Introduction

The Belief in a Just World (BJW) serves several different functions for individuals. One is that BJW is empowering (Lerner, 1980). Empowerment refers to the social-relational process by which individuals gain and maintain the ability to determine personally relevant resources and outcomes and exert influence over others’ resources and outcomes (see, for example, Anderson et al., 2012; Fiske & Berdahl, 2007; Keltner et al., 2003; Leach et al., 2017). Recently four experimental studies (Bartholomaeus et al., 2023a) provided, for the first time, empirical evidence that empowerment is indeed an important mechanism by which BJW contributes to the adaptive functioning of individuals in specific social contexts. However, while this experimental evidence appears to be robust, the studies were necessarily constrained. The empowering influence of BJW was limited specifically to the dependent variables of positive and negative affect. Further, due to the use of experimental designs, empowerment was tested in tightly controlled and highly specific scenarios. Thus, it is unclear whether the empowering function of BJW facilitates a broader number of positive attributes in everyday life and generalises to the psychological functioning of individuals outside the parameters of experimental contexts. In this paper we address these limitations by (a) testing associations between BJW, empowerment, and a broader range of positive and negative outcomes, and (b) using both cross-sectional and longitudinal designs with diverse participant samples.

The Belief in a Just World

BJW can be both manipulated and measured. In the case of the latter, it is conceptualized as a trait-level variable, insofar as individuals vary in the extent to which they see the world as a stable and orderly place and feel that they understand and can exert influence over the environment in which they live (Bartholomaeus et al., 2023b; Dalbert & Donat, 2015; Lerner, 1980). BJW is seated in the foundational notion that people get what they deserve and deserve what they get, i.e., cause-and-effect. The centrality of cause-and-effect to everyday functioning is established early in life as individuals enter into an intrapersonal implicit contract (Lerner et al., 1976). This contract states that in order to gain greater long-term rewards the individual must delay meeting their impulsive desires and invest in the future.

When treating BJW as an individual difference, researchers have distinguished between the belief that the world is a just place for others (BJW-other) and the world as a just place for the self (BJW-self; Lipkus et al., 1996). While BJW-other is generally associated with harsh social attitudes (for a review see Hafer & Sutton, 2016), BJW-self is consistently associated with many positive outcomes. For example, there is much support for Dalbert’s (1998) seminal claim that BJW-self functions as a coping resource (e.g., Kiral Ucar et al., 2022; Furnham, 2022). Specific to the present study, a large literature indicates a robust association between BJW-self and wellbeing and other affective outcomes (see Bartholomaeus & Strelan, 2019 for a review; and for more recent evidence see Chobthamkit et al., 2022; Goodwin & Williams, 2023; Hoolihan & Thomas, 2020; Ma et al., 2023).

BJW-self and Empowerment

The approach/inhibition theory of power (Keltner et al., 2003) accounts for the varied ways in which the experience of power influences affect, cognition, and behaviour. The theory suggests that increased power activates, and decreased power inhibits, behavioural motivations. An increased sense of power activates the behavioural activation system (BAS). BAS functioning is characterised by heightened positive responsiveness to rewards, an increased drive to pursue goals, and reward-seeking behaviour (Carver & White, 1994). In contrast, decreased power is associated with a lack of resources and a higher awareness of social constraints, and activates the behavioural inhibition system (BIS), which is associated with sensitivity towards potential threats inhibiting behaviour that may elicit punishment (Carver & White, 1994).

The approach/inhibition theory of power can be interpreted through a just world lens. Pursuing goals and increased reward-seeking behaviour necessitates a world in which there is a clear link between cause and effect. Further, both empowerment and the derivative facets of BJW (BJW-self & BJW-other) are conceptually related, at least intuitively, because they are inherently interpersonal in nature. The initial experiments that first established the BJW phenomenon involved a group of participants making social judgements about an innocent victim (Lerner & Simmons, 1966). In modern parlance, these innocent victim experiments likely engaged the BJW-other sphere, as participants were being asked to gauge the justness of the world of another person. However, the suffering of an innocent other is only disconcerting to the individual to the extent that it suggests that the observer too lives in an unjust world where they may be subject to undeserved punishment (Lerner, 1980). As such, both facets of BJW, at the core, are socially oriented, which in turn speaks to two aspects of empowerment. That is, (a) a belief that the individual gets what they deserve helps the individual to feel in control of personally relevant rewards and punishments (Leach et al., 2017), and (b) the belief that others get what they deserve helps the individual to feel as though they can understand and possibly influence the rewards and punishment of others (Keltner et al., 2003).

Further theorising specifically links BJW-self and empowerment. A strong endorsement of BJW-self enables people to confront the world as if it were a stable and orderly place (Lerner, 1980). A world that is stable and orderly is predictable and it is this predictability, combined with the expectation that effort will be fairly rewarded (Correia & Dalbert, 2007), that provides people with a sense of power over their present and future circumstances. Increased commitment to BJW-self, therefore, leads to a heightened sense of the ability to determine one’s own fate and the events that happen in their lives. Further, BJW-self enables people to exert influence over their interactions and relationships with others. For example, BJW-self is associated with one’s sense that they can influence the course of their romantic relationships through hard work (Bartholomaeus & Strelan, 2016), that they can help (or influence) the plight of others (Bègue et al., 2008), and that they can rise above interpersonal transgressions and forgive (see Strelan, 2018 for a review). In general, BJW-self also exerts an adaptive influence over one’s interpretations of social interactions, i.e., by insulating people’s wellbeing from the negative actions of others (Dzuka & Dalbert, 2007)Footnote 1.

Adaptive and Maladaptive Outcomes

As noted earlier, four experimental studies have demonstrated that BJW-self predicts positive and negative affect through empowerment (Bartholomaeus et al., 2023a). In the present study we selected a range of outcomes that represent specific aspects of psychological wellbeing to test the generalisability of the empowering function of BJW-self, specifically, life satisfaction, optimism, and resilience, and depression, anxiety, and stress. We chose these constructs because (a) there are firm theoretical grounds suggesting that they stem from the empowering function of BJW-self, and (b) they cover a range of healthy traits and positive processes, as well as symptoms of psychopathology.

Life satisfaction, optimism, and resilience are components of psychological health and wellbeing. Life satisfaction is one’s cognitive appraisal of the quality of their current life circumstances (Diener et al., 1985); trait optimism refers to the tendency to expect good things to happen (Carver et al., 2010); and resilience is the ability to bounce back or recover from stress (Smith et al., 2008). Conversely, depression, anxiety, and stress represent general dispositional subjective distress. The tripartite model suggests that depression, anxiety, and stress can all be characterised by high levels of negative affectivity (Henry & Crawford, 2005). Whereas stress is defined simply by the presence of negative affectivity, depression is suggested to also include feelings of low positive affectivity. In addition to negative affectivity, anxiety is characterised by feelings of hyperarousal. Thus, negative affectivity (or the experience of negative emotion) is a central component to the experience of depression, anxiety, and stress.

All aforementioned variables have been associated with BJW-self in previous studies. For instance, various indices of subjective wellbeing and affect are positively associated with BJW-self (for a review see Bartholomaeus & Strelan, 2019). Similarly, self-report levels of optimism and hope are also associated with BJW-self (Strelan & Callisto, 2020; Xie et al., 2011). Also, BJW-self has been shown to predict a variety of resilient behaviours in response to natural disasters such as decreased mental health symptoms (Otto et al., 2006), and an increased sense of hope (Xie et al., 2011). In a similar vein, BJW-self has been shown to be negatively associated with indices of distress, including decreased depressive symptoms (Kamble & Dalbert, 2012), lower anxiety (Otto et al., 2006), and reduced stress (Lipkus et al., 1996).

The empowering function can explain the association between BJW-self and a variety of wellbeing outcomes. Through the activation of BAS functioning a sense of power has been shown to be associated with general adaptive functioning (Taubitz et al., 2015) as well as subjective wellbeing (Leach & Weick, 2018). A sense of power is also associated with optimistic thinking (Fast et al., 2009), that is, people with power are more likely to see rewards and opportunities in ambiguous information (Keltner et al., 2003). Similarly, research suggests an empirical link between power and resilience insofar as trait resilience correlates directly with the reward responsiveness, drive, and fun-seeking subscales of the BAS scale (Genet & Siemer, 2011). The empowering function of BJW-self may also be related to people’s propensity to invest in long-term goals, another mechanism theorised to mediate the association between BJW and wellbeing (Dalbert, 2001). Individuals feeling empowered and experiencing increased BAS functioning are likely to invest in long-term goals—as a characteristic of BAS functioning is an increased drive to pursue goals (Carver & White, 1994)—while also experiencing the wellbeing outcomes associated with BAS functioning.

It is also plausible that those who have a weaker belief in the predictability and stability of the world experience a sense of disempowerment and, therefore, are more sensitive to potential threats. The experience of disempowerment (or low levels of power) has been linked to increased reports of negative affect (Anderson & Berdahl, 2002; Bombari et al., 2017). This increased negative affectivity—a component central to depression, anxiety, and stress—likely leads to increased symptoms of psychopathology. Conversely, empowerment has been associated with a reduced experience of depression, anxiety, and stress (Strelan et al., 2019).

The Present Study

The main objective of this study was twofold: first, to test if the empowering function of BJW-self generalises to a range of wellbeing outcomes and, second, to test whether the empowering function operates outside the controlled social context of experimental research designs. Formally, we hypothesise that BJW-self will be positively associated with life satisfaction, optimism, and resilience, and negatively associated with depression, anxiety, and stress indirectly through a sense of empowerment. These hypotheses are summarised in Fig. 1. Further, we hypothesise that the relations depicted in Fig. 1 are temporally stable and that BJW-self will causally influence empowerment and, in turn, life satisfaction, optimism, resilience, depression, anxiety, and stress over time.

Fig. 1
figure 1

Conceptual diagram of the hypothesised latent variable structural equation model.

Control of BJW-other not depicted.

Full size image

We used latent variable structural equation modelling (SEM) to model the associations between the variables and thus formally test the above hypotheses (Sample 1). We then replicated this model with a larger participant sample from a different population (Sample 2). Finally, we tested the temporal stability of the model over a one-year period using a subsample of participants from Sample 1 (Subsample 3).

Method

Participants

Sample 1

Participants were N = 355 undergraduate students from a large Australian university who participated in the first wave measurement of a longitudinal study in exchange for course credit. Due to often high attrition rates in longitudinal studies we aimed to recruit as many participants as possible. Respondents with incomplete data (n = 14); multiple survey attempts (n = 33); or who were underage (n = 1) were removed. The final sample consisted of n = 307 students (92 men, 212 women, 3 transgender, Mage = 20.7, SD = 5.93, age ranged from 18-to-62 years). The majority of participants were Australian (83%) and 88% spoke primarily English. Additionally, 71% of participants were not in a relationship and 53% reported having no religion.

Sample 2

Participants were N = 450 respondents from the Prolific website who participated for £1.00 (GBP; 246 men, 197 women, 3 transgender, 4 prefer not to answer, Mage = 32.1, SD = 11.6, age ranged from 18-to-73 years). Participants were limited to those who currently resided in Australia and who had a Prolific approval rating of over 90%, indicating satisfactory participation in previous studies. The majority of participants were Australian (76%) and 96% spoke primarily English. Half of the participants were not in a relationship (46%), and 34% were either married or in a long-term relationship; 60% reported having no religion.

Subsample 3

All participants from Sample 1 (T1) were invited to complete the same survey again approximately one year later (T2), N = 110 did so. Participants received $15.00 (AUD) and the opportunity to win one of five $100.00 (AUD) prizes. Respondents with > 30% incomplete data (n = 19); multiple survey attempts (n = 6); or who had not completed the T1 survey (n = 2) were removed. The final sample consisted of n = 83 participants (23 men, 59 women, 1 transgender, Mage = 22.1, SD = 8.92, age ranged from 18-to-62 years). The majority of participants were Australian or New Zealander (76%) and 90% spoke primarily English. Further, 67% were not in a romantic relationship and 55% reported having no religion.

Procedure

Participants completed all surveys online. Participants in Sample 1 accessed the survey from an internal university website; participants in Sample 2 accessed the survey on the Prolific website; and Subsample 3 participants accessed the survey via a link in a follow-up email. All participants first provided informed consent for their participation in the study. Participants in all samples completed measures of BJW-self, BJW-others, empowerment, life satisfaction, optimism, resilience, depression, anxiety, and stressFootnote 2. Participants in Samples 1 and 2 then provided demographic information and participants in Sample 1 were asked to opt-in to the second measurement. Participants in Subsample 3 completed the T2 measure approximately one year after the T1 measure (Mtime = 54.7 weeks, SD = 11.6 weeks). Participants were then thanked and debriefed. Ethical approval was obtained for this research from the University of Adelaide’s Human Research Ethics Committee (18/05).

Measures

Scale reliabilities ranged from ωt = 0.74 (Optimism) to ωt = 0.93 (Depression). Test-retest reliabilities range from r = .50 (Depression) to r = .76 (Empowerment). All scale reliabilities for are provided in Table 1.

Table 1 Reliability Estimates for Scales Used with the Three Samples
Full size table

Belief in a Just World

We measured BJW-self and BJW-other using Lipkus et al.’s (1996) sixteen-item scale. Participants responded on a seven-point Likert scale (1 = completely disagree, 7 = completely agree). Sample items are, “I feel that the world treats me fairly” (BJW-self); “I feel that people earn the rewards and punishments they get” (BJW-other). For all scales, higher scores indicated a greater endorsement of the construct.

Empowerment

Participants’ level of empowerment was measured using the eight-item Sense of Power Scale (Anderson et al., 2012). All items were responded to on a seven-point Likert scale (1 = disagree strongly, 7 = agree strongly). A sample item is, “I think I have a great deal of power”.

Life Satisfaction

We measured life satisfaction using the five-item Satisfaction with Life Scale (Diener et al., 1985). All items were answered on a seven-point Likert scale (1 = strongly disagree, 7 = strongly agree). A sample item is, “I am satisfied with my life”.

Optimism

We used the ten-item (including four filler items) Life Orientation Test-Revised to measure optimism (LOT-R; Scheier et al., 1994). Items were answered on a five-point Likert scale (0 = I disagree a lot, 4 = I agree a lot). An example item is, “I’m always optimistic about my future”.

Resilience

Resilience was measured using the six-item Brief Resilience Scale (BRS; Smith et al., 2008). Responses were on a five-point Likert scale (1 = strongly disagree, 5 = strongly agree). An example item is, “I tend to bounce back quickly after hard times”.

Depression, Anxiety, and Stress

We measured depression, anxiety, and stress with the 21-item Depression, Anxiety, and Stress Scale (DASS-21; Henry & Crawford, 2005). Participants responded on a four-point Likert scale (0 = did not apply to me at all, 3 = applied to me very much or most of the time). Sample items are, “I felt that life was meaningless” (depression); “I felt I was close to panic” (anxiety); and “I found it difficult to relax” (stress).

Statistical Analysis

Sample 1 Analysis

Analyses for Samples 1 and 2 were conducted in MPlus Version 8.3 (Muthén & Muthén, 1998–2017). We began by estimating a measurement model and then the structural model. We used a mean and variance adjusted weighted least squares estimator (WLSMV; Sellbom & Tellegen, 2019). Indirect effects of BJW-self on outcome variables were tested using bootstrap generated 95% confidence intervals (1,000 iterations). Criteria for model fit were CFI and TLI > 0.90; RMSEA < 0.06; and SRMR < 0.08 (Marsh et al., 2004). Criteria for judging path coefficients were; small > 0.05; moderate > 0.10; and large > 0.25 (Keith, 2006).

Sample 2 Analysis

We tested for differences between samples on all demographic variables. To test the generalisability of the proposed model with a different larger sample, we fit the same model from Sample 1 to Sample 2. To test the replicability of the structural model we first assessed overall model fit and then determined whether loadings, path estimates, indirect effects, and latent factor variances of the Sample 2 model fell within the CI99% of those estimates from the Sample 1 model.

Subsample 3 Analysis

Analysis for Subsample 3 was conducted in R (version 4.0.2). Initially we conducted a dropout analysis comparing differences on demographic and outcome variables. The sample size (n = 83) was too small to conduct latent variable SEM, therefore, we entered scale scores into a SEM model. Longitudinal mediation was estimated using a modified half-longitudinal model (Cole & Maxwell, 2003). Bootstrap generated 95% confidence intervals (1,000 iterations; bias corrected) were computed for all direct and indirect effects. The direct effects of T1 BJW-other on all T2 outcomes were controlled for.

To align with the analyses of Bartholomaeus et al. (2022), we treated BJW-self and BJW-other as separate latent variables (controlling for BJW-other in all analyses, which is also considered best practice in the field [Bartholomaeus & Strelan, 2016]). All de-identified individual participant data collected during the study, input analysis code, and statistical output are available here: https://osf.io/3p7gn/?view_only=17ee67bae8bd49598163c767b614adf1.

Results

Sample 1

The initial measurement model showed acceptable fit to the data χ2(1793) = 2937.6, p < .001, CFI = 0.929, TLI = 0.925, RMSEA = 0.046, CI90% [0.043, 0.049], SRMR = .065Footnote 3. The structural model (as shown in Fig. 1) showed acceptable fit to the data χ2(1799) = 3096.3, p < .001, CFI = 0.919, TLI = 0.915, RMSEA = 0.048, CI90% [0.046, 0.051], SRMR = 0.069Footnote 4. Path estimates (direct and indirect effects) and latent factor residuals are displayed in Table 2 (item loadings and errors are displayed in the supplementary material Table S1).

Table 2 Path Estimates, Indirect Effects, and Latent Factor Variances for the Structure Model fit with Samples 1 and 2
Full size table

BJW-self had a large and positive direct effect on empowerment. Empowerment had large positive direct effects on life satisfaction, optimism, and resilience. Empowerment had large negative direct effects on depression, anxiety, and stress. BJW-self had a large positive indirect effect via empowerment on life satisfaction and optimism, and a positive moderate indirect effect on resilience. BJW-self had a large negative indirect effect on depression and anxiety, and a moderate negative indirect effect on stress. The structural model explained an acceptable proportion of variance in each of the endogenous variables except resilience (which had a standardized disturbance > 0.90).

Sample 2

Comparing the Demographic Characteristics of Samples 1 and 2

Sample 1 had a higher proportion of females (70%) compared to Sample 2 (45%), χ2(3) = 47.4, p < .001. Participants in Sample 1 were younger than Sample 2, t(706.9) = -17.7, p < .001. Additionally, a higher proportion of participants in Sample 1 reported their ethnicity as Australian (83%) compared to Sample 2 participants (76%), χ2(1) = 5.78, p = .016. A smaller proportion of Sample 1 (88%) spoke English as their first language compared to Sample 2 (96%), χ2(1) = 20.2, p < .001, and a larger proportion of Sample 1 (71%) were not in a romantic relationship (46%), χ2(1) = 46.8, p < .001. In contrast, there were no significant differences between samples in participants’ religiousness, χ2(1) = 3.23, p = .072.

Structural Model Replication

The structural model was fit to Sample 2. The model showed acceptable fit to the data, χ2(1799) = 4150.7, p < .001, CFI = 0.918, TLI = 0.914, RMSEA = 0.054, CI90% [0.052, 0.056], SRMR = 0.065. BJW-self had a large and positive direct effect on empowerment. In turn, empowerment had large positive direct effects on life satisfaction, optimism, and resilience. Empowerment had large negative direct effects on depression, anxiety, and stress (Table 2). BJW-self demonstrated a large positive indirect effect on life satisfaction, optimism, and resilience. BJW-self had a large negative indirect effect on depression, anxiety, and stress.

Comparing Model Estimates

The majority of Sample 2 model estimates displayed overlap with those of Sample 1 (Table 2; supplementary material Table S1). The following items, however, loaded significantly higher on their respective latent constructs in Sample 2 compared to Sample 1: empowerment item 2 and item 8, life satisfaction item 3, optimism item 2, resilience item 5, depression item 2, and anxiety items 4 and 5. The direct effects of empowerment on optimism and the indirect effects of BJW-self on resilience were significantly stronger in Sample 2.

Subsample 3

Half-longitudinal Mediation Analysis

Drop out analysis indicated no significant differences between participants who responded and did not respond at T2 on gender, age, ethnicity, language, relationship status, religion, or any of the psychological variables (see supplementary material for full analysis). The direct effects of T1 predictors on T2 outcomes are depicted in Fig. 2. T1 BJW-self had a negligible direct effect on T2 empowerment. T1 empowerment had a large and significant direct effect on T2 optimism and a moderate and significant direct effect on T2 resilience. T1 empowerment had moderate and non-significant direct effects on all other outcome variables.

Fig. 2
figure 2

The direct effects of T1 predictors on T2 outcomes.

All path estimates are standardised beta coefficients with CI95% in square brackets. Control of BJW-other not depicted. BJW-self = belief in a just world for the self

Full size image

T1 BJW-self had negligible and non-significant indirect effects on all T2 outcome variables: T2 life satisfaction, β = .005, p = .81, CI95% [-.034, .072]; T2 optimism, β = .008, p = .81, CI95% [-.036, .049]; T2 resilience, β = .005, p = .81, CI95% [-.022, .032]; T2 depression, β = − .007, p = .80, CI95% [-.054, .027]; T2 anxiety, β = − .004, p = .83, CI95% [-.037, .010]; and T2 stress, β = − .005, p = .81, CI95% [− .038, .015].

T1 BJW-self had moderate but non-significant direct effects on T2 optimism, β = .110, p = .28, CI95% [-.070, .196]; T2 depression, β = − .212, p = .08, CI95% [-.303, .019]; and T2 anxiety, β = − .147, p = .29, CI95% [-.218, .061]. T1 BJW-self had a small direct effect on T2 stress, β = − .084, p = .52, CI95% [-.206, .089]. T1 BJW-self had negligible direct effects on T2 life satisfaction, β = .013, p = .90, CI95% [-.232, .208] and T2 resilience, β = .005, p = .96, CI95% [-.124, .152].

Discussion

Our findings provide both confirmatory and ambiguous evidence for the empowering function of BJW-self as demonstrated in experimental research of Bartholomaeus et al. (2023a). The findings presented here from Samples 1 and 2, suggest that the empowering function of BJW-self has a broad association with positive and healthy psychological functioning and that these associations are present and detectable outside of a highly controlled experimental setting. This finding was generalisable, insofar as the latent variable model largely replicated across two demographically different samples. Although eight of the total 69 item loadings (12%) and one direct and one indirect effect differed between Sample 1 and Sample 2, there was no discernible pattern to these differences. When comparing so many parameters a level or random variation is to be expected. It is likely that the differences between the models can be attributed to this random variation rather than inaccuracies in the model. These findings broadly cement the role of BJW-self as a positive psychological resource in everyday life (as suggested by Dalbert, 2001). These findings also align with past research which suggest that control, a conceptually similar construct to empowerment, mediates the association between BJW-self and positive outcomes (Fischer & Holz, 2010; Scholz & Strelan, 2020; Kiral Ucar et al., 2019; Yu et al., 2018).

Longitudinal Findings

The cross-sectional findings presented here, however, must be scrutinized in light of the null findings of the longitudinal analysis. It is difficult to understand why both experimental findings (Bartholomaeus et al., 2023a) and the cross-sectional findings presented here conform to the theoretical assertions of the empowering function of BJW-self and yet the longitudinal findings do not. There are two possible explanations for these contrary results.

The first explanation is statistical: the associations between these constructs overtime may have been masked by the mechanics of the statistical analysis. To expand, when predicting the association between BJW-self at T1 with T2 outcomes, we controlled for the T1 measurement of those outcome variables (as prescribed by Cole & Maxwell, 2003). Controlling for T1 measurements, however, ensures that T2 measurements effectively only represent changes in variance (or scores) between the two measurement points. In other words, the T2 measurements become estimates of change over time. If traits, such as the endorsement of BJW-self and empowerment are stable (i.e., change very little) over time, then there would be no increased variance at T2 to predict, as the T1 measure would account for the majority of the T2 measurement’s variance. Indeed, the test-retest correlation for empowerment indicates that it is highly stable over time. Thus, the stability in these dispositional constructs over time forms one explanation for the lack of observed associations between BJW-self and empowerment in the longitudinal data.

The second explanation that may account specifically for the null relationship between BJW-self and empowerment over time is that the measurement of these two constructs is highly conflated. To elaborate, in Sample 1, BJW-self had the strongest association with empowerment out of all the associations tested (BJW-self scores accounted for 56% of empowerment’s variance). Similarly, in Sample 2 BJW-self had a near perfect association with empowerment, with BJW-self scores accounting for 80% of empowerment’s variance. These strong associations indicate that the measurement tools used in our investigation may not adequately differentiate between the constructs of BJW-self and empowerment. Controlling for the T1 measure of empowerment in the longitudinal analysis removed the shared variance between the BJW-self and empowerment measurements. In the absence of the conflated measurement variance the association between the two constructs diminished, possibly resulting in the small observed association.

The measurement explanation above highlights one of the weaknesses in the field of dispositional BJW research, which is the issue of sound psychometric measurement. Beyond the initial validation studies of the Personal and General BJW scales (Dalbert, 1999) and the BJW-self/other scales (Lipkus et al., 1996)—which are the two most commonly used BJW measures (Bartholomaeus & Strelan, 2019)—only a small number of studies have investigated the psychometric properties of the Dalbert (1999) measures (Bartholomaeus et al., 2023; Esposito et al., 2023; Nartova-Bochaver et al., 2018). Considering the advances in psychometric techniques and understanding of self-report assessment over the last 20 years, it is likely that further investigation of just world measures, using modern techniques, will reveal important insights into the accuracy (or deficiencies) of these measures. Future research should prioritise the re-validation of the just world measures with large samples across cultures and other demographic groups. The accuracy of measurement will translate to more robust empirical findings.

Limitations and Future Directions

In the present investigation the primary confirmatory findings are drawn from cross-sectional samples. It is not possible to draw causal inferences from these data. However, we were able to replicate our model across two samples that differed significantly in gender, age, ethnicity, primary language, and relationship status. This indicates that the observed associations are largely applicable across populations, i.e., robust and generalisable.

The small sample size of Subsample 3 and the resultant lack of power to detect significant effects limits the insights that we can gain from the analysis of these data. However, this analysis warranted reporting as longitudinal research on the temporal change and predictive validity of BJW-self is sparse, but much needed. We are aware of only a handful of BJW studies that employ longitudinal designs (see Bartholomaeus & Strelan, 2019 for a review). Theorizing about the causal role of BJW-self (within the individual differences paradigm) is based predominately on cross-sectional studies. Therefore, longitudinal research is required to empirically validate the theoretical assertions surrounding BJW-self as the causal agent for a variety of adaptive and nonadaptive emotions, cognitions, and behaviours. Further, given that individual differences research in BJW grew out of findings generated by experimental manipulations, longitudinal research is required to further demarcate between the BJW outcomes that can be attributed to the characteristics of the person and the outcomes that can be attributed to the characteristics of the situation.

Conclusion

We have found that endorsement of BJW-self is associated with a broad range of variables indicative of adaptive psychological functioning. These associations were mediated by empowerment in cross-sectional samples but not in the longitudinal sample. Thus, our attempt to generalise the experimental findings of the empowering function of BJW-self to a broader range of outcome variables outside of an experimental setting was partially supported, but also resulted in uncovering some measurement and methodological limitations that are pertinent to the field and should be considered in future research. Specifically, careful investigation into the measurement of BJW as a dispositional variable is one crucially important area for future investigation. Nevertheless, at the individual level, it would appear that humans’ fundamental value of justice has far-reaching implications for their health and wellbeing, and the mechanism of empowerment, while not consistently observed here, appears to play a supporting role in this dynamic.