twice as strong for the men in our sample after 3 years. Likewise, higher SDO was associated with a decreased probability of switching from an HE to an HA occupation after 3 and 5 years. Furthermore, working in an HE occupation was also associated with higher SDO 1 and 5 years later, with Cohen’s d effect sizes of 0.06 and 0.08, respectively (Table S4 online). Although one of our subhypotheses was not supported (namely, Hypothesis 2b), the broader pattern of our results indicates that (a) people are sorted into occupations that match their preferences for intergroup hierarchies over medium-to-long periods and (b) the norms of an HE occupation may socialize antiegalitarian beliefs among its members over both short and long periods.

Theoretical and Practical Implications

This study adds to the literature by providing the first empirical test of the longitudinal associations between SDO and occupation type using a nationwide sample of adults working in a wide range of HE and HA occupations. Although the pioneering work by Holland (1959) argues that occupations may be an expression of people’s personality or values, the present findings demonstrate that fit may also be a product of people’s sociopolitical attitudes, including their intergroup beliefs (Haley & Sidanius, 2005). These findings extend our understanding of how sociopolitical fit occurs between people and their occupations by demonstrating both occupational assortment and occupational socialization among adults already in the course of their careers. Given that occupations are aggregates of individuals, these associations may be largely recursive (Holland, 1959; Schneider, 1987). That is, people are attracted to, and selected into, HE occupations partly based on their SDO. However, occupational directives and norms are usually preexisting structures. As such, people may also be socialized by institutional norms, experiencing a gradual drift toward higher antiegalitarianism throughout their tenure. Irrespective of which process comes first, this cycle may become self-perpetuating over time (e.g., Schneider, 1987).

The present findings also have important practical implications. The bidirectional nature of the relationships between SDO and HE occupations should concern practitioners who seek to create workplaces that redress social inequality, which is becoming increasingly important in public discourse and among institutions seeking to mitigate reputational harm (Bapuji et al., 2020). Notably, management scholars have long recognized that SDO is linked to a number of detrimental work outcomes, including unethical decision making (Son Hing et al., 2007), hiring discrimination (Umphress et al., 2008), and abusive supervision (Khan et al., 2018). Moreover, occupational entities not only have the potential to influence social hierarchies within their ranks (e.g., hiring discrimination that creates occupational segregation; Pratto et al., 1997) but also for the customer bases they serve (King, Shapiro, Hebl, Singletary, & Turner, 2006). Some examples of these include racial, ethnic, and gender biases in mortgage lending (e.g., Houkamau & Sibley, 2015; Robinson, 2002), insurance criteria and claims (e.g., Bradford, Reisner, Honnold, & Xavier, 2013), and rental leases (e.g., Carpusor & Loges, 2006). Put simply, occupations provide a platform to either enact or challenge antiegalitarian beliefs and actions.

In light of our findings, practitioners should endeavor to tackle antiegalitarian norms and behaviors within their workplaces. Although we found that SDO was significantly associated with increases in the probability of working in an HE occupation after 3 and 5 years (i.e., an occupational assortment effect), only a small number of people switched occupation type over time. However, a small number of individuals can, and do, have an outsized negative (or positive) effect on occupational norms (e.g., “bad apples” vs. “bad barrels”; Dunlop & Lee, 2004; Kish-Gephart, Harrison, & Treviño, 2010), especially if they are in positions of leadership (Cialdini, Li, Samper, & Wellman, 2019). Indeed, recent research shows that most instances of discrimination on university campuses originate from only a small number of students, even though this has a corrosive effect on the overall campus environment (Campbell & Brauer, in press). Thus, when making hiring decisions, practitioners may want to consider individuals lower on SDO, particularly if they apply for positions where they wield influence over others in the workplace.

The low base rate at which people changed their occupation type also means that occupational environments have more opportunities to socialize (anti)egalitarian beliefs among their members. This may become more pronounced as the ongoing coronavirus pandemic limits people’s occupational choices and, as a result, may undermine their ability to choose work environments that resonate with their sociopolitical beliefs (Carnevale & Hatak, 2020). Given our findings of occupational socialization at both shorter and longer intervals (1 and 5 years, respectively), tackling the aggregate effects of occupational socialization is more likely to make a substantive difference than would screening out individuals who are high on SDO (Haley & Sidanius, 2005). Evidence for viable solutions shows that simple directives to stick to fair decision criteria may lessen the impact of SDO on, for example, selection discrimination (Umphress et al., 2008). Extrapolating from this point, occupations should develop processes and structures to ensure they treat both colleagues and wider society in a fair manner. This is particularly important for institutions that rely on public and consumer trust to achieve their mission effectively (Tyler, 2006). However, more research investigating ways to combat antiegalitarian norms and behaviors in the workplace is needed to develop targeted interventions to this problem.

Finally, the nature of these occupational distinctions is not static and can change over time (albeit incrementally) with concerted efforts. An example of this is the education sector: In Western societies just over a century ago, education was arguably an institution that enhanced (rather than mitigated) disparities between social groups, as those who were lower socioeconomic status, ethnic minorities, and women were generally excluded from institutions of higher learning (Lucas, 2006). Although there are still issues of accessibility in this sector today, higher education is more open to different groups, especially in societies where it is financially subsidized (e.g., in New Zealand, where the current study is situated) or free under certain citizenship conditions (e.g., Denmark). The good news is that the occupational socialization effect found in this study implies that incumbents will also be responsive to more egalitarian occupational norms (i.e., decrease in SDO over time). Therefore, with the efforts of practitioners, norms can transform an occupation from one that enhances inequities between social groups to one that actively mitigates these differences.

Limitations and Future Research

The use of a large, national sample that examines both HE and HA occupations lends high external validity to our study. Also, that the associations between SDO and occupation type emerged with both the inclusion of crucial covariates (age, gender, ethnicity, income, RWA, and Big Five personality traits) and their exclusion suggests our results are not affected by systematic variability (e.g., Weston et al., 2019). Despite these strengths, there are some notable limitations to our study. First, our measures capture only broad occupational assortment and occupational socialization processes. Specifically, we could not extract the unique occurrences of self-selection, occupational selection, or differential attrition in these data (Van Laar et al., 1999). Nor could we identify the particular factors that contribute to occupational socialization, including the SDO of peers or supervisors, incentive structures, or different levels of abstraction, such as the organizational structure (e.g., Bauer & Green, 1998; Kim, Cable, & Kim, 2005). However, our study provides the foundations upon which future research can build to assess the intra-institutional factors that facilitate sociopolitical fit.

Second, though unsurprising given the labor market in New Zealand (Statistics New Zealand, 2018), occupation type was highly stable among participants, as evidenced by the low base rate of switching between HE and HA occupations. This may have made it difficult to detect occupational assortment. Indeed, meta-analytic data showing weak-to-modest correlations between misfit and turnover suggests that people will often stay with their institutions even when mismatched (Verquer, Beehr, & Wagner, 2003), as there are myriad reasons for why people remain in an occupation (e.g., Cooper-Thomas & Wright, 2013). That we still found evidence of occupational assortment in our sample is compelling, given that a mismatch between people and their occupation is likely a common phenomenon that does not necessarily result in turnover (Vogel et al., 2016). Indeed, our data suggest that SDO is a key intergroup belief that predicts working in occupations that resonate with people’s orientations toward hierarchy between groups. However, given that the effect of misfit is relatively underresearched (Sidanius et al., 2017), these are important questions that should be addressed in future studies.

Third, it should be noted that we extracted a sample of participants who were already working in either HE or HA occupations. As such, the results of this study cannot generalize to initial occupational assortment or occupational socialization, for example, among students graduating from university and moving into their first career occupation. This may be important, as sociopolitical beliefs, such as SDO, may not foster occupational fit in initial vocation choice, although prior research among university students’ professional preferences provides some evidence that it does play a role (e.g., Sidanius et al., 2003). Nevertheless, age may be an important moderator of the associations between SDO and occupation type. According to the impressionable-years hypothesis (e.g., Osborne, Sears, & Valentino, 2011), people’s sociopolitical attitudes are most malleable in late adolescence and early adulthood. This suggests that the occupational socialization effects we identify here may be stronger for younger participants. Likewise, as people age and their economic and familial responsibilities increase, other aspects of an occupation, including its flexibility, remuneration, and location, may become more integral to occupational assortment than a general preference for social hierarchy. However, because sociopolitical attitudes crystallize as people age (e.g., Henry & Sears, 2009), socialization effects may conversely be stronger among older workers who have stayed in their occupations. Given these conflicting possibilities, future research will need to examine the potential moderating effects of age on the associations reported in the current study.

Fourth, the effect sizes for our focal relationships of interest are modest. Yet, methodologists have argued that small effect sizes derived from large samples—especially when controlling for other variables—should be normalized, given that complex phenomena are usually multidetermined (Adachi & Willoughby, 2015; Götz et al., in press). Moreover, small margins can ultimately yield large differences in outcomes. For example, even a handful of individuals high in SDO within an occupational environment may be disruptive to their colleagues (e.g., Khan et al., 2018). Furthermore, given that occupations are governed by their own norms (Cohn et al., 2014), they have the potential not only to impact social disparities within their ranks but also to affect the way they deploy resources to wider society (Vallas & Cummins, 2014). If SDO becomes gradually elevated over the tenure of working in an HE occupation, incumbents may begin making resourcing decisions that differentially discriminate against groups of people (e.g., Umphress et al., 2007). As such, small effect sizes can be socially meaningful when aggregated over large populations, particularly if they also accumulate over time (Greenwald, Banaji, & Nosek, 2015; Ofosu, Chambers, Chen, & Hehman, 2019). Although subtle, the cumulative effects of even small acts can perpetuate intergroup inequality in profound ways.

Finally, although longitudinal data are critical for establishing the temporal order of variables that are difficult to manipulate, causal inferences from a single study based on survey data should be made with caution. However, we believe that using survey data outside of student samples and artificial lab environments is a notable strength of the study because it is unfeasible to pursue randomized experiments based on SDO and occupation type without using proxy manipulations with low external validity (Rohrer, 2018). Nevertheless, consensus on these questions will be best served by evidence accrued through diverse research methods and over long periods (Weston et al., 2019). By using longitudinal data to test the bidirectional relationships between SDO and occupation type over time, we view our study as one integral piece of a larger literature.

Conclusion

Social dominance theory argues that people’s SDO should correspond to the hierarchy-relevant nature of their occupations and that this sociopolitical fit occurs through both occupational assortment and occupational socialization. To date, no study has examined these associations among a range of occupations that use longitudinal data from a nationwide sample of adults already in the course of their careers. We addressed these gaps by using data from a longitudinal, nationwide, random sample of adults to test whether SDO predicted increases in the probability of working in an HE occupation over 1, 3, and 5 years and, conversely, whether working in an HE occupation resulted in higher SDO over these same periods. We found support for a reciprocal relationship between occupational assortment and occupational socialization, though occupational assortment tended to emerge over longer intervals. Collectively, our results demonstrate the power of occupations to shape—and be shaped by—intergroup beliefs, such as SDO.

Supplemental Material

sj-docx-1-jom-10.1177_01492063211004993 – Supplemental material for People and the Place: Social Dominance Orientation Is Reciprocally Associated With Hierarchy-Enhancing Occupations Over Time

Supplemental material, sj-docx-1-jom-10.1177_01492063211004993 for People and the Place: Social Dominance Orientation Is Reciprocally Associated With Hierarchy-Enhancing Occupations Over Time by Elena Zubielevitch, Gordon W. Cheung, Chris G. Sibley, Nikhil Sengupta and Danny Osborne in Journal of Management

Footnotes

Acknowledgements

Preparation of this manuscript was supported by a University of Auckland Doctoral Scholarship awarded to the first author, a Templeton Religion Trust grant (No. 0196) awarded to the third author, a University of Auckland FRDF grant (No. 3709123) awarded to the last author, and PBRF grants jointly awarded to the third and last authors.

Authors’ note:

The syntax for the analyses reported in this article are available on the New Zealand Attitudes and Values Study (NZAVS) website: . A deidentified data set containing the variables analyzed in this manuscript is available upon request from Chris Sibley or any member of the NZAVS advisory board for the purposes of replicating the analyses reported here.

ORCID iDs

Elena Zubielevitch

Gordon W. Cheung

Supplemental material for this article is available with the manuscript on the JOM website.

References

Adachi

Willoughby

2015. Interpreting effect sizes when controlling for stability effects in longitudinal autoregressive models: Implications for psychological science. European Journal of Developmental Psychology, 12: 116-128.

Altemeyer

1996. The authoritarian specter. Cambridge, MA: Harvard University Press.

Amis

J. M.

Munir

K. A.

Lawrence

T. B.

Hirsch

McGahan

2018. Inequality, institutions and organizations. Organization Studies, 39: 1131-1152.

Anteby

Chan

C. K.

DiBenigno

2016. Three lenses on occupations and professions in organizations: Becoming, doing, and relating. Academy of Management Annals, 10: 183-244.

Asparouhov

Muthén

2010. Bayesian analysis using Mplus: Technical implementation. Retrieved from https://www.statmodel.com/download/Bayes3.pdf.

Bapuji

Ertug

Shaw

J. D.

2019. Organizations and societal economic inequality: A review and way forward. Academy of Management Annals, 14: 60-91.

Bapuji

Patel

Ertug

Allen

D. G.

2020. Corona crisis and inequality: Why management research needs a societal turn. Journal of Management, 46: 1205-1222.

Bauer

T. N.

Green

S. G.

1998. Testing the combined effects of newcomer information seeking and manager behavior on socialization. Journal of Applied Psychology, 83: 72-83.

Bechky

B. A.

2011. Making organizational theory work: Institutions, occupations, and negotiated orders. Organization Science, 22: 1157-1167.

10.

Bergman

M. E.

Jean

V. A.

2016. Where have all the “workers” gone? A critical analysis of the unrepresentativeness of our samples relative to the labor market in the industrial-organizational psychology literature. Industrial and Organizational Psychology, 9: 84-113.

11.

Bernerth

J. B.

Aguinis

(2016). A critical review and best-practice recommendations for control variable usage. Personnel Psychology, 69: 229-283.

12.

Bollen

K. A.

1989. Structural equations with latent variables. New York: Wiley.

13.

Bradford

Reisner

S. L.

Honnold

J. A.

Xavier

2013. Experiences of transgender-related discrimination and implications for health: Results from the Virginia Transgender Health Initiative Study. American Journal of Public Health, 103: 1820-1829.

14.

Buckley

2018. Powerful Hollywood women unveil anti-harassment action plan. New York Times, January 1. Retrieved from https://www.nytimes.com/2018/01/01/movies/times-up-hollywood-women-sexual-harassment.html.

15.

Campbell

M. R.

Brauer

. in press. Is discrimination widespread? Testing assumptions about bias on a university campus. Journal of Experimental Psychology: General. https://doi.org/10.1037/xge0000983

16.

Carnevale

A. P.

Rose

S. J.

Cheah

. 2011. The college payoff: Education, occupations, lifetime earnings. Washington, DC: Georgetown University Center on Education and the Workforce.

17.

Carnevale

J. B.

Hatak

2020. Employee adjustment and well-being in the era of COVID-19: Implications for human resource management. Journal of Business Research, 116: 183-187.

18.

Carpusor

A. G.

Loges

W. E.

2006. Rental discrimination and ethnicity in names. Journal of Applied Social Psychology, 36: 934-952. https://doi.org/10.1111/j.0021-9029.2006.00050.x.

19.

Casella

George

E. I.

1992. Explaining the Gibbs sampler. American Statistician, 46: 167-174.

20.

Cheung

G. W.

Rensvold

R. B.

2002. Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9: 233-255.

21.

Cialdini

Y. J.

Samper

Wellman

2019. How bad apples promote bad barrels: Unethical leader behavior and the selective attrition effect. Journal of Business Ethics, 168: 1-20.

22.

Cobb

2016. The matter of Black lives: A new kind of movement has found its moment. What will its future be? New Yorker, March 14. Retrieved from https://www.newyorker.com/magazine/2016/03/14/where-is-black-lives-matter-headed.

23.

Cohn

Fehr

Maréchal

M. A.

2014. Business culture and dishonesty in the banking industry. Nature, 516: 86-89.

24.

Cooper-Thomas

H. D.

Van Vianen

Anderson

2004. Changes in person–organization fit: The impact of socialization tactics on perceived and actual P-O fit. European Journal of Work and Organizational Psychology, 13: 52-78.

25.

Cooper-Thomas

H. D.

Wright

2013. Person-environment misfit: The neglected role of social context. Journal of Managerial Psychology, 28: 21-37.

26.

Dambrun

Guimond

Duarte

2002. The impact of hierarchy-enhancing vs. attenuating academic major on stereotyping: The mediating role of perceived social norm. Current Research in Social Psychology, 7: 114-136.

27.

Dambrun

Kamiejski

Haddadi

Duarte

2009. Why does social dominance orientation decrease with university exposure to the social sciences? The impact of institutional socialization and the mediating role of “geneticism.” European Journal of Social Psychology, 39: 88-100.

28.

De Cooman

De Gieter

Pepermans

Hermans

Du Bois

Caers

Jegers

. (2009). Person–organization fit: Testing socialization and attraction–selection–attrition hypotheses. Journal of Vocational Behavior, 74: 102-107.

29.

DeLuca

K. M.

Lawson

Sun

2012. Occupy Wall Street on the public screens of social media: The many framings of the birth of a protest movement. Communication, Culture & Critique, 5: 483-509.

30.

DiStefano

Motl

R. W.

2006. Further investigating method effects associated with negatively worded items on self-report surveys. Structural Equation Modeling, 13: 440-464.

31.

Donnellan

M. B.

Oswald

F. L.

Baird

B. M.

Lucas

R. E.

2006. The mini-IPIP scales: Tiny-yet-effective measures of the Big Five factors of personality. Psychological Assessment, 18: 192-203.

32.

Dormann

Griffin

M. A.

2015. Optimal time lags in panel studies. Psychological Methods, 20: 489-505.

33.

Duckitt

2001. A dual-process cognitive-motivational theory of ideology and prejudice. In Zanna

(Ed.), Advances in experimental social psychology, Vol. 33: 41-113. San Diego, CA: Academic Press. https://doi.org/10.1016/S0065-2601(01)80004-6.

34.

Duckitt

Sibley

C. G.

2017. The dual process motivational model of prejudice. In Barlow

F. K.

Sibley

C. G.

(Eds.), The Cambridge handbook of the psychology of prejudice: 188-221.

35.

Dunlop

P. D.

Lee

2004. Workplace deviance, organizational citizenship behavior, and business unit performance: The bad apples do spoil the whole barrel. Journal of Organizational Behavior, 25: 67-80.

36.

Eid

2000. A multitrait-multimethod model with minimal assumptions. Psychometrika, 65: 241-261.

37.

Equal Employment Opportunity Commission. 2016. Select task force on the study of harassment in the workplace. Retrieved from https://www.eeoc.gov/select-task-force-study-harassment-workplace#_Toc453686298.

38.

Gatto

Dambrun

Kerbrat

De Oliveira

2010. Prejudice in the police: On the processes underlying the effects of selection and group socialisation. European Journal of Social Psychology, 40: 252-269.

39.

Geiser

Eid

Nussbeck

F. W.

2008. On the meaning of the latent variables in the CT-C (M-1) model: A comment on Maydeu-Olivares and Coffman (2006). Psychological Methods, 13: 49-57.

40.

Goff

P. A.

Lloyd

Geller

Raphael

Glaser

2016. The science of justice: Race, arrests, and police use of force. Retrieved from http://policingequity.org/research/1687-2.

41.

Götz

Gosling

Rentfrow

. in press. Small effects: The indispensable foundation for a cumulative psychological science. Perspectives on Psychological Science. https://doi.org/10.31234/osf.io/hzrxf.

42.

Gray

Kish-Gephart

J. J.

2013. Encountering social class differences at work: How “class work” perpetuates inequality. Academy of Management Review, 38: 670-699.

43.

Greenwald

A. G.

Banaji

M. R.

Nosek

B. A.

2015. Statistically small effects of the Implicit Association Test can have societally large effects. Journal of Personality and Social Psychology, 108: 553-561.

44.

Guimond

Dambrun

Michinov

Duarte

2003. Does social dominance generate prejudice? Integrating individual and contextual determinants of intergroup cognitions. Journal of Personality and Social Psychology, 84: 697-721.

45.

Haley

Sidanius

2005. Person-organization congruence and the maintenance of group-based social hierarchy: A social dominance perspective. Group Processes & Intergroup Relations, 8: 187-203.

46.

Hayduk

L. A.

1987. Structural equation modeling with LISREL: Essentials and advances. Baltimore, MD: John Hopkins University Press.

47.

Henry

P. J.

Sears

D. O.

2009. The crystallization of contemporary racial prejudice across the lifespan. Political Psychology, 30: 569-590.

48.

A. K.

Sidanius

Kteily

Sheehy-Skeffington

Pratto

Henkel

K. E.

. . . Stewart

A. L.

2015. The nature of social dominance orientation: Theorizing and measuring preferences for intergroup inequality using the new SDO7 scale. Journal of Personality and Social Psychology, 109: 1003-1028.

49.

Holland

J. L.

1959. A theory of vocational choice. Journal of Counseling Psychology, 6: 35-45.

50.

Hoofs

van de Schoot

Jansen

N. W.

Kant

2018. Evaluating model fit in Bayesian confirmatory factor analysis with large samples: Simulation study introducing the BRMSEA. Educational and Psychological Measurement, 78: 537-568.

51.

Houkamau

C. A.

Sibley

C. G.

2015. Looking Māori predicts decreased rates of home ownership: Institutional racism in housing based on perceived appearance. PLoS One, 10: e0118540.

52.

Cheung

G. W.

2008. Eight general questions on measurement equivalence/invariance that bother me most. Academy of Management Proceedings, 2008: 1-6.

53.

L. T.

Bentler

P. M.

1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6: 1-55.

54.

Hudson

S. K. T. J.

Cikara

Sidanius

2019. Preference for hierarchy is associated with reduced empathy and increased counter-empathy towards others, especially outgroup targets. Journal of Experimental Social Psychology, 85: 103871. doi:10.1016/j.jesp.2019.103871

55.

Khan

A. K.

Moss

Quratulain

Hameed

2018. When and how subordinate performance leads to abusive supervision: A social dominance perspective. Journal of Management, 44: 2801-2826.

56.

Kim

T. Y.

Cable

D. M.

Kim

S. P.

2005. Socialization tactics, employee proactivity, and person-organization fit. Journal of Applied Psychology, 90: 232-241.

57.

King

E. B.

Shapiro

J. R.

Hebl

M. R.

Singletary

S. L.

Turner

2006. The stigma of obesity in customer service: A mechanism for remediation and bottom-line consequences of interpersonal discrimination. Journal of Applied Psychology, 91: 579-593.

58.

Kish-Gephart

J. J.

Harrison

D. A.

Treviño

L. K.

2010. Bad apples, bad cases, and bad barrels: Meta-analytic evidence about sources of unethical decisions at work. Journal of Applied Psychology, 95: 1-31.

59.

Kteily

N. S.

Sidanius

Levin

2011. Social dominance orientation: Cause or “mere effect”? Evidence for SDO as a causal predictor of prejudice and discrimination against ethnic and racial outgroups. Journal of Experimental Social Psychology, 47: 208-214.

60.

Larson

L. M.

Rottinghaus

P. J.

Borgen

F. H.

2002. Meta-analyses of Big Six interests and Big Five personality factors. Journal of Vocational Behavior, 61: 217-239.

61.

Little

T. D.

Preacher

K. J.

Selig

J. P.

Card

N. A.

2007. New developments in latent panel analyses of longitudinal data. International Journal of Behavioral Development, 31: 357-365.

62.

Lucas

C. J.

2006. American higher education: A history. Berlin: Springer.

63.

Marsh

H. W.

Scalas

L. F.

Nagengast

2010. Longitudinal tests of competing factor structures for the Rosenberg Self-Esteem Scale: Traits, ephemeral artifacts, and stable response styles. Psychological Assessment, 22: 366-381.

64.

Marsman

Wagenmakers

E. J.

2017. Three insights from a Bayesian interpretation of the one-sided P value. Educational and Psychological Measurement, 77: 529-539.

65.

McDonald

R. P.

M. H. R.

2002. Principles and practice in reporting structural equation analyses. Psychological Methods, 7: 64-82.

66.

Muheljic

Drace

2018. University socialization and the acceptance of anti-egalitarian ideology: The underlying role of extrinsic life goals. European Journal of Social Psychology, 48: O73-O80.

67.

Muthén

2010. Bayesian analysis in Mplus: A brief introduction. Retrieved from https://www.statmodel.com/download/IntroBayesVersion%203.pdf.

68.

Muthén

Muthén.

Asparouhov

2015. Estimator choices with categorical outcomes: Technical Implementation. Retrieved from https://www.statmodel.com/download/EstimatorChoices.pdf.

69.

Muthén

L. K.

Muthén

B. O

. 1998-2017. Mplus user’s guide. 8th ed. Los Angeles: Muthén & Muthén.

70.

Nicol

A. A.

Charbonneau

Boies

2007. Right-wing authoritarianism and social dominance orientation in a Canadian military sample. Military Psychology, 19: 239-257.

71.

Ofosu

E. K.

Chambers

M. K.

Chen

J. M.

Hehman

2019. Same-sex marriage legalization associated with reduced implicit and explicit antigay bias. Proceedings of the National Academy of Sciences, 116: 8846-8851.

72.

Osborne

Satherley

Little

T. D.

Sibley

C. G

. in press. Authoritarianism and social dominance predict annual increases in generalized prejudice. Social Psychological and Personality Science. https://doi.org/10.1177/1948550620969608.

73.

Osborne

Sears

D. O.

Valentino

N. A.

2011. The end of the solidly Democratic South: The impressionable-years hypothesis. Political Psychology, 32: 81-108.

74.

Pascarella

E. T.

Edison

Nora

Hagedorn

L. S.

Terenzini

P. T.

1996. Influences on students’ openness to diversity and challenge in the first year of college. Journal of Higher Education, 67: 174-195.

75.

Peterson

B. E.

Lane

M. D.

2001. Implications of authoritarianism for young adulthood: Longitudinal analysis of college experiences and future goals. Personality and Social Psychology Bulletin, 27: 678-690.

76.

Pratto

Espinoza

2001. Gender, ethnicity, and power. Journal of Social Issues, 57: 763-780.

77.

Pratto

Sidanius

Levin

2006. Social dominance theory and the dynamics of intergroup relations: Taking stock and looking forward. European Review of Social Psychology, 17: 271-320.

78.

Pratto

Sidanius

Stallworth

L. M.

Malle

B. F.

1994. Social dominance orientation: A personality variable predicting social and political attitudes. Journal of Personality and Social Psychology, 67: 741-763.

79.

Pratto

Stallworth

L. M.

Sidanius

Siers

1997. The gender gap in occupational role attainment: A social dominance approach. Journal of Personality and Social Psychology, 72: 37-53.

80.

Robinson

J. K.

2002. Race, gender, and familial status: Discrimination in one US mortgage lending market. Feminist Economics, 8: 63-85.

81.

Rohrer

J. M.

2018. Thinking clearly about correlations and causation: Graphical causal models for observational data. Advances in Methods and Practices in Psychological Science, 1: 27-42.

82.

Satherley

Milojev

Greaves

L. M.

Huang

Osborne

Bulbulia

Sibley

C. G.

2015. Demographic and psychological predictors of panel attrition: Evidence from the New Zealand attitudes and values study. PLoS One, 10: e0121950. https://doi.org/10.1371/journal.pone.0121950.

83.

Schneider

1987. The people make the place. Personnel Psychology, 40: 437-453.

84.

Selig

J. P.

Little

T. D.

2012. Autoregressive and cross-lagged panel analysis for longitudinal data. In Laursen

Little

T. D.

Card

N. A.

(Eds.), Handbook of developmental research methods: 265-278. New York: Guilford Press.

85.

Sibley

C. G.

2020. Sampling procedure and sample details for the New Zealand Attitudes and Values Study. NZAVS Technical Documents, e01. Retrieved from https://cdn.auckland.ac.nz/assets/psych/about/our-research/nzavs/NZAVSTechnicalDocuments/NZAVS-Technical-Documents-e01-Sampling-Procedure-and-Sample-Details.pdf.

86.

Sibley

C. G.

Duckitt

2008. Personality and prejudice: A meta-analysis and theoretical review. Personality and Social Psychology Review, 12: 248-279.

87.

Sibley

C. G.

Wilson

M. S.

Duckitt

2007. Antecedents of men’s hostile and benevolent sexism: The dual roles of social dominance orientation and right-wing authoritarianism. Personality and Social Psychology Bulletin, 33: 160-172.

88.

Sidanius

Pratto

Sinclair

Van Laar

1996. Mother Teresa meets Genghis Khan: The dialectics of hierarchy-enhancing and hierarchy-attenuating career choices. Social Justice Research, 9: 145-170.

89.

Sidanius

Cotterill

Sheehy-Skeffington

Kteily

Carvacho

2017. Social dominance theory: Explorations in the psychology of oppression. In Sibley

C. G.

Barlow

F. K.

(Eds.), The Cambridge handbook of the psychology of prejudice: 149-187. New York: Cambridge University Press.

90.

Sidanius

Liu

J. H.

Shaw

J. S.

Pratto

1994. Social dominance orientation, hierarchy attenuators and hierarchy enhancers: Social dominance theory and the criminal justice system. Journal of Applied Social Psychology, 24: 338-366.

91.

Sidanius

Pratto

1999. Social dominance: An intergroup theory of social hierarchy and oppression. New York: Cambridge University Press.

92.

Sidanius

Pratto

2001. Social dominance: An intergroup theory of social hierarchy and oppression. New York: Cambridge University Press.

93.

Sidanius

Pratto

2003. Social dominance theory and the dynamics of inequality: A reply to Schmitt, Branscombe, & Kappen and Wilson & Liu. The British Journal of Social Psychology, 42: 207-213.

94.

Sidanius

van Laar

Levin

Sinclair

2003. Social hierarchy maintenance and assortment into social roles: A social dominance perspective. Group Processes & Intergroup Relations, 6: 333-352.

95.

Simmons

A. L.

Watkins

M. B.

Umphress

E. E.

2015. Are male tokens bad for females? The effect of social dominance orientation on high-status tokens’ evaluations of low-status members. Journal of Applied Social Psychology, 45: 55-66.

96.

Sinclair

Sidanius

Levin

1998. The interface between ethnic and social system attachment: The differential effects of hierarchy-enhancing and hierarchy-attenuating environments. Journal of Social Issues, 54: 741-757.

97.

Son Hing

L. S.

Bobocel

D. R.

Zanna

M. P.

McBride

M. V

. 2007. Authoritarian dynamics and unethical decision making: High social dominance orientation leaders and high right-wing authoritarianism followers. Journal of Personality and Social Psychology, 92: 67-81.

98.

Statistics New Zealand. 2018. Survey of working life: 2018. Retrieved from https://www.stats.govt.nz/reports/survey-of-working-life-2018.

99.

Thomsen

Green

E. G.

A. K.

Levin

van Laar

Sinclair

Sidanius

2010. Wolves in sheep’s clothing: SDO asymmetrically predicts perceived ethnic victimization among White and Latino students across three years. Personality and Social Psychology Bulletin, 36: 225-238.

100.

Tyler

T. R.

2006. Why people obey the law. Princeton, NJ: Princeton University Press.

101.

UBS/PwC. 2020. Billionaires report: Riding the storm. Market turbulence accelerates diverging fortunes. Retrieved from https://www.pwc.ch/en/publications/2020/UBS-PwC-Billionaires-Report-2020.pdf.

102.

Umphress

E. E.

Simmons

A. L.

Boswell

W. R.

Triana

M. D. C.

2008. Managing discrimination in selection: The influence of directives from an authority and social dominance orientation. Journal of Applied Psychology, 93: 982-993.

103.

Umphress

E. E.

Smith-Crowe

Brief

A. P.

Dietz

Watkins

M. B.

2007. When birds of a feather flock together and when they do not: Status composition, social dominance orientation, and organizational attractiveness. Journal of Applied Psychology, 92: 396-409.

104.

Unzueta

M. M.

Knowles

E. D.

G. C.

2012. Diversity is what you want it to be: How social-dominance motives affect construals of diversity. Psychological Science, 23: 303-309.

105.

Vallas

Cummins

2014. Relational models of organizational inequalities: Emerging approaches and conceptual dilemmas. American Behavioral Scientist, 58: 228-255.

106.

Vandenberg

R. J.

Lance

C. E.

2000. A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3: 4-70.

107.

Van Assen

M. A.

van Aert

Wicherts

J. M

. 2015. Meta-analysis using effect size distributions of only statistically significant studies. Psychological Methods, 20: 293-309.

108.

Van Laar

Sidanius

Rabinowitz

J. L.

Sinclair

. 1999. The three Rs of academic achievement: Reading,’riting, and racism. Personality and Social Psychology Bulletin, 25: 136-148.

109.

Van Maanen

J. E.

Schein

E. H

. 1977. Toward a theory of organizational socialization. In Staw

B. M.

(Ed.), Research in Organizational Behavior, 1: 209-265.

110.

Van Vianen

A. E

. 2018. Person–environment fit: A review of its basic tenets. Annual Review of Organizational Psychology and Organizational Behavior, 5: 75-101.

111.

Verquer

M. L.

Beehr

T. A.

Wagner

S. H.

2003. A meta-analysis of relations between person–organization fit and work attitudes. Journal of Vocational Behavior, 63: 473-489.

112.

Vogel

R. M.

Rodell

J. B.

Lynch

J. W.

2016. Engaged and productive misfits: How job crafting and leisure activity mitigate the negative effects of value incongruence. Academy of Management Journal, 59: 1561-1584.

113.

Wang

2012. Structural equation modeling: Applications using Mplus. Chichester, UK: Wiley/Higher Education Press.

114.

Weston

S. J.

Ritchie

S. J.

Rohrer

J. M.

Przybylski

A. K.

2019. Recommendations for increasing the transparency of analysis of preexisting data sets. Advances in Methods and Practices in Psychological Science, 2: 214-227.

115.

Yuan

MacKinnon

D. P.

2009. Bayesian mediation analysis. Psychological Methods, 14: 301-322.

116.

Zyphur

M. J.

Oswald

F. L.

2015. Bayesian estimation and inference: A user’s guide. Journal of Management, 41: 390-420.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB