En Route to Depression: Self-Esteem Discrepancies and Habitual Rumination Wendy J. Phillips and DonaldW. Hine University of New England Abstract Dual-process models of cognitive vulnerability to depression suggest that some individuals possess discrepant implicit and explicit self-views, such as high explicit and low implicit self-esteem (fragile self-esteem) or low explicit and high implicit self-esteem (damaged self-esteem). This study investigated whether individuals with discrepant self-esteem may employ depressive rumination in an effort to reduce discrepancy-related dissonance, and whether the relationship between self-esteem discrepancy and future depressive symptoms varies as a function of rumination tendencies. Hierarchical regressions examined whether self-esteem discrepancy was associated with rumination in an Australian undergraduate sample at Time 1 (N = 306; Mage = 29.9), and whether rumination tendencies moderated the relationship between self-esteem discrepancy and depressive symptoms assessed 3 months later (n = 160). Damaged self-esteem was associated with rumination at Time 1.As hypothesized, rumination moderated the relationship between self-esteem discrepancy and depressive symptoms at Time 2, where fragile self-esteem and high rumination tendencies at Time 1 predicted the highest levels of subsequent dysphoria. Results are consistent with dual-process propositions that (a) explicit self-regulation strategies may be triggered when explicit and implicit self-beliefs are incongruent, and (b) rumination may increase the likelihood of depression by expending cognitive resources and/or amplifying negative implicit biases. Cognitive models of depression presume that vulnerable indi- viduals possess negative self-beliefs that manifest in depres- sion under the stress of negative life events (Abramson, Metalsky, & Alloy, 1989; Beck, 1987). Recent dual-process models of cognitive vulnerability to depression distinguish between the predictive roles of explicit and implicit self- beliefs, which represent outputs of two cognitive systems that interact to precipitate depression (Beevers, 2005; Carver, Johnson, & Joormann, 2008; Haeffel et al., 2007). Like dual- process theories from other psychological arenas, implicit self- beliefs are posited to arise from a system that is automatic, spontaneous, effortless, and affect oriented, whereas explicit self-beliefs arise from a system that is deliberate, controlled, effortful, and directed by rule-based learning (for a review, see Evans, 2008). According to Beevers’s (2005) dual-process model, a depression-vulnerable individual’s initial negative cognitive responses to a stressful event reflect activated negative self- schemas associated with the implicit system (Bower, 1981; Ingram, Bernet, & McLaughlin, 1994; Smith & DeCoster, 2000; Teasdale, 1988). Effortful explicit processing may over- ride negative automatic responses and relieve negative affect. However, depression may arise if effortful processes fail to curtail negative implicit responses. Together, hyperactive nega- tive implicit self-referential cognitions and ineffective correc- tive explicit processing may create a feedback loop that results in a downward spiral into depression. Haeffel and colleagues (2007) proposed a similar dual-process model that made the additional proposition that maladaptive explicit processes may independently confer vulnerability to depression, even in the presence of positive implicit self-cognitions. In recent years, researchers have found considerable evi- dence consistent with the view that both implicit and explicit processes play important roles in the etiology of depression (Glashouwer, de Jong, & Penninx, 2012; Haeffel et al., 2007; Joormann, 2009; Phillips, Hine, & Bhullar, 2012). At the same time, parallel lines of research have investigated other potential cognitive risk factors, such as self-esteem discrepancies (Briñol, Petty, & Wheeler, 2006; Creemers, Scholte, Engels, Prinstein, & Wiers, 2013; Pavlickova, Turnbull, & Bentall, 2014; Schröder-Abé, Rudolph, & Schütz, 2007) and depres- sive rumination (Ciesla & Roberts, 2007; Donaldson, Lam, & Mathews, 2007; Spasojevic & Alloy, 2001; Williams & Moulds, 2010), which may be usefully integrated with dual- process models. Integrating these lines of research may Correspondence concerning this article should be addressed to Wendy J. Phillips, School of Behavioural, Cognitive and Social Sciences, University of New England, Armidale, NSW, 2351, Australia. Email: wphilli4@une.edu.au. Journal of Personality ••:••, •• 2014 © 2014 Wiley Periodicals, Inc. DOI: 10.1111/jopy.12141 Journal of Personality 84:1, February 2016 V C 2014 Wiley Periodicals, Inc. DOI: 10.1111/jopy.12141