Implicit and explicit self-esteem and their reciprocal relationship with symptoms of depression and social anxiety: A longitudinal study in adolescents

https://doi.org/10.1016/j.jbtep.2013.09.007Get rights and content

Highlights

  • Explicit self-esteem predictive of symptoms of depression at two-year follow-up.

  • Explicit self-esteem predictive of symptoms of social anxiety two years later.

  • Implicit self-esteem not predictive of symptoms of depression and social anxiety.

  • Symptoms of depression and social anxiety not predictive of low self-esteem.

  • Explicit self-esteem interventions may be most effective in preventing symptoms.

Abstract

Background and Objectives

A negative self-view is a prominent factor in most cognitive vulnerability models of depression and anxiety. Recently, there has been increased attention to differentiate between the implicit (automatic) and the explicit (reflective) processing of self-related evaluations. This longitudinal study aimed to test the association between implicit and explicit self-esteem and symptoms of adolescent depression and social anxiety disorder. Two complementary models were tested: the vulnerability model and the scarring effect model.

Method

Participants were 1641 first and second year pupils of secondary schools in the Netherlands. The Rosenberg Self-Esteem Scale, self-esteem Implicit Association Test and Revised Child Anxiety and Depression Scale were completed to measure explicit self-esteem, implicit self-esteem and symptoms of social anxiety disorder (SAD) and major depressive disorder (MDD), respectively, at baseline and two-year follow-up.

Results

Explicit self-esteem at baseline was associated with symptoms of MDD and SAD at follow-up. Symptomatology at baseline was not associated with explicit self-esteem at follow-up. Implicit self-esteem was not associated with symptoms of MDD or SAD in either direction.

Limitations

We relied on self-report measures of MDD and SAD symptomatology. Also, findings are based on a non-clinical sample.

Conclusions

Our findings support the vulnerability model, and not the scarring effect model. The implications of these findings suggest support of an explicit self-esteem intervention to prevent increases in MDD and SAD symptomatology in non-clinical adolescents.

Introduction

Cognitive vulnerability models of depression and anxiety attempt to identify risk factors that increase the likelihood of disorder onset and maintenance. Common across models is the salient role of a negative self-view construct (e.g., self-esteem, self-concept) that stems from the negative inferential style that is characteristic of both depression and anxiety (Sutton et al., 2011). Self-esteem is considered as the baseline self-view from which fluctuations may occur in a given context (Kernis, Grannemann, & Barclay, 1989). During early adolescence, self-views become more negative as the positivity bias that is present during childhood decreases (Baumeister & Tice, 1986), and identity confusion peaks (Erikson, 1968). Therefore, decreased levels of self-esteem during the identity confusion phase may increase vulnerability for depression and anxiety in adolescents. Two prominent mental health disorders during adolescence are social anxiety disorder (SAD) and major depressive disorder (MDD; Wittchen, Nelson, & Lachner, 1998), and lifetime prevalence rates by the age of 18 are 5.5% and 11.2%, respectively (Merikangas et al., 2010). Adolescent SAD and MDD increase the risk for more severe depression and anxiety symptoms in adulthood, as well as suicidal behaviour (Lim et al., 2012, Pine et al., 1998, Wittchen et al., 1999, Zisook et al., 2007). Further knowledge into how symptoms of SAD and MDD develop during adolescence and particularly the role of self-esteem herein may highlight possible areas for prevention.

Previous cross-sectional studies have consistently observed lower self-esteem in those with relatively higher levels of depression and anxiety symptoms, both in adults (e.g., Ginsburg et al., 1998, Hammond and Romney, 1995) and adolescents (e.g., De Jong et al., 2012, Moksnes et al., 2010). Longitudinal studies are more apt for testing vulnerability models. A recent meta-analysis of 95 longitudinal studies (77 on depression, 18 on anxiety) suggests that low self-esteem was predictive of both symptoms of depression and anxiety (Sowislo & Orh, 2013). SAD symptoms, specifically, have also been found to be predicted by self-esteem in adulthood (Acarturk et al., 2009). Longitudinal studies looking at SAD and self-esteem in adolescents could not be found. Although Sowislo and Orh (2013) found age not to be a moderator of the effect size in the relationship between self-esteem and symptoms of depression, it is also important to acknowledge that findings observed in adulthood and late adolescence might not be observed in a younger adolescent sample, particularly since several studies argue that adulthood depression and anxiety differ aetiologically and neurologically from adolescent and child depression and anxiety (e.g., Kaufman, Martin, King, & Charney, 2001).

While the vulnerability model in the current context suggests that relative decreases in self-esteem increases risk for later symptoms of psychopathology, a longitudinal relationship could also, theoretically, occur in the opposite direction. The “scar hypothesis” refers to residual negative cognitions following a depressive episode (Lewinsohn, Steinmetz, Larson, & Franklin, 1981). In the current context, a model based on the scar hypothesis would suggest that self-esteem is lowered as a consequence of depression and anxiety (Zeigler-Hill, 2011). In the meta-analysis by Sowislo and Orh (2013), a significant reciprocal relationship was observed where prediction of self-esteem by depression was weaker than the prediction of depression by self-esteem. Also for anxiety there was a significant reciprocal relationship with self-esteem, yet in this case with both unidirectional relationships being equally strong. As such, there appears to be some support for both the vulnerability and the scar model. If symptoms of SAD and MDD in adolescence affect subsequent self-esteem, it may highlight the need for interventions that target residual negative self-related thoughts following increases in symptomatology.

While a vulnerability and scar model involving self-esteem has been extensively researched, it is important to note that the majority of studies utilise self-report measures of self-esteem like the Rosenberg's (1989) Self-esteem Scale. Self-report measures of self-esteem rely on a person's explicit reflection of their self-worth. Based on current dual processing models (e.g., Beevers, 2005), when cognitive resources are limited and purposeful reflection is not possible, automatic heuristic based processes are adopted. Implicit self-esteem refers to the automatically elicited self-evaluation in a given context that guides and influences behaviour. Usually, one is aware of the output (e.g., a ‘gut’ feeling) while not being aware of the trigger. Explicit self-esteem is said to be adopted when there is motivation, time and cognitive resources to do so. Correlations between implicit and explicit self-esteem tend to be moderate at best (Bosson, Swann, & Pennebaker, 2000), suggesting that for the most part, explicit and implicit self-esteem are two independent concepts. The distinction between explicit and implicit self-esteem suggests the possibility of differential roles in the aetiology of MDD and SAD. While a person with low explicit self-esteem may be able to actively counter negative self-related thoughts by considering positive propositions, implicit self-esteem suggests that the process itself is non-intentional and therefore harder to change. Therefore, differential roles of implicit and explicit self-esteem in the aetiology of MDD and SA would suggest the need for differential interventions.

As implicit self-esteem is a relatively new concept, only few longitudinal studies have been conducted to date. The available studies appear to support the inclusion of implicit self-esteem in the depression vulnerability model. Franck, De Raedt, and Houwer (2007) found that while at a cross-sectional design there were no significant differences between former, current, and never depressed adults, a six month follow-up of depression scores in the former and never depressed adults indicated a significant prediction by implicit, and not explicit, self-esteem. Steinberg, Karpinski, and Alloy (2007) conducted a 4 month longitudinal study with undergraduate psychology students in which (marginally significant) results indicated that implicit self-esteem moderated the association between depressive symptoms and negative life events amongst those with high cognitive vulnerability (i.e., displaying negative cognitive style and high dysfunctional attitudes), while explicit self-esteem failed to contribute to the prediction. With regard to the inclusion of implicit self-esteem in a scar model, the available studies are not particularly conclusive. Franck et al. (2007) found that former depressed and never depressed adults did not differ in implicit self-esteem, suggesting that the scar model does not hold true. However, Risch et al. (2010) found that previously depressed individuals with more than two depressive episodes had significantly lower implicit self-esteem than remitted individuals with less than three episodes. This also partially supports the notion of a reciprocal relationship between self-esteem and depression. If low self-esteem led to high levels of depression, which in turn lowered self-esteem, this would imply a downward spiral of deteriorating symptoms and self-esteem. It is therefore important to note that a vulnerability model and a scar model could occur simultaneously. The main aim of the current study is to test the vulnerability model and the scarring effect of implicit and explicit self-esteem in adolescent MDD and SAD symptomatology. As the scar hypothesis refers to the effect of a depressive episode specifically, we use the term scarring effect to refer to the effect of symptomatology on self-esteem.

Although previous studies support that implicit and explicit self-esteem are two distinct constructs (e.g., Bosson et al., 2000), specific patterns of the two may also be predictors of SAD and MDD symptomatology. Discrepant self-esteem refers to a negative relationship between implicit and explicit self-esteem. Specifically, fragile (also known as “defensive” or “discrepant high”) self-esteem refers to high explicit coupled with low implicit self-esteem and has been linked to narcissistic behaviour (Jordan, Spencer, Zanna, Hoshino-Browne, & Correll, 2003). Damaged (or “discrepant low”) self-esteem, on the other hand, refers to high implicit coupled with low explicit self-esteem and has been linked to more depressive symptoms in adults (Creemers, Scholte, Engels, Prinstein, & Wiers, 2012). Following a social threat activation, damaged self-esteem has been linked to SAD in adolescents (aged 14–20; Schreiber, Bohn, Aderka, Stangier, & Steil, 2012). Discrepant self-esteem in adolescent MDD and SAD (before a social threat activation) has not been researched longitudinally. In order to look at whether discrepant self-esteem adds to the predictability of SAD and MDD symptoms in adolescents, the interaction between implicit and explicit self-esteem is included in the vulnerability model.

There are two broad research aims in the present study of adolescent SAD and MDD symptomatology. First, we aim to test the vulnerability model with implicit and explicit self-esteem as two distinct constructs being potential predictors of subsequent symptoms of MDD and SAD in a large longitudinal cohort study in adolescents. To test the potential added prediction of damaged and/or fragile self-esteem, the interaction between implicit and explicit self-esteem is also included in the vulnerability model. Second, we aim to test the scarring effect with both SAD and MDD symptomatology being predictive of explicit and implicit self-esteem. Further, as gender differences have not only been consistently observed in SAD and MDD (Merikangas et al., 2010), but also in self-esteem (Kling, Hyde, Showers, & Buswell, 1999), we will include gender in the analysis of the vulnerability model and scarring effect in order to control for the possibility of gender differences in the aetiology of MDD and SAD.

Section snippets

Participants

Adolescents in the first and second year of secondary school in the Northern part of the Netherlands (n = 5318) were invited to participate in a large longitudinal study: Prevention of Adolescent Social and Test Anxiety (PASTA; www.projectpasta.nl). Invitations were sent out to adolescents and their parents via the school. Consent from participants and one of their parents/guardians was obtained from 1811 (34%) of those invited. 97.1% were of Dutch nationality, and 68% came from a rural area as

Missing data analysis

Four t-tests were conducted in order to check for differences between those who were present at follow-up and those who dropped out of the study (n = 576). In order to limit the number of t-tests conducted, we checked for differences in variables relevant to the analysis (i.e., baseline IAT, RSES, MDD and SAD). A Bonferroni correction was adopted to account for conducting four t-tests, and therefore we used a significance level of .012. Results can be seen in Table 2. Although the drop-outs had

Discussion

The main findings of the present study can be summarised as follows: (i) There was partial support for the vulnerability model in that low explicit (but not implicit) self-esteem was predictive of relative high MDD and SAD symptomatology at follow-up, even when controlling for baseline symptomatology; (ii) Discrepant self-esteem did not add to the prediction of follow-up MDD and SAD symptomatology as the interaction between explicit and implicit self-esteem showed no independent predictive

Conclusion

In keeping with old and new cognitive models, the present large-scale longitudinal study in adolescents highlights the significance of self-views in the prognosis of MDD and SAD symptomatology. Although it remains to be seen if self-esteem is an aetiological factor, the significant association between explicit self-esteem and future symptoms of MDD and SAD highlights where preventative interventions may potentially be most effective. Persistent negative self-related thoughts would, logically,

Acknowledgements

This research was supported by grant 62200027, ZonMw (Translated: The Netherlands Organisation for Health Research and Development), The Netherlands.

References (54)

  • B. Mayer et al.

    Brief report: direct and indirect relations of risk factors with eating behavior problems in late adolescent females

    Journal of Adolescence

    (2009)
  • K.R. Merikangas et al.

    Lifetime prevalence of mental disorders in U.S. adolescents: results from the national comorbidity survey replication–adolescent supplement (NCS-A)

    Journal of the American Academy of Child & Adolescent Psychiatry

    (2010)
  • U.K. Moksnes et al.

    The association between stress and emotional states in adolescents: the role of gender and self-esteem

    Personality and Individual Differences

    (2010)
  • C.J. Murray et al.

    Alternative projections of mortality and disability by cause 1990–2020: global burden of disease study

    The Lancet

    (1997)
  • S.A. Reijneveld et al.

    Area deprivation affects behavioral problems of young adolescents in mixed urban and rural areas: the TRAILS study

    Journal of Adolescent Health

    (2010)
  • A.K. Risch et al.

    Implicit self-esteem in recurrently depressed patients

    Journal of Behavior Therapy and Experimental Psychiatry

    (2010)
  • F. Schreiber et al.

    Discrepancies between implicit and explicit self-esteem among adolescents with social anxiety disorder

    Journal of Behavior Therapy and Experimental Psychiatry

    (2012)
  • C. Acarturk et al.

    Incidence of social phobia and identification of its risk indicators: a model for prevention

    Acta Psychiatrica Scandinavica

    (2009)
  • L.S. Aiken et al.

    Multiple regression: Testing and interpreting interactions

    (1991)
  • R.F. Baumeister et al.

    How adolescence became the struggle for self: a historical transformation of psychological development

  • T.E. Bodner

    What improves with increased missing data imputations?

    Structural Equation Modeling

    (2008)
  • J.K. Bosson et al.

    Stalking the perfect measure of implicit self-esteem: the blind men and the elephant revisited?

    Journal of Personality and Social Psychology

    (2000)
  • R.N. Carleton et al.

    Brief fear of negative evaluation scale-revised

    Depression and Anxiety

    (2006)
  • C.S. Carver et al.

    Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: the BIS/BAS scales

    Journal of Personality and Social Psychology

    (1994)
  • P.J. De Jong et al.

    Co-occurrence of social anxiety and depression symptoms in adolescence: differential links with implicit and explicit self-esteem?

    Psychological Medicine

    (2012)
  • E.H. Erikson

    Identity: Youth and crisis

    (1968)
  • K. Fiedler et al.

    Unresolved problems with the ‘I’, the ‘A’, and the ‘T’: a logical and psychometric critique of the implicit association test (IAT)

    European Review of Social Psychology

    (2006)
  • Cited by (0)

    View full text