Having unduly positive self-views is not conducive to physical health

Self-enhancement and physical health: A meta-analysis. Constantine Sedikides. British Journal of Social Psychology, September 6 2022. https://doi.org/10.1111/bjso.12577

Abstract: A prior meta-analysis yielded a positive relation between self-enhancement and psychological health. This article presents the first meta-analysis of the association between self-enhancement and physical health (k = 87; N = 22,415). The meta-analysis relied predominantly on social desirability as an operationalization of self-enhancement and secondarily on comparative judgement and narcissism. Further, the meta-analysis operationalized physical health in terms of self-rated health, symptoms and biomarkers. Overall, self-enhancement yielded a near-zero association with physical health, r = .01. However, this association was more pronounced for comparative judgement (r = .18, k = 6) than social desirability (r = .03, k = 41) or narcissism (r = −.0001, k = 8), and for self-rated health (r = .09, k = 9) than symptoms (r = .01, k = 29) or biomarkers (r = −.13, k = 17). The association between self-enhancement and physical health fluctuates across measures of both constructs calling for more focussed and nuanced investigations.

GENERAL DISCUSSION

I asked whether self-enhancement conduces to physical health and reported the first meta-analysis of self-enhancement association with physical health, a critical outcome that underlies daily functioning and longevity. After aggregating across 87 studies, which included 22,415 participants, the overall association of self-enhancement with physical health was near-zero, with little direct evidence of publication bias. Furthermore, although moderation tests suggest that this association is influenced by the type of self-enhancement and physical health measure implicated, the obtained near-zero effect size was generally robust to different methodological factors and sample characteristics.

Implications

The meta-analysis makes several contributions to the self-enhancement literature. First, by uniquely focussing on physical health and aggregating across a large and diverse set of studies, the meta-analysis provided a comprehensive estimate of the adaptiveness of self-enhancement. It indicated that the overall association of self-enhancement with physical health is rather negligible. Self-enhancement does not appear to reap substantial physical health benefits.

Second, the meta-analysis demonstrated that the association between self-enhancement and physical health fluctuates across measures of both constructs. Self-enhancement yielded a small positive association with self-rated health, but this association was likely inflated by common method variance (i.e. both measures were obtained from the same source; Strickhouser et al., 2017) or the possibility that self-enhancement contaminates ratings of one's health. Consistent with this argument, self-enhancement yielded a near-zero association with other health outcomes, such as diseases, symptoms and obesity, and indeed yielded a small negative association with biomarkers that were statistically significant (i.e. had a 95% confidence interval that excluded 0). Thus, physical health outcomes assessed via self-reports (self-rated health) yielded positive associations with self-enhancement, but physical health outcomes that are assessed more objectively (diseases, biomarkers) yielded near-zero or even negative associations.

Results were partially inconsistent when examining the different conceptualizations and measures of self-enhancement. In support of the argument that different conceptualizations of self-enhancement reflect the same higher-order construct (Sedikides, 2021b; Sedikides & Gregg, 2008; Taylor & Brown, 1988), no significant difference in effect size emerged across these concepts (i.e. narcissism, optimistic bias, social desirability). However, when zeroing on the specific scale used to measure self-enhancement, significant fluctuations in effect size were evident. Frequently used measures such as the MCSD and NPI yielded near-zero associations with physical health, but the BIDR yielded a small positive association with physical health and comparative judgements yielded a small-to-medium positive association with physical health. Although these results should be interpreted with caution due to a relatively small number of effect sizes for some scales (k < 9), they provide preliminary evidence that associations of self-enhancement with physical health are more detectable when measured via some scales than others. Comparative judgement and BIDR may be more direct measures of self-enhancement than the MCSD and NPI, which appear to involve other constructs in addition to self-enhancement (e.g. concern for social approval, global self-esteem, status).

Third, unlike the significant moderation observed for measures of self-enhancement and physical health, effect sizes were largely constant across a variety of sample and methodological characteristics. Along these lines, the association between self-enhancement and physical health was negligible across age, gender, race ethnicity and country. In addition, near-zero effects were obtained regardless of whether prior studies used correlational or between-subjects designs and whether they reported cross-sectional or prospective associations between self-enhancement and physical health. Taken together, the near-zero association between self-enhancement and physical health was largely robust across different samples and methods examined in the literature so far.

Fourth, and more broadly, the meta-analysis synthesized research on self-enhancement and physical health across several disciplines, including social, personality, health, clinical and biological psychology as well as public health, medicine and sociology. Prior reviews of this topic were qualitative and focussed on only a portion of the available research literature (Konrath & Bonadonna, 2014; Segerstrom & Roach, 2008; Taylor & Sherman, 2008). In this comprehensive analysis, studies used a wide variety of measures, samples and research practices. Little direct evidence of publication bias was obtained, with many published effects being close to zero. Nonetheless, substantial variability in effect sizes was observed, with associations between self-enhancement and physical health ranging from medium-to-large negative effects to large positive effects (Figure 2). Taken together, the results indicate that the effects of self-enhancement are variable and context dependent.

Limitations and future directions

Although the meta-analysis provides a comprehensive estimate of the association between self-enhancement and physical health, limitations necessitate additional research. Most of the included studies used socially desirable responding as the measure of self-enhancement (k = 54; 62%). More importantly, of the 87 effect sizes, 41 (47%) were derived from studies that examined the relation between the MCSD and physical health markers. This may be problematic. First, the MCDS is not widely regarded as a hallmark index of self-enhancement. The scale is very similar to the BIDR's impression management subscale; so, the MCSD may assess more closely other-deceptive, rather than self-deceptive, self-enhancement: It may predominantly capture style rather than substance (bias). To the extent that it assesses substance, the scale may be pertinent to defensiveness or neuroticism (Andrews & Meyer, 2003; McCrae & Costa, 1983; Weihs et al., 2000), tapping self-protection (Sedikides, 2012; vanDellen et al., 2011) as opposed to self-enhancement strivings. Finally, the MCDS, despite its overall adequacy, has met with some criticism regarding its validity (Ballard, 1992; Ballard et al., 1988) and reliability (Beretvas et al., 1992; O'Grady, 1988). Of note, the other operationalisation of social desirability, the BIDR, evinced a positive association with physical health (r = .11) and was stronger than that of the MCSD; thus, the BIDR may be a better proxy for self-enhancement than the MCSD.

Narcissism and optimistic bias as well as comparative judgement yielded a small positive association with physical health across six studies. Follow-up work could derive robust estimates of effect size for other, more direct indices of self-enhancement. A review published in 2010 identified 60 of such indices (Hepper et al., 2010) and since then more than a dozen additional indices have been documented (Sedikides, 2020). Some examples include favourable interpretation of ambiguous feedback, selectively approaching individuals who are likely to deliver positive feedback, assuming credit for the successes of the dyad or group, comparing favourably the present self with the past self, ‘holier than thou’ perceptions, and resorting to counterfactual thinking.

The meta-analysis located only a small number of studies in conjunction with each physical health outcome. Follow-up work could clarify the nature and robustness of the association of self-enhancement with each of these specific outcomes, particularly biomarkers, which yielded a small negative association across 17 studies, and self-rated health, which yielded a small positive association across nine studies.

The meta-analysis was also limited to the inclusion of studies published in English. Very few studies examine associations of self-enhancement with physical health outside of Western societies (k = 3; 3%). Research in other countries and cultures is needed to assess whether the near-zero association of self-enhancement with physical health is universal (Church et al., 2014). Moreover, as the current meta-analysis found that most studies in this literature are cross-sectional, longitudinal studies are needed to test whether self-enhancement is associated with changes in physical health across time. Assuming they exist, small effects of self-enhancement on physical health may take years or even decades to manifest. Thus, research testing whether self-enhancement in adolescence predicts later physical health would be especially informative (Steiger et al., 2014). Research examining associations of self-enhancement with health behaviours (Davidson & Prkachin, 1997), such as eating, exercise and sleep, is also necessary to test the adaptiveness of self-enhancement and expand the literature.

Given that this meta-analysis yielded considerable heterogeneity in associations between self-enhancement and physical health, much of it unexplained, primary research is needed to uncover additional moderators. Such research should test whether associations vary across specific aspects or sub-components of self-enhancement. Evidence indicates that psychological health is more strongly associated with (a) inflated views of one's social skills than inflated views of one's intelligence (communal vs. agentic narcissism; Rentzsch & Gebauer, 2019), (b) self-promoting aspects of narcissism (grandiosity) than defensive aspects (hypersensitivity; Edelstein et al., 2012) and (c) perceiving as opposed to merely presenting the self in a positively biased manner (Paulhus, 2002). Thus, it is possible that some aspects of self-enhancement are more strongly associated with physical health than others. Moreover, the meta-analysis suggests that self-enhancement may have both positive and negative pathways to health (that cancel each other out) or may have no association with health (Esterhuizen & Thabane, 2016). Future work would do well to test possible pathways and the conditions under which they occur.

Another consideration for follow-up research is the specificity or match between measures of self-enhancement and physical health. As of present, studies have primarily examined the association between general measures of self-enhancement across domains and specific health outcomes (Taylor et al., 2003). However, the association between attitudes and behaviour is more pronounced when measures of attitudes and behaviour are matched in their specificity (Ajzen & Fishbein, 2005). Further, although global self-esteem often fails to predict specific behaviours, specific self-concepts are more predictive of these behaviours (Swann Jr. et al., 2007). Thus, future work should examine whether specific indices of self-enhancement (e.g. overestimations of one's cardiovascular or metabolic health) predict relevant physical health outcomes over time (e.g. heart disease or diabetes). Researchers should also use measures of self-enhancement at the individual level (e.g. narcissism, social desirability) either instead of or in addition to measures of self-enhancement at the aggregate level (e.g. above average effects), given that the former is better matched to health outcomes, which are also assessed at the individual level, than the latter.