The idea that a specific motivation to justify the system is required to explain instances of system justification, beyond personal/group interest, has been challenged by Owuamalam et al. (2018, 2019a, 2019b) in their Social Identity Model of System Attitudes (SIMSA). SIMSA assumes that system justification, especially among the disadvantaged, can be explained by the traditional motives of personal/group interests, without recourse to an independent system motivation explanation. For example, SIMSA assumes that some instances of the puzzling justification of disadvantageous systems, sometimes seen among members of low-status groups, can result from them paying attention to their identity needs but at a more inclusive level of social categorization (e.g., their nation). So, for instance, African-Americans may justify disadvantageous realities in America, if their attention is strongly focused on the needs that are tied to their superordinate identity as Americans rather than the needs that are tied to their subgroup (African) identity. Therefore, African-Americans may justify disadvantageous systems (e.g., the American government) that regulate/oversee the institutional huddles confronting fellow group members (e.g., fatal law enforcement), if their attention is narrowly focused on their national (superordinate) identity as Americans. A similar process should also operate for high-status groups. In this sense, system justification is likely nothing more than a favorable evaluation of one’s superordinate in-group.

But, this superordinate in-group bias explanation is not the only one on offer under the social identity umbrella (i.e., social identity theory (SIT); Tajfel and Turner, 1979), especially given that SIMSA and its predecessor (SJT) do not currently say much about system justification of members of groups that are placed in an intermediate position (i.e., those who are disadvantaged but can nevertheless realize downward comparison) relative to those who are clearly advantaged or disadvantaged. This is the vacuum that the Triadic Social Stratification Theory (TSST, Caricati, 2018; Caricati and Owuamalam, 2020) fills, also drawing from SIT (Tajfel and Turner, 1979).

In particular, TSST offers one distinct reason why people might support their social systems, especially the puzzling instances of system justification among the disadvantaged. TSST, similar to its parent framework (SIT), assumes that people are motivated by a need for positive self-worth to improve their social position both personally (i.e., by upward individual mobility) and collectively (by upward social mobility) and that sometimes this goal can be reached by comparing the outcomes of an individual (or the outcomes of an individual’s social group) with those of others. Evidence shows that people are often motivated to enhance their own social position by upward individual mobility unless this goal is impeded in some way (Wright et al., 1990; Ellemers, 1993; Ellemers et al., 1993; Jackson et al., 1996). So, for example, an African-American may choose to compare him/herself with other in-group members (even their immediate family) who are not doing so well and may embrace America and consequently support its systems simply because it has afforded him/her the opportunity to rise above his/her parents or other members of their African-American community. Beyond the foregoing social comparisons, it is also possible for individuals to compare their outcomes against different time points in their life, so that a favorable comparison is achieved when individuals believe that they are doing better now than they did in the past (i.e., temporal comparison, Blanz et al., 1998).

TSST is, therefore, currently unique in its emphasis on the social comparison provision of SIT, arguing that so long as a given social or temporal stratification allows for intermediate positioning (whether it be within groups or between time points), that people may be motivated by the need for positive self-worth to support the status quo in which this was made possible. That is, they do so because (1) they are better off than others (individuals or groups) and (2) they are better off now than they were in the past.

Accordingly, it has been shown that people are more likely to justify their societal systems (Caricati and Sollami, 2018) and are even less likely to question these realities (Becker, 2012) when the status quo permits a downward comparison. Similarly, a comparison between actual and past conditions (Mummendey et al., 1996, 1999; Zagefka and Brown, 2005) has been suggested as an important factor when considering the extent to which people support their social systems. For example, people who successfully realize (or believe in) individual mobility may be more likely to support existing social arrangement and, consequently, also more likely to deny that prejudice or discrimination exists (Ng and Chiu, 2001; Derks et al., 2011, 2015; Owuamalam et al., 2017). Similarly, people who believe that their standard of life has improved with respect to the past may be motivated to justify the system that allowed for such improvement to materialize (Caricati and Owuamalam, 2020).

Although research has shown that social and temporal comparisons [as well as superordinate (national) identification] are able to bolster system support, none of these previous studies have examined, so far, these variables together, meaning that their unique contributions with regard to system justification have yet to be determined. This is important for two reasons at least. On the one hand, the national in-group provides the context to appraise the life outcomes of an individual relative to fellow nationals (including family members). From this perspective, it is possible to argue that a downward social comparison with other nationals could enhance the extent to which an individual observes (and appreciates) the benefits of greater (psychological) investments in a national in-group that made this favorable (self-worth-boosting) comparison possible. Therefore, in this situation, system support could be due to social comparison alone, or it could be due to its joint action with national identification. System support could also be due to “love of country” and not due to a favorable social/temporal comparison. In short, understanding the unique contribution of these explanations is difficult, if both are not simultaneously accommodated within the same model, to allow the partitioning of variances in system support that is due to each explanation.

At the same time, it is also possible that “love of country” (i.e., superordinate identification) could bolster entitlement feelings, and such feelings could poison the normally self-enhancing effect of downward temporal comparisons (e.g., when white working-class Americans compare their outcomes at present to what it was in the past, amidst the influx of competing migrant groups). From this perspective, then, it is possible that the ordinarily positive effect of favorable temporal comparison on system justification might cease (or be suppressed) when superordinate identification is taken into account. No other investigation has systematically unpacked these processes by considering them in tandem. Therefore, to be more certain that the social and temporal comparisons of TSST, as well as the superordinate in-group bias explanation of SIMSA are independent influencers of the system justification effect, one should demonstrate that they offer unique insights when considered together. In this investigation, therefore, we focused on three key self/group-interested predictors of system justification, namely, superordinate (national) identification (as per SIMSA), social comparison, with income as an indicator of relative social advantage, and temporal comparison (as per TSST).

First, we expected that system justification would be positively related to national identification over and beyond the alternative explanations (i.e., social/temporal comparisons) because, according to SIMSA, people would be more likely to support the national system to the extent that they identify with their nation (Hypothesis 1). Second, based on TSST, we expected that all otherwise being equal, the advantaged (as well as the intermediately positioned) would be more likely than their relatively more disadvantaged counterparts to justify the system (Hypothesis 2). This is because, in this situation, such individuals can obtain positive personal and group comparisons from a system that enabled their relatively advantaged social position. Finally, based on TSST, we considered the consequence of comparing own standard of living to those of one’s parents over time, which we used to proxy temporary comparison. It is important to note that this instance of intragroup comparison is consistent with evidence that people tend to prefer intragroup comparison over intergroup comparison and often engage in comparison with past outcomes as a means of dealing with social identity-based challenges (Major and Forcey, 1985; O’Brien and Major, 2009; Akfirat et al., 2016). Therefore, we expected that system justification would increase when people believe that they are better off at present than in the past, especially when this comparison is tied to those people that one ordinarily look up to (e.g., one’s parents, Hypothesis 3).

We used the 7th Wave of the World Value Survey (WVS) with 70,867 participants from 49 countries worldwide. We considered only participants with no missing values¹ on the measures that were relevant to our analysis (refer to the explanation below), and this consideration reduced both the N-size (now down to 55,721 participants) and the number of nations (down to 40 countries: 47.7% men, mean age = 42.16, SD = 15.97, range = 16–103; Table 1).

System justification has been operationalized in many ways by its principal proponent, such as out-group favoritism (Jost et al., 2004), general and economic system justification (Jost and Thompson, 2000; Kay and Jost, 2003), and trust/confidence in government (Jost et al., 2003) among others. In this study, we focused on the last operationalization (i.e., trust in government), which we assessed with four items asking participants to indicate the extent to which they were confident in the institutions of governance of their society, namely, parliament, government, political parties, and justice system/courts (1 = a great deal, 4 = not at all, reverse scored). We focused on trust in government (and its apparatuses) because it satisfies several auxiliary conditions that should enable the system motive to manifest. Because this system motive is theorized to be in conflict with the personal/group interests of people in disadvantaged groups (Jost et al., 2003, 2004), it should create an obstacle for our personal/group-interested predictions to operate. Specifically, systems of governance are institutions that objectively high-, intermediate-, and low-status people are often highly dependent on, also because these entities are stable and inescapable realities of citizens’ existence (Kay et al., 2009; Friesen et al., 2019; Jost, 2019). The inability to escape governments that regulate sub-systems that undermine people’s outcomes could cause a sense of personal control to decrease (Kay and Friesen, 2011; Laurin et al., 2013), and these situations should allow the system motive to take a prime position, while personal and group motives should be relegated to the rear position (based on SJT), meaning that it should be more difficult to find supportive evidence for the interest-based predictions derived from SIMSA and TSST, especially for low- and intermediate-status groups. A multilevel confirmatory factor analysis revealed that the four items on this scale had adequate reliability both within (a = 0.82) and between nations (a = 0.97). Items were then averaged so that higher scores indicate higher system justification (i.e., trust in governance).

National identification was measured with a single item asking participants “how close do you feel to your country?” (1 = very close, 4 = not close at all), which was reversed so that higher scores indicated high levels of national identification.

A within-group temporal comparison was measured with a single item asking participants “comparing your standard of living with your parents’ standard of living when they were about your age, would you say that you are better off, worse off, or about the same?” (1 = better off, 2 = worse off, and 3 = about the same).

An income-based social comparison was measured with a single item asking participants “On this card is an income scale on which 1 indicates the lowest income group and 10 the highest income group in your country. We would like to know in what group your household is. Please, specify the appropriate number, counting all wages, salaries, pensions, and other incomes that come in.”

To control for the effect of other levels of social advantage and being, thus, able to better estimate the effect of variables of interest to our test, we added some level-1 and level-2 covariates. At level-1, we considered gender, age, and the higher level of education attained by participants (0 = Early childhood education, 8 = Doctoral or equivalent)². At the country level, we considered the GINI index and GDP PPP (Gross Domestic Product based on Purchasing Power Parity) to account for the potential effect of objective wealth inequality in the nation (higher GINI indicates higher inequality) and national wealth level (both GDP and GINI were taken from the WVS database).

We performed a series of multilevel hierarchical models in which country was the nesting variable. Model 1 considered income (i.e., social comparison) and national identification as the predictors of system justification (i.e., trust in national governance). In Model 2, we added temporal comparison as a predictor. The income-based social comparison was coded using two dummy variables considering “the same” as a reference category. Model 3 added interactions between national identification and temporal and social comparisons to control for potential interactive effects. In all models, level-1 continuous variables were centered within nations while level-2 covariates were grand-mean centered. The slopes of all level-1 predictors were allowed to have random variation across nations, while covariates were treated as fixed effects. Analyses were performed with restricted maximum likelihood estimation using the lme4 package (Bates et al., 2015) in R (R Core Team, 2021).

Table 2 reports both level-1 and level-2 zero-order correlations and descriptive statistics of measured constructs. At the individual level (refer to the correlation coefficients above the diagonal), the resulting associations between the principal variables were generally weak in magnitude but in the expected direction. Notably, system justification was positively correlated with (1) positive temporal comparison, (2) national identification, and (3) income. With respect to the control variables: women, older and less educated people appeared to be more likely to justify the system.

Supporting the use of multilevel modeling, a null model in which only the intercept varied randomly across nations revealed that 36.4% of the variance in system justification was due to the different nations represented in the survey [Intraclass correlation (ICC) = 0.364; χ²(1) = 25638.00, p < 0.001].

Table 3 depicts results from the estimated models. First, models indicated that people, on average, did not express much trust in the system of governance of their nation (i.e., their level of system justification was low). More importantly, as indicated, both income and national identification were significantly and positively associated with system justification in all models. Model 2 indicated, as expected, a main effect of temporal comparison F(2, 37) = 21.22, p < 0.001 so that those who believed that their situation has worsened now than what it was in the past justified the system significantly less (M = 2.13, SE = 0.072) than people who believed that their social condition was the same (at least relative to their parents, M = 2.22, SE = 0.075). Moreover, those who believed that their situation has somewhat stagnated (i.e., “the same” group) were also less likely to justify the system compared with those who reported being better off now than in the past (M = 2.26, SE = 0.079).

Considering covariates, results indicated that women justified their national systems of governance more than men did. Interestingly, more educated people were less (not more) likely to justify their societal systems of governance. GDP (i.e., the objective index of societal wealth) and GINI (i.e., the objective index of societal-level inequality) appeared to have no significant main effects on system justification in the current data.

On an exploratory basis, we considered, in Models 3, the interactions between national identification and temporal and social comparisons, given our a priori speculation that national identification could actually fade (or suppress) the effects of social/temporal comparisons on system justification. Results revealed that national identification × income interaction was not significant (b = 0.002, SE = 0.002, p = 0.235), suggesting that social comparison is by-and-large a unique explanation for system justification that may not necessarily be contingent on superordinate identification (at least in this case). However, national identification interacted with temporal comparison to predict system justification, F(2, 43,0341) = 3.867. p = 0.021. When we decomposed this interaction by examining the association between temporal comparisons and system justification when superordinate (national) identification was high (M + 1SD) vs. low (M-1SD), we found, consistent with our speculation, that positive temporal comparisons seem to work best, in terms of its boosting effect on system justification, when national identification was low (Δb_same–worse = 0.07, SE = 0.02, p = 0.0003; Δb_{better–same} = 0.05, SE = 0.02, p = 0.006). Meanwhile, when superordinate (national) identification was high, positive temporal comparisons significantly predicted an increase in system justification also, but only in relation to the same vs. worse-off contrast (Δb_same–worse = 0.11, SE = 0.02, p < 0.0001) and the better-off vs. worse-off contrast (Δb_{better–worse} = 0.15, SE = 0.02, p < 0.0001), but not in relation to the better-off vs. same contrast (Δb_{same–better} = −0.03, SE = 0.02, p = 0.120).

This research, as any other correlational research, does not allow causal inference, and thus, some caution is needed in this respect. For example, it is possible that people who strongly trust their national institutions may identify strongly with their national in-group, rather than the opposite. However, when we examined this possibility, we found that a model in which national identification was the dependent variable and system justification (i.e., trust in governance) was the predictor, also produced a positive relationship, b = 0.118, SE = 0.02, p < 0.001, although the fit of this latter model was reliably poorer than the preferred reverse (i.e., Model 2), based on poorer fit indices (difference between Model 2 and alternative model: ΔAIC = −16948.90, ΔBIC = −16948.90, ΔICC = 0.24, Table 3). Note that even though this reverse causation is plausible, it would imply that a credible system provides individuals with a reason to identify with their nation, and this outcome will be more consistent with the rationality implied in the social identity perspective than with the irrationality implied in competing frameworks that assume system justification has less (if at all anything) to do with social identity needs.