Personality and Depression
Already in ancient times, the postulate of a strong association between personality characteristics and depression peaked thinkers’ curiosity. Its roots can be traced back to Hippocrates’ theory of melancholic humor, and it continued through Galen’s theory until the rise of psychological science in the twentieth century. Today, personality traits are often regarded as potential vulnerability factors that may contribute to the development of psychopathology, including depression. However, this is only one hypothesis regarding how personality and depression could be interlinked. In this chapter, first the concepts of depression and personality are explained, and cross-sectional evidence supporting the assumed connection between the two constructs is summarized. Thereafter, different ways how depression and personality could be interrelated are described. After that, empirical evidence from recent longitudinal studies that favor specific models is presented. Finally, challenges that researchers face investigating this relationship are discussed.
Depressive disorders are among the most common mental diseases. The lifetime prevalence is estimated at 14.6% for high-income and 11.1% for low- to middle-income countries in the WHO World Mental Health (WMH) surveys (Kessler and Bromet 2013). The consequences of depressive disorders are severe and disabling, including considerable loss of quality of life for patients and their relatives, increased mortality rates, high levels of service use, and immense economic costs.
The DSM-5 (American Psychiatric Association 2013) distinguishes between several depressive disorders and sets diagnostic criteria for each. For instance, the DSM-5 lists nine main symptoms characterizing major depressive disorder (MDD). In order to be characterized as suffering from MDD, the patient must experience not less than five of the following symptoms every day for at least 2 weeks. One of these five symptoms must be either depressed mood or loss of interest or pleasure in activities. The nine symptoms are (1) depressed mood (e.g., sadness, emptiness, hopelessness), (2) markedly diminished interest or pleasure in all or almost all activities, (3) significant weight loss when not dieting or significant weight gain, (4) inability to sleep or oversleeping, (5) psychomotor agitation or retardation, (6) fatigue or loss of energy (7) feelings of worthlessness or excessive or inappropriate guilt (which may be delusional), (8) diminished ability to think or concentrate or indecisiveness, and (9) recurrent thoughts of death (not just fear of dying), recurrent suicidal ideation without a specific plan, or a suicide attempt or a specific plan for committing suicide. Furthermore, it is mandatory that the symptoms cause clinically significant distress or impairment in social, occupational, or other important areas of functioning and that the episode is not due to the effects of a substance or a medical condition. Lastly, the occurrence of these symptoms is not better explained by other disorders (e.g., psychotic disorders), and there has never been any manic or hypomanic episode.
In addition, dysthymic disorder (or dysthymia) and other chronic forms of depression (e.g., “double depression”) have been merged into one single diagnosis of persistent depressive disorder (PDD) in the updated DSM-5 (American Psychiatric Association 2013). The essential feature of a PDD is a depressed mood that occurs for most of the day, for more days than not, for a minimum duration of at least 2 years.
Personality and Personality Disorder
Personality refers to individual differences in characteristic patterns of thinking, feeling, and behaving. Here, we briefly outline several models and approaches to (maladaptive) personality traits that are most relevant in the context of depression.
The five-factor model of personality (FFM) assumes that personality is ordered hierarchically from a large number of specific traits to five general characteristics, extraversion (E), agreeableness, conscientiousness, neuroticism (N), and openness to experience, also known as the Big Five personality traits (McCrae and Costa 1997). Many scholars have suggested that individual differences, especially in N (or the closely related concept of negative emotionality (NE)) or E (or the closely related concept of positive emotionality (PE)), may confer vulnerability to, or resilience against, depression. N/NE is characterized by a general tendency to experience fear, anger, and sadness and a susceptibility to the effects of stress on mood. E/PE is characterized by a general tendency to experience positive emotions (e.g., joy, exuberance), as well as being gregarious and engaged with the environment. In the following, the abbreviations N/NE and E/PE will be used to refer to these constructs.
Aside from the traditional personality field, scholars with a clinical background have suggested various trait-like constructs to describe specific dispositions of personality to depression. These “clinical traits” (Klein et al. 2011) are similar to traditional personality facets insofar that they also emerge in an interactive process that unfolds across the life span and show similar stability. Influential examples are Blatt’s (1974) dimensions of interpersonal relatedness and self-definition and their maladaptive expressions of dependency and self-criticism. Individuals with high levels of dependency rely on others to maintain their sense of well-being through closeness, intimacy, and/or reassurance of their worth. On the other hand, individuals with high levels of self-criticism derive well-being from meeting (or exceeding) their standards or those imposed by others, thereby achieving or maintaining acceptance. Similarly, Beck (1983) suggested two personality trait dimensions strongly related to depression: whereas individuals high in sociotropy are motivated to form and maintain relationships, individuals high in autonomy focus on goals associated with the development of the self, such as achievement. Both pairs of constructs are suggested to be central to the phenotypic variability in depression. An elevated need for interpersonal relatedness (i.e., dependency/sociotropy) results in an increased risk for depression in the face of stressors such as loss or loneliness. In contrast, an increased need for self-definition within interpersonal relationships (i.e., self-criticism/autonomy) may lead to depression when a person experiences performance-related failure, feelings of inferiority, or diminished self-worth. Relating the interpersonal relatedness and self-definition to the FFM, interpersonal relatedness and self-definition are both typically moderately correlated with N, but there are consistently secondary relationships with one or more of the other Big Five personality traits. Hence, interpersonal relatedness and self-definition are not reducible to N but rather reflect different neurotic styles (Luyten and Blatt 2011).
Finally, personality disorders (PDs) can be viewed as extreme variants of personalities and have repeatedly been associated with depression. According to the current model, diagnostic criteria for all PDs require that the personality patterns are pervasive and inflexible, have an early onset (in adolescence or early adulthood), and are stable and of long duration. Currently, the classification system for PDs in the DSM-5 (American Psychiatric Association 2013) is based on ten distinct categories in three clusters: (1) Cluster A, including the paranoid, schizoid, and schizotypal categories; (2) Cluster B, including the antisocial, borderline, histrionic, and narcissistic categories; and (3) Cluster C, including the avoidant, dependent, and obsessive–compulsive categories. It is notable that there have been considerable discussions about the optimal classification system for PDs. As a consequence, the DSM-5 also provides a separate section including an alternative classification system of PDs. Specifically, the alternative PD model in DSM-5 requires both impairments in personality functioning (Criterion A) and maladaptive personality traits (Criterion B) to be present when diagnosing a PD. The trait taxonomy in DSM-5 comprises five broad domains, namely, negative affectivity, detachment, antagonism, disinhibition, and psychoticism. A growing body of research reveals that the DSM-5 traits – with the exception of psychoticism – are maladaptive variants of the Big Five personality traits. For example, negative affectivity shows a large overlap with N/NE, and detachment is inversely related to E/PE. Thus, it is safe to say that the alternative model implies a dimensional model very similar to the FFM that could cover most of the current criteria defining a personality disorder (Krueger and Markon 2014).
In this chapter, only the most influential concepts of personality in relation to depression are described, and the focus is on the scientific literature of the last few years. A more comprehensive review regarding the relationship between personality and depression referring to older literature can be found elsewhere (Klein et al. 2011).
Evidence for Cross-Sectional Associations Between Personality/PD and Depression
Maladaptive personality traits (or excessively high scores on normal personality traits) as well as personality disorders have frequently been associated with depression, and a wealth of research investigated cross-sectional relationships between the two constructs. Of special interest is a meta-analysis by Kotov et al. (2010) on the relationship between the FFM and depression, which revealed that MDD is strongly associated with high N/NE (Cohen’s d = 1.33) and low conscientiousness (d = −0.90). The link to low E/PE was more modest (d = −0.62) and inconsistent, with some studies finding even positive effects. For openness (d = −0.21) and agreeableness (d = −0.14), the effects were negligible. Interestingly, the associations of personality traits with dysthymic disorder (in the nomenclature of the DSM-5: PDD) were all more pronounced (N/NE, d = 1.93; E/PE, d = 1.47; conscientiousness, d = 1.24; agreeableness d = 0.26; openness, d = 0.57).
With regard to relatedness and self-definition, results of a study by Luyten et al. (2007) indicated that patients with MDD showed elevated levels of both dependency and self-criticism compared to normal controls. Furthermore, patients with MDD also showed higher levels of dependency than mixed psychiatric patients. Levels of self-criticism, however, did not differ between patients with MDD and mixed psychiatric patients. Thus, whereas dependency may be more specifically related to MDD, it seems that self-criticism may play a role in a broader range of psychopathology.
With regard to personality disorders, a recent meta-analysis on the comorbidity of PD and MDD (Friborg et al. 2014) indicated that 45% of the patients with a current MDD also received a PD diagnosis. Cluster C diagnoses (i.e., avoidant, dependent, and obsessive–compulsive PD; 30% combined prevalence) and borderline PD (14%) revealed to be especially common. Conversely, using borderline PD as an example, more than four out of five persons who meet diagnostic criteria for borderline PD also report to have a lifetime history of depression (Tomko et al. 2014).
While the abovementioned studies clearly illustrate a link between personality/personality disorders and depression, cross-sectional studies or studies on comorbidity do not sufficiently explain this relationship. For this purpose, theory-driven, longitudinal studies are of great importance. Therefore, we first want to provide an overview of the suggested models of the relationship between personality and depression and subsequently present recent empirical evidence from longitudinal studies in support of the various models.
Theoretical Explanations for the Relationship Between Personality and Depression
In their review on the relationship between personality and depression, Klein et al. (2011) identified seven basic models: (1) the common cause model, (2) the continuum/spectrum model, (3) the precursor/prodrome model, (4) the concomitants model, (5) the consequence/scar model, (6) the predisposition model, and (7) the pathoplasticity model. It is important to note that these models have not been formulated exclusively for the association between personality and depression but for the association between personality and common mental disorders in general (Ormel et al. 2013). Moreover, these models are not mutually exclusive, and different models may apply to different subgroups. With regard to the existence and direction of a causal relationship between the two constructs, the categories explicated in these models can be summarized in three broader categories (e.g., Widiger 2011). Models (1) to (3) assume that specific personality traits and depression are influenced by another variable in terms of having common causes. All these models postulate that there is no causal relation between personality and depression: in the common cause model (1), it is assumed that personality and depression are not directly related, and the association is due to a shared third variable. In the continuum/spectrum model (2), it is assumed that there is a conceptual overlap between depressive disorders and certain personality traits. Both specific personality traits and depression are on the same continuum and arise from a similar set of causal factors. In the precursor/prodrome model (3), it is assumed that specific personality traits are an early manifestation of depressive disorders and both are caused by similar etiologic factors.
Models (4) and (5) assume that a depressive disorder influences personality in the sense that personality features are state-dependent concomitants of depressive episodes or consequences (“scars”) of depressive episodes. More specifically, the concomitants (or state-dependent) model (4) assumes individuals’ reports of personality facets or observations of their personality-relevant behavior are influenced by their current mood state/depression and that the personality returns to its baseline level after recovery from an episode. The consequence (or scar) model (5) postulates also that depression is causal for personality features. However, it differs from the concomitants (or state-dependent) model insofar that depressive episodes have an enduring effect on personality and that these changes persist after recovery from a depressive episode.
Models (6) and (7) assume that personality traits influence disorders in terms of predisposing the development of depressive disorders or having pathoplastic effects on depression. The predisposition model (6) assumes that personality plays a causal role in the onset of depressive disorders. In this view, people with personality disorders have an increased risk of subsequently developing depression, whereas other variables could play a role in mediating or moderating this transition (e.g., stress, negative life events). The diathesis–stress model, as the most common example, conceptualizes personality as the diathesis and stress as a moderator that triggers the onset of depressive disorder. Finally, the psychoplasticity model (7) assumes that personality traits influence the expression of the disorder after onset (severity or pattern of symptomatology, course, and response to treatment). In this view, personality traits can be used to explain variation among depressed individuals in their clinical presentation or treatment outcome.
Empirical Evidence in Support of the Different Models
The strongest support for the common cause (and continuum/spectrum) models comes from twin and family studies. Twin studies indicate that there are substantial associations between the propensity for N/NE and MDD but only weak associations between the genetic liabilities for E/PE and MDD (e.g., Kendler et al. 2006). Additionally, an extensive Norwegian twin study found that depressive, avoidant, and borderline personality disorders and MDD share substantial genetic variance, but that there was also evidence for specific genetic factors (Ørstavik et al. 2007; Reichborn-Kjennerud et al. 2010).
There are numerous studies that are in support of the predisposition model. For example, a recent longitudinal cohort study over 30 years by Hengartner et al. (2016) showed that N/NE assessed in 1988 significantly predicted subsequent MDD and the use of depression treatment in the following 15-year period. Kopala-Sibley and colleagues (2016) recently provided further support for the predisposition model: 318 women completed measures of N/NE and E/PE 5 years prior to a natural disaster, i.e., a hurricane. Adjusting for lifetime history of depressive disorders, higher levels of stress associated with the hurricane predicted elevated levels of depressive symptoms, but only in participants with high levels of N/NE or with low levels of E/PE. These findings add to the previously mentioned literature by demonstrating that N/NE moderates the effects of stressors that occur up to 5 years later. Similarly, in a recent prospective study over 12 years in early adulthood, Wilson and colleagues (2014) found that higher levels of N/NE were associated with the subsequent development of MDD. Lastly, Noteboom et al. (2016) found in the NESDA (Netherlands Study of Depression and Anxiety) study that N/NE predicts both first and recurrent episodes of depression. In addition to this evidence in support of the predisposition model, Conway et al. (2016) investigated the prospective influence of personality traits on the onset and recurrence of depressive disorders over a 4-year interval. In this study, N/NE was shown as a robust predictor of both onset and recurrences of internalizing disorders. Interestingly, the authors found differential predictions of onset and recurrence by lower-order personality traits: whereas self-harm (i.e., reflecting low self-esteem and self-destructive tendencies) predicted recurrences, dependency (i.e., the predisposition toward an external locus of control and limited self-reliance) predicted new onsets. Recently, a study by Kendall and colleagues (2015) provided the first evidence that low E/PE predicted the prospective development of depressive disorders (and also social anxiety disorder and generalized anxiety disorder). However, follow-up analyses revealed that these effects were largely accounted for by the overlap of E/PE with N/NE.
With regard to personality disorders, Skodol et al. (2011) found that PDs seem to have a negative effect on the naturalistic course of depression (i.e., progression independent of treatment). Using epidemiological data, avoidant, borderline, histrionic, paranoid, schizoid, and schizotypal personality disorders were found to be associated with more persistent MDD over 3 years of follow-up.
Consistent with the pathoplasticity model, Wilson and colleagues found that lower E/PE was associated with chronic/recurrent, but not remitting, MDD (Wilson et al. 2014). Furthermore, Naragon-Gainey and colleagues (2013) found in a large treatment-seeking sample with mood and/or anxiety disorders that, after accounting for initial symptom severity and for shared variance between N/NE and E/PE, the time-invariant proportion of N/NE was significantly predictive of change in depression. More specifically, the authors showed that higher levels of N/NE predicted less symptom reduction over time. These results suggest that, if state variance from the data of clinically distressed individuals is removed, the effect of personality on psychopathology shows even more strongly, clearly supporting the pathoplasticity model (i.e., specific personality traits influence the course, severity, or expression of existing psychopathology). Additionally, regarding treatment outcome, in a very recent study by Kushner and colleagues (2016), N/NE was the only pretreatment personality trait that showed a direct effect on treatment outcome, such that higher levels of N/NE predicted poorer treatment outcomes. Furthermore, the authors found a significant indirect effect of agreeableness on treatment outcome via early and late therapeutic alliance. The pathoplasticity model was further supported by a recent review and meta-analysis, showing that patients with comorbid personality disorder and depression are associated with more than double the odds of a poor outcome in depression treatments compared to those with no personality disorder (Newton-Howes et al. 2014).
Finally, there are several studies that have investigated the concomitants (or state-dependent) model or scar models; however, none of these studies found convincing support for these models (De Fruyt et al. 2006; Hengartner et al. 2016; Morey et al. 2007; Ormel et al. 2004).
With some exceptions, most of the studies do not directly compare competing models regarding the association between personality and depression. A hindering fact is that the models presented above are not mutually exclusive. However, regarding N/NE, Ormel and colleagues (2013) tested the prospective relationship of N/NE and depression in a comprehensive meta-analysis. First, the authors found marked differences in effect size between prospective and cross-sectional studies and concluded that approximately half of the cross-sectional associations are due to relations with current mental state. According to the authors, this is inconsistent with the vulnerability model but in line with the state and spectrum models and neutral with regard to the common cause model. Furthermore, none of the tested models in this study were capable to account for all of the prospective N/NE–depression associations. Nevertheless, neither does the evidence completely rule out any of these models, except for the state and the scar models. Ranking the different models based on the existing evidence, the authors infer to have found the strongest evidence for the common cause model and the vulnerability model and the weakest support for the spectrum model. It is noteworthy that this study did not test the pathoplasticity model. Unfortunately, to our knowledge, there are no comparable comprehensive meta-analyses regarding the prospective association between E/PE and depression or other personality traits and depression, making it impossible to conclude that the patterns found in the Ormel et al. study can be generalized to these personality traits.
Challenges for Conceptualization and Assessment
There are at least four significant challenges when it comes to research on the relationship between personality and depression. First, based on the marked heterogeneity of depression symptoms, recent research may question the validity of the current depression diagnosis (e.g., Fried 2015). Consequently, diagnosis-based analyses of the relationship between personality and depression diagnosis risk failing to capture the marked heterogeneity of the syndrome.
Second, a wealth of personality traits and constructs are or could be investigated with regard to their associations with depression. In this chapter, the focus was on the most common personality traits of the FFM and, to some degree, on interpersonal relatedness and self-definition. However, it should be noted that so far most studies have focused on single traits, whereas it may be more important to investigate patterns of traits and/or the interactions between several traits.
Third, whereas there is substantial evidence for a significant relationship between personality and depression, there seems to be a lack of diagnostic specificity regarding the FFM. For example, a recent symptom-based analysis by Watson and Naragon-Gainey (2014) showed that N/NE was strongly positively related to every disorder, conscientiousness was substantially negatively associated with every disorder, and agreeableness and openness were largely unrelated to these disorders. This may be explained by the overlap between many personality and psychopathology constructs (e.g., NE/N with depression and anxiety) and by the fact that many items in widely used personality and psychopathology measures are similar. Nevertheless, it is important to note that several recent studies have demonstrated that personality–psychopathology associations remain after accounting for overlapping item content (e.g., Naragon-Gainey et al. 2013). Relatedly, several authors have suggested that a refinement of the higher-order constructs of N/NE (e.g., Ormel et al. 2013) and E/PE (e.g., Watson et al. 2015) would be helpful to target specific facets of these constructs. With regard to the lack of specificity found for several personality dimensions, a study by Caspi and colleagues (2014) is worth mentioning. Examining the longitudinal structure of psychopathology from adolescence to midlife, the authors found evidence for a general psychopathology factor or p factor (in analogy to the g factor in intelligence research). They speculated that it is difficult to find risk factors specific to disorders because most disorder variance is primarily due to p, leaving little specificity among disorders to be predicted in the first place. As a consequence, future studies should include the p factor (or its personality-related counterpart: Criterion A of the alternative DSM-5 model for PD) in order to control for this nonspecific risk factor.
Lastly, Durbin and Hicks (2014) raised the concern that current models may be seen as being too static and not take into account dynamic developments of personality traits and disorders over time. Arguing from a more developmental perspective, they propose points that could inform future studies on the relationship between personality and psychopathology: among others, these principles entail that trait deviations can only be understood relative to age-related norms, that trait–disorder associations are dynamic and vary depending on developmental context as well as an individual’s disorder-related history, and that the timing and impact of developmental pressures vary across individuals. In sum, the authors suggest that research on the relationship between personality and psychopathology in general needs to move from solely focusing on structures to focusing more on developmental processes.
Overall, a strong link between personality traits and depression has been shown, and current empirical evidence seems to favor the common cause and predisposition models. However, there are still many open questions and this conclusion must remain tentative. It is possible that the introduction of the DSM-5 (e.g., the clear distinction between MDD and PDD or the adoption of the alternative model for PDs with its similarities to the FFM) and the integration of a more developmental perspective lead to a clearer picture of this relationship in the future.
- Beck, A. T. (1983). Treatment of depression: Old controversies and new approaches. In P. J. Clayton & J. E. Barrett (Eds.), Cognitive therapy of depression: New perspectives (pp. 265–290). New York: Raven.Google Scholar
- Hengartner, M., Ajdacic-Gross, V., Wyss, C., Angst, J., & Rössler, W. (2016). Relationship between personality and psychopathology in a longitudinal community study: A test of the predisposition model. Psychological Medicine, 46(8), 1693–1705. https://doi.org/10.1017/S0033291716000210.CrossRefPubMedGoogle Scholar
- Kendall, A. D., Zinbarg, R. E., Mineka, S., Bobova, L., Prenoveau, J. M., Revelle, W., & Craske, M. G. (2015). Prospective associations of low positive emotionality with first onsets of depressive and anxiety disorders: Results from a 10-wave latent trait-state modeling study. Journal of Abnormal Psychology, 124(4), 933–943. https://doi.org/10.1037/abn0000105.CrossRefPubMedPubMedCentralGoogle Scholar
- Klein, D. N., Kotov, R., & Bufferd, S. J. (2011). Personality and depression: Explanatory models and review of the evidence. Annual Review of Clinical Psychology, 7, 269–295. https://doi.org/10.1146/annurev-clinpsy-032210-104540.CrossRefPubMedPubMedCentralGoogle Scholar
- Kopala-Sibley, D. C., Kotov, R., Bromet, E. J., Carlson, G. A., Danzig, A. P., Black, S. R., & Klein, D. N. (2016). Personality diatheses and Hurricane Sandy: Effects on post-disaster depression. Psychological Medicine, 46(4), 865–875. https://doi.org/10.1017/S0033291715002378.CrossRefPubMedGoogle Scholar
- Krueger, R. F., & Markon, K. E. (2014). The role of the DSM-5 personality trait model in moving toward a quantitative and empirically based approach to classifying personality and psychopathology. Annual Review of Clinical Psychology, 10, 477–501. https://doi.org/10.1146/annurev-clinpsy-032813-153732.CrossRefPubMedGoogle Scholar
- Kushner, S. C., Quilty, L. C., Uliaszek, A. A., McBride, C., & Bagby, R. M. (2016). Therapeutic alliance mediates the association between personality and treatment outcome in patients with major depressive disorder. Journal of Affective Disorders, 201, 137–144. https://doi.org/10.1016/j.jad.2016.05.016.CrossRefPubMedGoogle Scholar
- Newton-Howes, G., Tyrer, P., Johnson, T., Mulder, R., Kool, S., Dekker, J., & Schoevers, R. (2014). Influence of personality on the outcome of treatment in depression: Systematic review and meta-analysis. Journal of Personality Disorders, 28(4), 577–593. https://doi.org/10.1521/pedi_2013_27_070.CrossRefPubMedGoogle Scholar
- Reichborn-Kjennerud, T., Czajkowski, N., Røysamb, E., Ørstavik, R., Neale, M., Torgersen, S., & Kendler, K. (2010). Major depression and dimensional representations of DSM-IV personality disorders: A population-based twin study. Psychological Medicine, 40(9), 1475–1484. https://doi.org/10.1017/S0033291709991954.CrossRefPubMedGoogle Scholar
- Tomko, R. L., Trull, T. J., Wood, P. K., & Sher, K. J. (2014). Characteristics of borderline personality disorder in a community sample: Comorbidity, treatment utilization, and general functioning. Journal of Personality Disorders, 28(5), 734–750. https://doi.org/10.1521/pedi_2012_26_093.CrossRefPubMedPubMedCentralGoogle Scholar