Self-esteem and depression are linked across adolescence and adulthood, but the relationship is more informative than a simple statement that people with depression think poorly of themselves. Prospective studies repeatedly show that individuals reporting lower global self-esteem tend to report more depressive symptoms later, even after earlier depressive symptoms are taken into account. Researchers call this the vulnerability model. A second possibility, the scar model, proposes that depressive episodes undermine subsequent self-esteem. Both processes can occur, and their estimated strength depends partly on how investigators measure change, which populations they study, and whether they distinguish persistent differences between people from fluctuations within the same person. The most influential longitudinal meta-analysis found a stronger prospective association from self-esteem to depression than in the reverse direction, but these observational findings do not establish a single causal mechanism (Sowislo and Orth, 2013; Orth and Robins, 2013).
The clinical question is therefore twofold: how self-evaluation may contribute to vulnerability, and how depressive illness can reshape a person's sense of worth during an episode and recovery. Effective depression care addresses mood, functioning, safety, and maintaining processes, while beliefs about worth and self-criticism may be relevant individualized treatment targets. Low self-esteem is a psychological construct, not a diagnosis, and an isolated self-esteem score cannot establish whether someone has major depressive disorder (NICE, 2022, reviewed 2026; WHO, 2026).
What self-esteem and depression mean in psychological research
Global self-esteem: the evaluative relationship with oneself
Global self-esteem is a person's relatively broad evaluation of personal worth, acceptance, or value. It answers questions such as whether someone generally regards themselves as worthy of respect and regards their life and abilities as having value. Researchers commonly measure global self-esteem using the Rosenberg Self-Esteem Scale and related questionnaires. Such measures capture subjective self-evaluations, and their scores depend on item wording, response context, culture, and the period being assessed. Self-esteem can show both substantial rank-order stability and meaningful change across life (Robins, Hendin, and Trzesniewski, 2001; Orth, Erol, and Luciano, 2018).
Low self-esteem is a lower position on this evaluative dimension. It can involve chronic feelings of inadequacy, an unfavorable view of the self, or a strong tendency to use mistakes as evidence of global personal failure. The severity, persistence, and consequences of these experiences vary. A person may feel inadequate in one valued domain while maintaining a stable overall sense of worth. Conversely, someone may report satisfactory confidence at work and still regard themselves as fundamentally unlovable. These distinctions matter because research often finds stronger prospective links between depression and global self-esteem than between depression and narrowly defined domain-specific evaluations (Orth and Robins, 2013).
For the broader conceptual foundation, see Self-Esteem in Psychology: Definition, Types, and Evidence. This article concentrates on the longitudinal association with depressive symptoms, the interpretive limits of that association, and its relevance to clinical care.
Depression: symptoms, episodes, and disorders
Depression is used in several related ways. A depressed mood is an emotional state; depressive symptoms comprise a measurable pattern of experiences such as reduced pleasure, sadness, fatigue, disturbed sleep, concentration difficulties, guilt, hopelessness, and negative self-evaluation. Major depressive disorder is a diagnostic category established through clinical assessment of symptom pattern, duration, distress or impairment, history, and alternative explanations. Depression can also occur in other clinical contexts. The same questionnaire score may have different meanings depending on physical illness, bereavement, medication, bipolar history, social stressors, and individual baseline functioning (WHO, 2026; VA/DoD, 2022).
Low worth and excessive or inappropriate guilt are possible depressive symptoms, so some conceptual overlap between self-esteem questionnaires and depression questionnaires is expected. Investigators must consider whether an apparent statistical relationship reflects a meaningful pathway, shared negative affect, overlapping wording, or a combination. A longitudinal analysis that controls the earlier level of depression offers more information about temporal precedence than a cross-sectional correlation, but overlapping content and unmeasured influences can persist.
Self-esteem, self-concept, and self-criticism
Self-concept consists of beliefs and representations about one's attributes, identities, roles, and capacities. Self-esteem evaluates that represented self. Self-criticism refers to a style of judging or responding to perceived shortcomings, sometimes with harsh or punitive internal language. These constructs overlap empirically and conceptually yet can follow different trajectories. A person can describe their strengths accurately while criticizing themselves relentlessly; another may experience low global worth without frequent conscious self-attack. Depression research benefits from separating what people believe about themselves, how they evaluate that information, and how they respond emotionally to setbacks.
Association, prediction, vulnerability, and causation
A concurrent association shows that two variables co-occur at the same time. A prospective association shows that one measured variable predicts the later value of another. A vulnerability factor increases expected risk under a statistical model, and a causal vulnerability mechanism implies that changing the proposed factor would change the outcome under otherwise comparable circumstances. Each claim requires progressively stronger evidence. Most self-esteem–depression evidence concerns correlation and prospective prediction; interventions and sophisticated longitudinal models can contribute to causal understanding, but no single published observational coefficient resolves causation.
Three models of the self-esteem–depression relationship
The vulnerability model: low self-esteem precedes later symptoms
The vulnerability model proposes that low self-esteem contributes to susceptibility to depression. Negative self-evaluations may shape the interpretation of ordinary failures, amplify sensitivity to criticism, make social setbacks feel globally diagnostic, and sustain repetitive negative thought. For example, a disappointing exam may be interpreted as information about one performance, or as proof that the entire self lacks worth. The second interpretation can encourage rumination, withdrawal, and hopelessness, especially under continuing stress. The model predicts that lower self-esteem at one measurement occasion will be associated with greater depression later after accounting for earlier depressive symptoms (Orth, Robins, and Roberts, 2008; Kuster, Orth, and Meier, 2012).
The hypothesis does not imply that low self-esteem is sufficient to produce clinical depression. Depression has multifactorial determinants, including prior illness, biological vulnerabilities, developmental histories, adversity, ongoing stress, sleep disruption, relationships, and access to support. A risk factor increases probability within a population; individual outcomes are neither predetermined nor reducible to a self-evaluation score.
The scar model: depression changes later self-esteem
The scar model proposes the reverse direction. A depressive episode can disrupt participation, employment, relationships, daily routines, and valued activities. The person may then remember an extended period of impaired functioning as evidence of personal inadequacy. Depressive thinking may also reduce perceived agency and diminish the subjective value of accomplishments. Even when symptoms improve, negative beliefs learned or consolidated during illness may persist. The scar hypothesis predicts that higher depression at an earlier assessment will forecast lower self-esteem at a later assessment after controlling for prior self-esteem.
A depressive episode may temporarily reduce state self-esteem without producing a persistent trait-level scar. Studies that assess self-esteem only during acute illness cannot determine whether the difference reflects an enduring shift. Other studies find small prospective or longer-run scar associations, and the size of those associations may depend on follow-up, developmental period, severity, measurement, and analytic approach (Sowislo and Orth, 2013; Orth and Robins, 2013).
Reciprocal and common-cause accounts
A reciprocal model allows each construct to influence the other over time, potentially creating self-reinforcing cycles in some circumstances: low self-worth can increase distress following failure, while a period of distress and reduced functioning can further undermine self-worth. An association may also arise when both variables reflect common liabilities, such as relatively persistent negative affect, adverse family environments, chronic stress, or response tendencies. Statistical control for previously measured symptoms and self-esteem can reduce some ambiguities, but it does not eliminate all third variables.
The evidence supports using these models as hypotheses about mechanisms and timescales. A population-average prospective effect does not show that every depressed individual enters the condition through low self-esteem, or that everyone leaving an episode develops a durable self-esteem scar. Clinical assessment must examine the individual's history and context rather than assign a theoretical model from a questionnaire alone.
What the longitudinal evidence establishes
The landmark meta-analysis of prospective studies
Sowislo and Orth synthesized data from 77 longitudinal studies examining depression and 18 studies examining anxiety. They used random-effects meta-analytic models and prospective estimates that controlled for prior levels of the predicted outcome. For depression, the average standardized prospective estimate from self-esteem to later depression was approximately β = −0.16, whereas the reverse estimate from depression to later self-esteem was approximately β = −0.08. Lower self-esteem therefore predicted more depressive symptoms later, and the average association was stronger in that direction. The authors reported no significant moderation of the vulnerability effect by gender, age, measurement instruments, or assessment lag in their examined analyses (Sowislo and Orth, 2013).
A standardized coefficient of −0.16 is a population-level association adjusted in the models included in the synthesis. It does not mean that low self-esteem explains 16% of depression, nor that increasing anyone's self-esteem by a specified amount will reliably prevent a depressive episode. Effects are conditioned by model specification, reliability, sampled populations, and the measures used. Although the reverse estimate was smaller, its nonzero average leaves room for scar processes. Furthermore, the pooled studies were heterogeneous observational investigations rather than randomized tests of changing self-esteem.
Adolescent and young-adult evidence
Orth, Robins, and Roberts analyzed two large longitudinal data sets with repeated assessments across adolescence and young adulthood, including four waves in each study. Earlier self-esteem predicted later depressive symptoms, whereas the reverse prospective pathways were weak or absent in the reported cross-lagged models. Results persisted when the investigators addressed overlap between depression and self-esteem measures. This work helped shift the question from whether the constructs correlate to whether their associations unfold asymmetrically over time (Orth, Robins, and Roberts, 2008).
An adolescent study may capture vulnerabilities during identity exploration, increased sensitivity to peer evaluation, educational transitions, and changing autonomy. It cannot by itself prove that the same pathway is identical in later life, when occupational roles, long-term relationships, physical health, and accumulated depressive episodes can change the context of self-evaluation. Evidence from wider age ranges is important for assessing generalizability.
Longer-term stability and life-course consequences
Longitudinal work following individuals across multiple years shows that self-esteem has a relatively persistent component and changes systematically across development. In the Longitudinal Study of Generations, Orth, Robins, and Widaman analyzed five assessments over 12 years among participants spanning ages 16 to 97; self-esteem showed prospective associations with several later outcomes, including depression. Such studies situate depression within the larger development of relationships, health, satisfaction, and functioning (Orth, Robins, and Widaman, 2012).
A separate meta-analysis of longitudinal self-esteem development spanning ages 4 to 94 reported average increases from childhood through much of adulthood, relative stability across some intervals, and declines at very advanced ages. These mean trajectories describe populations rather than individual destinies. They help explain why risk estimation should account for age, cohort, and the changing meaning of valued roles, without assuming that age alone causes either low self-esteem or depression (Orth, Erol, and Luciano, 2018).
Refinements beyond a simple correlation
The vulnerability interpretation has been examined in studies attempting to separate genuine global self-esteem from narcissistic self-enhancement. Orth, Robins, Meier, and Conger analyzed six longitudinal samples including 2,717 participants and meta-analytically combined the prospective estimates. Their work indicated that the self-esteem–depression association was not merely an artifact of self-aggrandizing responses or narcissistic tendencies (Orth et al., 2016).
Other studies ask whether mean self-esteem level, short-term instability, and contingency on events have independent predictive significance. Such research is important because two individuals may give similar global scores while differing in how strongly their self-worth fluctuates after social feedback. Findings have not established that instability or contingency universally supersede global level as the main longitudinal depression predictor. Both theory and practice require precision about whether one is studying chronic negative evaluation, sensitivity to particular experiences, or variability over time (Sowislo, Orth, and Meier, 2014).
Why newer longitudinal methods matter
Traditional cross-lagged panel models estimate whether people scoring relatively low at one point tend to score relatively high or low on another variable later. Yet conventional models may combine stable between-person differences with within-person deviations. If a participant is chronically lower in self-esteem than peers and chronically higher in depressive symptoms than peers, a conventional lagged coefficient may partly reflect enduring differences rather than a process in which a temporary fall in that person's self-esteem causes a subsequent rise in their depression.
Random-intercept cross-lagged panel models attempt to separate these levels. Within-person analyses ask whether an individual being lower than their own typical self-esteem at one occasion predicts them having greater-than-usual depression later. This is a more direct test of dynamic change but still depends on adequate repeated measurements, lag spacing, measurement invariance, and confounding control. A Norwegian six-wave study following 985 participants from ages 13 to 23 reported vulnerability associations at ages 13–15 and smaller scar associations during later adolescence, alongside strong correlations between stable components of both constructs. These results emphasize developmental timing and the distinction between within-person and between-person processes (Jørgensen, Kristensen, and Breivik, 2023).
The appropriate conclusion is methodological as well as substantive: robust between-person prospective differences need not equal a universal short-term within-person causal mechanism. An article that presents every cross-lagged coefficient as a direct intervention effect would misrepresent the design.
Mechanisms that may connect self-evaluation to depression
Rumination and repetitive negative thinking
Rumination involves repetitive, often passive attention to distress, its causes, and its consequences. Low self-esteem may make negative autobiographical material feel more personally salient and harder to disengage from. A person experiencing rejection might repeatedly ask what the event proves about their fundamental inadequacy, limiting opportunities for problem-solving, rest, or rewarding activity. Kuster, Orth, and Meier examined this pathway across five waves in a sample of 663 people assessed over eight months. Their longitudinal mediation analyses supported rumination as one pathway through which lower self-esteem forecast later depressive symptoms (Kuster, Orth, and Meier, 2012).
Mediation in observational data is an explanatory model, not an experimental demonstration. Rumination can itself be affected by depression, trauma history, personal habits, and ongoing stress. Nevertheless, its role is clinically relevant because psychological therapies often help patients recognize and respond differently to repetitive negative thinking.
Negative self-schemas and interpretation of ambiguous events
Self-schemas are organized patterns of beliefs about oneself that guide attention, memory, and interpretation. Negative self-schemas can promote broad conclusions from limited evidence: a criticism becomes confirmation of global failure; a neutral expression becomes evidence of rejection; an unsuccessful effort becomes proof that future efforts will be pointless. These interpretations can reduce perceived control and hope. Depression can additionally affect memory accessibility, motivation, concentration, and perceptions of reward, so the proposed direction of influence may vary across individuals and phases of illness.
A clinician might ask whether the patient thinks, “I made a mistake,” “I tend to make mistakes in this setting,” or “I am worthless.” The three statements differ in globality, stability, and the possibility of corrective learning. The last is most closely aligned with the global negative self-appraisal relevant to self-esteem. Assessing these distinctions helps identify mechanisms without turning every unfavorable self-belief into a psychiatric symptom.
Withdrawal, reduced activity, and lost reinforcement
Low self-worth may discourage approaching opportunities that could lead to positive social contact or mastery experiences. Avoidance can temporarily protect against anticipated embarrassment while reducing activity, accomplishment, and access to reinforcement. Depressive symptoms can independently cause fatigue, anhedonia, and withdrawal. Thus, reduced engagement may be both an antecedent and consequence of depressive distress. Behavioral activation addresses activity patterns by rebuilding meaningful contact with valued and potentially rewarding experiences, one reason it is included among recommended depression treatments (NICE, 2022).
Social comparison, interpersonal expectations, and rejection sensitivity
Self-worth develops in interpersonal and cultural contexts. Repeated exclusion, hostile evaluation, unstable support, discrimination, or coercive relationships may affect both self-esteem and mood. People who already doubt their worth may also approach uncertain social interactions expecting rejection, which can shape attention and behavior. These processes can create feedback loops, but the evidence for any specific mechanism depends on setting and design. The specialized psychological dynamics of platform-mediated evaluation are addressed in Social Comparison on Social Media: Mental Health, Body Image, and Self-Evaluation.
Self-criticism, shame, and the meaning of failure
Self-criticism is a plausible pathway from perceived personal shortcomings to sadness, shame, helplessness, and avoidance. Yet shame is not identical to low self-esteem, and self-compassion is not simply “high self-esteem.” Self-compassion concerns how one responds to suffering or perceived inadequacy, including acknowledgment of difficulty, care rather than hostility, and recognition that imperfection is part of human life. Intervention research indicates potential benefits for depressive symptoms, but direct evidence that a change in self-compassion interrupts the particular self-esteem–depression causal pathway remains more limited than the evidence for co-occurring improvement (Han and Kim, 2023).
Why psychological mechanisms should be specified, not assumed
The same self-esteem score can reflect different histories. One person may hold longstanding global negative beliefs after chronic criticism; another may experience a temporary decline after illness; a third may report negative self-evaluation as part of a major depressive episode. A mechanism-based assessment attends to timing, triggers, coping patterns, functioning, social context, and prior episodes. Such specificity protects against simplistic explanations that blame the individual for illness or suggest that positive self-statements alone constitute treatment.
Developmental, cultural, and situational variation
Adolescence and emerging adulthood
During adolescence, social evaluation, autonomy, school transitions, peer belonging, and identity exploration become prominent developmental tasks. Self-esteem can respond to changing bodies, aspirations, social feedback, and opportunities for competence. Depression risk is also influenced by family circumstances, genetic and temperamental vulnerabilities, adversity, sleep, substance use, and experiences of exclusion. A developmental account should therefore examine cumulative conditions and trajectories, not assume that adolescent self-doubt inevitably becomes a depressive disorder. Prospective research involving adolescents supports a modest vulnerability association on average, with variation between individuals and contexts (Orth, Robins, and Roberts, 2008; Sowislo and Orth, 2013).
The distinction between ordinary self-evaluative uncertainty and clinically concerning depressive symptoms is especially important for younger people. Persistent loss of pleasure, marked changes in functioning, pervasive hopelessness, or thoughts of suicide require a more comprehensive assessment than a conversation about confidence. When working with adolescents, competent assessment also considers developmental stage, caregivers, privacy, peer environment, neurodevelopmental variation, and the person's own account of what has changed.
Adult transitions and recurring depression
In adulthood, career changes, financial insecurity, parenthood, separation, caregiving, disability, retirement, and changes in belonging can alter the evidence people use to assess their worth. The psychological risk may be particularly high when personal value has become narrowly contingent on one role or external outcome. These links are plausible and clinically observable, but the predictive magnitude of any specific life event varies substantially. Self-esteem is partly stable and partly responsive to changing circumstances, making repeated assessment more informative than a single score (Orth, Robins, and Widaman, 2012).
Recurrent depression poses a related temporal problem. Low self-esteem before a first episode may reflect one vulnerability pattern; low self-esteem during an episode may partly reflect current symptoms; and low self-esteem after remission may indicate residual symptoms, a stable preexisting characteristic, a lasting scar, or ongoing adversity. Only appropriate longitudinal history, symptom assessment, and follow-up can begin to distinguish these possibilities. Treating every post-episode self-evaluation as proof of damage from depression overstates the scar hypothesis.
Culture, social position, and measurement context
Global self-esteem scales were developed within particular research traditions and languages. Terms such as worth, pride, acceptance, failure, and self-respect may be understood differently depending on culture, socioeconomic position, family norms, religious traditions, and social expectations. Some respondents express positive self-evaluation directly, while others avoid strong positive statements even when they feel secure in their value. Language adaptation alone does not guarantee that a scale measures precisely the same construct across groups.
Likewise, social disadvantage should not be psychologically reduced to deficient self-esteem. Exclusion, poverty, discrimination, unsafe environments, and limited access to care are material experiences that can contribute to distress irrespective of how people evaluate themselves. Researchers should test measurement invariance, contextual confounding, and the representativeness of samples before claiming that an association has the same explanation everywhere. The 2013 longitudinal meta-analysis found no significant moderation across some demographic variables examined, but absence of detected moderation is not evidence that all mechanisms are culturally identical (Sowislo and Orth, 2013).
Digital environments and changing social feedback
Online environments can increase exposure to visible measures of approval, curated achievements, comparison opportunities, and hostile feedback. Repeated exposure may affect both self-evaluation and depressive mood in some people, but the direction and magnitude depend on individual experience, platform use, social context, and research design. Cross-sectional correlations between time online, self-esteem, and mood cannot identify a universal platform effect. Digital feedback may amplify already established vulnerabilities, and offline sources of support can moderate experience. The platform-specific literature has its own scientific questions and should remain distinct from the general longitudinal self-esteem–depression model.
Measuring the association without confusing constructs
Rosenberg Self-Esteem Scale and related global measures
The Rosenberg Self-Esteem Scale is among the most widely used tools for global self-esteem research. Its short set of positively and negatively worded items provides an economical summary of global self-evaluation. Construct validation supports its usefulness for group research, but individual score interpretation must attend to translation, response styles, item polarity, reliability, and the selected norm group. Different versions and scoring conventions are used in practice; published cutoffs are context-specific rather than universal diagnostic thresholds. Validation research comparing the Rosenberg scale with a single-item global measure illustrates the importance of construct validity and convergent evidence (Robins, Hendin, and Trzesniewski, 2001).
The instrument should not be used to diagnose depression, infer a history of trauma, or decide whether treatment is necessary. A low score may motivate further discussion about beliefs, distress, and everyday impact. Longitudinal research additionally requires sufficient reliability to detect genuine change rather than random fluctuation, especially when differences are small. Repeated measurement and sensitivity analyses help estimate whether outcomes depend on scale design.
Depression severity questionnaires and clinical interviews
The Patient Health Questionnaire-9 is a common self-report measure of depressive symptom frequency and severity. Its original validation established utility for assessing depressive symptom burden in clinical settings; it does not replace the clinical judgment needed to diagnose a depressive disorder, evaluate differential diagnoses, assess functioning, or determine risk. The PHQ-9 includes a question about thoughts of death or self-harm, requiring an appropriate clinical follow-up when endorsed (Kroenke, Spitzer, and Williams, 2001).
A study that uses a symptom questionnaire and a study that uses a structured diagnostic interview may answer different questions. The first can estimate whether self-esteem predicts symptom score changes; the second may examine incident or recurrent episodes meeting explicit diagnostic criteria. Converting any statistically significant symptom association into a claim about major depressive disorder incidence without evidence of diagnostic outcomes is a category error.
The shared-content problem
Consider a self-esteem item expressing global dissatisfaction with oneself and a depression item expressing feelings of failure or worthlessness. A participant experiencing a period of negative affect may endorse both, partly because the underlying language is related. This overlap can inflate simultaneous correlations. Prospective associations are more compelling when they remain after statistical or measurement adjustments for shared content, but even those adjustments depend on item-level modeling choices. Researchers can use alternative scales, remove overlapping items, model latent variables, and compare cognitive-affective versus somatic depression symptoms. Such safeguards were important in the adolescent prospective research and its subsequent theoretical synthesis (Orth, Robins, and Roberts, 2008; Orth and Robins, 2013).
State, trait, and symptom contamination
Trait self-esteem denotes a relatively persistent pattern of self-evaluation. State self-esteem concerns more immediate feelings that can fluctuate after feedback or mood changes. A scale administered during acute depression may partly reflect state effects, even if the questionnaire was designed as a global trait measure. Person-centered longitudinal analyses, repeated assessments after remission, and latent state-trait models can help distinguish temporary shifts from stable individual differences. Their interpretations remain conditional on sampling intervals and measurement quality.
What longitudinal coefficients mean
When self-esteem at time one predicts depression at time two after adjusting for depression at time one, investigators describe a prospective path. Its coefficient is influenced by initial stability, measurement error, the distribution of scores, follow-up duration, and the inclusion or omission of covariates. Two studies with identical substantive mechanisms can produce different coefficients if one measures change over weeks and the other over years. Conversely, a similar coefficient can reflect several processes. Standardized betas do not directly convey individual risk or prove that a given amount of improvement in self-esteem would produce a clinically meaningful decrease in depression.
Attrition, missing data, and selection effects
Longitudinal cohorts often lose participants over time. If those with more severe symptoms, unstable housing, lower social support, or particular response patterns are more likely to discontinue, the remaining sample may distort estimates. Statistical methods can handle missing data under assumptions, but cannot fully repair every form of selective attrition. Prospective studies also vary in whether they recruit general community samples, students, clinic patients, or individuals with recurrent illness. A finding from community adolescents may be highly relevant to vulnerability while offering only indirect evidence about acute treatment in hospitalized adults.
Confounding and alternative explanations
Self-esteem and depression are associated with numerous other constructs, including negative affectivity, neuroticism, perceived support, stress exposure, coping style, and cognitive distortions. Investigators can statistically account for measured covariates, although adjustment itself may remove part of a genuine pathway if a covariate is a mediator rather than a prior cause. Unmeasured common causes remain possible. Researchers should identify the hypothesized temporal sequence before choosing covariates, avoid treating every adjusted estimate as causal, and report whether competing specifications change conclusions.
Clinical assessment: from self-evaluation to a comprehensive formulation
Questions that distinguish vulnerability from an episode-related change
Clinical history asks when negative self-evaluation first became prominent, whether it preceded depressive symptoms, whether it worsened after repeated episodes, and what it looks like during periods of relative wellness. It also explores specific content: perceived failure, unlovability, social comparison, shame, guilt, loss of agency, or beliefs that achievements never count. Some patients describe harsh self-criticism without a consistently low global self-esteem score; others report persistent worthlessness even when current depressive symptom intensity is modest. Both patterns warrant attention in context.
A clinically useful formulation also considers depressive symptom duration, loss of interest, sleep and appetite changes, psychomotor changes, concentration, energy, medical conditions, substance use, medications, trauma history, anxiety, prior mania or hypomania, and interpersonal and occupational functioning. These topics can change diagnosis and treatment selection. For example, a history of hypomania requires careful evaluation for bipolar-spectrum illness rather than assuming that every depressive period represents unipolar major depressive disorder (VA/DoD, 2022).
Low self-esteem as a symptom versus a risk factor
A person who has criticized themselves for many years and later develops a depressive episode presents a different temporal question from someone whose usual stable self-regard deteriorates sharply when depressive symptoms begin. The first history may be compatible with preexisting vulnerability; the second with symptom-associated state decline, and sometimes a scar. Neither chronology conclusively identifies causation. Clinical work can nevertheless make practical use of the distinction by targeting persistent beliefs, acute symptoms, or both while observing how each responds over time.
The importance of function and impairment
An individual's distress is assessed in relation to everyday function, not the level of perceived worth alone. Someone can live with low self-regard while maintaining relationships and employment, although their experience may still be painful. Another person with normally positive self-esteem may have clinically severe depression marked by anhedonia, exhaustion, disturbed sleep, or suicidality. There is no self-esteem threshold that rules depression in or out.
Suicide risk and immediate safety
Thoughts of being worthless can occur in depression, but they do not automatically indicate an intention to die. Similarly, apparently confident behavior does not rule out suicidal thinking. If someone reports thoughts of suicide, self-harm, hopelessness, or feeling that others would be better off without them, clinicians must assess current thoughts, intent, plans, access to means, prior behavior, protective factors, intoxication, and immediate circumstances using an individualized safety approach. Immediate danger requires emergency assessment and locally available emergency services. Routine symptom monitoring and appropriate response to suicidality are part of guideline-concordant depression care (NICE, 2022; WHO, 2026).
The broader evidence about a statistical connection between self-esteem and depression should never be substituted for an acute suicide-risk assessment. Likewise, a self-esteem worksheet or motivational message is not an appropriate response to a person facing imminent danger.
Treatment implications: what evidence supports
Treating depression as the primary clinical condition
Evidence-based care for depression depends on severity, duration, recurrence, preference, comorbidity, safety, prior treatment, and availability. Clinical guidelines identify psychological interventions such as cognitive behavioral therapy, behavioral activation, interpersonal psychotherapy, and other structured approaches; medications are considered according to clinical presentation and shared decision-making. The NICE guideline differentiates less severe and more severe depression and provides distinct options and recommendations for each. The 2022 VA/DoD guideline similarly addresses initial assessment, treatment selection, monitoring, and advanced care management (NICE, 2022, reviewed 2026; VA/DoD, 2022).
A self-esteem intervention may be helpful alongside depression care when global self-devaluation is persistent and clinically meaningful. The existence of a prospective vulnerability pathway does not justify withholding established depression treatment in order to work on self-esteem first. Conversely, symptom improvement without attention to enduring self-critical assumptions may leave an important source of distress unaddressed. Treatment decisions should be based on the patient's formulation and clinical needs, not a universal one-way theory.
Cognitive behavioral therapy and self-evaluation
Cognitive behavioral work can examine the evidence used for global negative beliefs, the circumstances in which self-criticism is activated, and the behaviors that maintain those beliefs. A person who concludes “I always fail” after one unsuccessful interaction may learn to differentiate event-specific outcomes from global identity judgments, observe selective attention to mistakes, and test alternative interpretations through experience. Behavioral experiments are especially relevant when low self-esteem sustains avoidance of situations that could provide disconfirming feedback.
The clinical aim is calibrated self-appraisal and flexible functioning. Replacing every negative feeling with exaggerated positive affirmations is not the defining mechanism of evidence-based CBT. Moreover, psychotherapy studies commonly target depressive symptoms while measuring self-esteem as a secondary outcome, making it important to distinguish treatment of depression from a specific test of changing self-esteem as the active ingredient.
Meta-analysis of psychotherapy for depression and self-esteem outcomes
Bhattacharya and colleagues conducted a 2023 systematic review and meta-analysis of randomized trials of psychotherapy for adult depression that also measured self-esteem. Nineteen studies including 3,423 participants met their criteria. Pooled effects favored psychotherapy for depressive symptoms and showed a moderate improvement in self-esteem relative to control conditions (reported Hedges' g approximately 0.63 for self-esteem). The authors also found a relationship between the magnitude of improvement in depression and self-esteem. Interpretation requires caution: heterogeneity was very high, relatively few trials were rated low risk of bias, and the confidence interval for sustained self-esteem effects at later follow-up crossed zero (Bhattacharya et al., 2023).
This review supports the clinically useful observation that effective depression psychotherapy can improve self-esteem. It does not prove that the improvement in self-esteem necessarily caused the reduction in depressive symptoms. Both may respond to broader treatment processes such as restored activity, reduced rumination, improved relationships, emotional regulation, and shifts in negative cognitive patterns.
Interventions designed for low self-esteem
Kolubinski and colleagues examined CBT interventions based on Fennell's model of low self-esteem. Their 2018 systematic review located eight relevant studies, with seven contributing to quantitative synthesis. Interventions delivered as weekly sessions showed larger reported self-esteem changes than brief one-day workshops, and depressive symptoms also improved. Nevertheless, the modest evidence base and differences in study design restrict firm statements about general effectiveness or whether these approaches have mechanisms distinct from conventional depression treatment (Kolubinski et al., 2018).
Intervention selection should consider the specificity and burden of the person's difficulties. If low self-esteem is closely linked to trauma, domestic abuse, ongoing discrimination, or severe socioeconomic stress, work solely on thought content may be insufficient. Support, safety, practical assistance, and appropriate specialist care can be as important as cognitive exercises.
Behavioral activation and opportunities for mastery
Behavioral activation focuses on how mood and behavior interact, helping patients notice avoidance and reintroduce meaningful activities despite reduced motivation. It does not require that self-esteem be high before action is possible. For a person who has withdrawn because every effort seems destined to fail, gradual activity can offer direct experience that changes both depressive mood and self-evaluation. A successful action may support a more differentiated appraisal, although the direct degree to which self-esteem mediates the treatment effect remains an empirical question. Behavioral activation is included among guideline-supported treatments for depression (NICE, 2022).
Self-compassion-oriented approaches
A 2023 meta-analysis of 56 randomized controlled trials found that interventions targeting self-compassion produced small to medium short-term improvements in depressive symptoms, anxiety, and stress compared with controls, with smaller follow-up effects for some outcomes. Overall risk of bias was high, and comparisons with active control interventions were less common than studies with weaker comparators. These findings suggest a potentially useful therapeutic approach but do not establish equivalence to first-line treatments for major depressive disorder or prove that increasing self-compassion prevents future episodes through increased self-esteem (Han and Kim, 2023).
Self-compassion-oriented skills may help an individual respond to errors and emotional pain without treating those experiences as evidence of personal worthlessness. They can be incorporated into a broader treatment plan when clinically appropriate, while acknowledging that the relevant intervention literature contains heterogeneity and methodological limitations.
Medication, combination treatment, and shared decisions
Antidepressant medication can be appropriate for certain depressive presentations, particularly where symptom severity, recurrence, patient preference, previous response, or comorbidity support its use. Psychological treatment and medication may also be combined. Changes in self-esteem observed during improvement with medication should not be interpreted automatically as pharmacologic correction of a self-esteem trait; they may reflect changes in mood, energy, cognitive processes, or functioning. The clinical decision concerns the full depressive condition and the person's priorities, with follow-up for benefit, adverse effects, adherence, and safety (NICE, 2022; VA/DoD, 2022).
Monitoring improvement across two domains
Separate outcome tracking is more useful than assuming that all self-related variables change together. A patient may experience improved sleep, activity, and anhedonia before feeling consistently worthy of respect. Another may become less self-critical while residual fatigue and low mood continue. Repeated symptom measures, discussion of personal functioning, and carefully interpreted self-esteem instruments can identify the need to revise the formulation. If suicidal thoughts, severe impairment, emerging mania, or worsening symptoms appear, clinical reassessment takes precedence over a routine self-esteem exercise.
Prevention and relapse: what can responsibly be inferred
Can higher self-esteem prevent depression?
Prospective research makes a credible case that low global self-esteem is a risk marker and possible contributing vulnerability. It is a further causal claim to state that raising global self-esteem would prevent depression. That claim requires adequately powered intervention studies showing that a specified self-esteem intervention changes subsequent depression incidence, preferably with durable follow-up, appropriate comparators, and tests of the proposed mediating pathway. The observational vulnerability literature alone cannot provide that evidence. Current support is stronger for prospective prediction than for any universal prevention prescription (Sowislo and Orth, 2013; Orth and Robins, 2013).
Preventive approaches can reasonably include reducing harsh self-criticism, strengthening stable social support, fostering realistic self-acceptance, protecting against bullying and abuse, and supporting engagement in meaningful activities. Their potential benefit should be framed according to the intervention and population studied, rather than as a promise that individuals can eliminate depression risk by achieving a particular self-esteem level.
Recovery from an episode and possible residual vulnerability
Following symptom remission, patients may still hold durable beliefs of worthlessness, avoid valued situations, or interpret normal setbacks as proof that relapse is inevitable. Such residual processes can be assessed in the context of broader relapse planning. Guidance on relapse prevention recommends attention to warning signs, previous stressors, maintaining coping skills, and considering continuation or additional treatment when clinically appropriate. Monitoring should also explore positive change: some individuals recover a stable self-evaluation as symptoms remit and functioning improves (NICE, 2022).
A setback in self-confidence or self-esteem is not itself a depressive relapse. Relapse concerns the re-emergence of a clinically significant symptom pattern and its impact. An individualized plan can specify when to seek reassessment and which supports to use, without transforming ordinary fluctuations in self-worth into pathology.
Realistic self-regard versus compulsory positivity
Therapeutic progress is better characterized by flexible, evidence-sensitive self-evaluation than by a permanent feeling of being exceptional. Stable self-regard permits recognition of mistakes, moral responsibility, disappointment, and realistic limits while preserving personal dignity. For many individuals, learning that an unsuccessful effort does not define the entire self is more sustainable than repeatedly trying to feel globally superior. The evidence on global self-esteem and depression does not require maximizing every self-rating; it points toward understanding the persistence and consequences of harsh global negative evaluations.
Evidence strength and unresolved scientific questions
Established: a robust association and prospective asymmetry
A negative association between global self-esteem and depressive symptoms is well documented. Large longitudinal syntheses indicate a statistically significant prospective relationship in which lower self-esteem tends to predict greater future depressive symptoms, with an average path stronger than the reverse depression-to-self-esteem path in the landmark meta-analysis. This is established at the level of observational prediction and literature synthesis (Sowislo and Orth, 2013).
Supported but variable: scar effects and reciprocal processes
Some analyses identify a smaller depression-to-self-esteem pathway or support a reciprocal explanation, while others show little evidence of a lasting scar effect after adjustments. Different follow-up periods and the separation of stable between-person differences from within-person variation can change results. The scientific question is not settled by declaring a universal one-directional mechanism (Orth and Robins, 2013; Jørgensen, Kristensen, and Breivik, 2023).
Preliminary: specific mediators and intervention mechanisms
Rumination, avoidance, self-criticism, interpersonal expectations, and negative interpretations offer plausible explanatory pathways. Longitudinal mediation has specifically implicated rumination, but a definitive causal pathway requires stronger methods. Intervention meta-analyses show that psychotherapy for depression can improve self-esteem and that some self-esteem-focused interventions may reduce depressive symptoms. Whether self-esteem change is an independent active mechanism, rather than a correlated treatment outcome, remains incompletely established (Kuster, Orth, and Meier, 2012; Bhattacharya et al., 2023).
Limited: individual forecasting and diagnostic use
Existing population-level coefficients do not justify individualized depression prediction from a global self-esteem score. They do not yield a clinically validated cutoff for depression, a universal causal interpretation, or a guarantee about who will benefit from a specific therapy. Better person-centered data, ecological momentary assessment, diverse clinical populations, and studies that carefully separate traits from states are valuable scientific directions.
Contested: the size and interpretation of within-person causal effects
The most important live methodological disagreement concerns what conventional lagged effects actually represent. Some of the predictive association may stem from enduring differences between people, whereas an intervention targets changes within a person. Random-intercept and other dynamic models improve identification of those questions but bring their own assumptions. An evidence-sensitive review therefore reports both the strong prospective literature and the limits of transferring its coefficients to individual causal claims.
Frequently asked questions
Does low self-esteem cause depression?
Lower global self-esteem predicts higher subsequent depressive symptoms in many longitudinal studies, including a large meta-analysis. This supports low self-esteem as a vulnerability factor. Observational prediction does not prove that low self-esteem is the sole cause, a sufficient cause, or necessarily the direct cause in a particular individual. Depression arises through multiple interacting influences (Sowislo and Orth, 2013).
Can depression permanently lower self-esteem?
Some people report lower self-worth during and after depressive episodes. The scar model proposes a lasting depression-to-self-esteem effect, and some longitudinal estimates support it, but average scar associations are generally smaller than the vulnerability association and are not consistently detected in all studies. Many people regain their usual self-evaluation as they recover. Determining a persistent individual change requires assessment across time.
Is low self-esteem a symptom of major depressive disorder?
Feelings of worthlessness or excessive guilt can be symptoms within a depressive episode, but low global self-esteem also occurs outside depression. It is not a stand-alone diagnostic criterion for a disorder, and a questionnaire score is not an evaluation of depression. Diagnosis relies on the broader symptom pattern, course, differential diagnosis, and impairment (WHO, 2026).
Why can a confident or successful person become depressed?
Occupational competence, public confidence, domain-specific self-efficacy, and global self-esteem are distinct. Depression also involves affective, behavioral, cognitive, biological, and social processes beyond self-evaluation. Someone who performs confidently may still experience profound anhedonia, distress, hopelessness, or a depressive disorder. Self-esteem is one factor among many, not a universal explanation.
Does treating depression improve self-esteem?
Randomized psychotherapy trials synthesized in a 2023 meta-analysis found improvements in both depressive symptoms and self-esteem, with considerable variation across studies and methodological concerns. Symptom relief, changed habits, therapeutic relationships, and changes in self-evaluation may evolve together. The observed improvement does not establish the direction of mediation for every treatment or person (Bhattacharya et al., 2023).
Is CBT for self-esteem the same as CBT for depression?
The approaches can overlap in addressing negative beliefs, rumination, avoidance, and behavioral patterns. Depression treatment covers a broader clinical condition, including mood, function, risk, and comorbidity. CBT interventions specifically based on a low-self-esteem model have some supportive evidence, although their distinctive added effects beyond depression-focused therapies require further study (Kolubinski et al., 2018).
Do unstable or contingent self-esteem predict depression more strongly than low self-esteem?
Instability and contingency are scientifically relevant characteristics, but research does not consistently show that either overrides global self-esteem level as the strongest prospective marker. Comparisons require matched longitudinal designs and sound measurement. A person with relatively stable but low self-esteem and another with high but highly contingent self-esteem may have different psychological difficulties (Sowislo, Orth, and Meier, 2014).
What should someone do if negative self-worth accompanies persistent depression?
Persistent sadness, loss of pleasure, significant functional decline, or other depressive symptoms warrant assessment by a qualified health professional. Care can address current depression, ongoing stressors, and recurring patterns of harsh self-evaluation together. Treatment should be selected through assessment and shared decision-making. Immediate danger, including current intention to act on suicidal thoughts, warrants emergency help rather than reliance on a self-esteem exercise (NICE, 2022; VA/DoD, 2022).
Scientific synthesis and clinical implications
The connection between self-esteem and depression is best understood as a relationship between a relatively broad evaluative self-representation and a complex, fluctuating domain of symptoms and illness. Longitudinal data establish that lower self-esteem often precedes higher depressive symptoms and that the vulnerability pathway is, on average, stronger than the reverse in influential syntheses. Depression can also reshape self-evaluation, especially during illness, and the persistence of such changes varies. Common causes, overlapping measurement content, and differences between stable personal characteristics and within-person change qualify every simple causal account.
For clinical practice, the most useful conclusion is an integrated one. Assessment identifies both depressive symptoms and the form, history, and consequences of negative self-evaluation. Treatment addresses the depressive condition through established approaches while examining persistent self-critical beliefs, rumination, avoidance, social adversity, and functional patterns where relevant. Progress is evaluated through safety, mood, functioning, and self-evaluation as distinguishable domains. This framework preserves the scientific contribution of the vulnerability model while keeping clinical decisions grounded in the whole person (Sowislo and Orth, 2013; Bhattacharya et al., 2023; NICE, 2022).
Related Articles
Cluster Self-Concept, Identity and Self-Esteem — A research-wide treatment of self-concept, personal identity, self-esteem, measurement, and related processes.
Self-Esteem in Psychology: Definition, Types, and Evidence — The core conceptual and measurement foundations for global self-esteem.
Low Self-Esteem: Signs, Causes, and Psychological Effects — Broader negative self-evaluation patterns, influences, and everyday consequences.
Self-Esteem vs. Self-Confidence: Meaning and Differences — Why confidence in an ability and evaluation of personal worth require separate interpretation.
Self-Esteem After a Breakup: Rejection and Recovery — Relationship-loss-specific changes in self-worth and emotional recovery.
Self-Esteem After Rejection: Psychological Responses and Recovery — Explore social rejection, threatened belonging, self-evaluation, and evidence-informed recovery.
Self-Esteem and Social Anxiety: Evaluation, Avoidance, and Self-Worth — Examine fear of evaluation, self-worth, avoidance, and associations with social anxiety.
Self-Acceptance: Meaning, Processes, and Mental Health — Examine acceptance of personal attributes, self-regard, psychological flexibility, and clinical boundaries.
References
Bhattacharya, S., Kennedy, M., Miguel, C., Tröger, A., Hofmann, S. G., and Cuijpers, P. (2023). Effect of psychotherapy for adult depression on self-esteem: A systematic review and meta-analysis. Journal of Affective Disorders, 325, 572–581. https://doi.org/10.1016/j.jad.2023.01.047
Han, A., and Kim, T. H. (2023). Effects of self-compassion interventions on reducing depressive symptoms, anxiety, and stress: A meta-analysis. Mindfulness, 14, 1553–1581. https://doi.org/10.1007/s12671-023-02148-x
Jørgensen, M., Kristensen, S. M., and Breivik, K. (2023). Scar, vulnerability, or both? A longitudinal study of the association between depressive tendencies and global negative self-esteem from early adolescence to young adulthood with gender as a moderating factor. Personality and Individual Differences, 214, 112349. https://doi.org/10.1016/j.paid.2023.112349
Kolubinski, D. C., Frings, D., Nikčević, A. V., Lawrence, J. A., and Spada, M. M. (2018). A systematic review and meta-analysis of CBT interventions based on the Fennell model of low self-esteem. Psychiatry Research, 267, 296–305. https://doi.org/10.1016/j.psychres.2018.06.025
Kroenke, K., Spitzer, R. L., and Williams, J. B. W. (2001). The PHQ-9: Validity of a brief depression severity measure. Journal of General Internal Medicine, 16(9), 606–613. https://doi.org/10.1046/j.1525-1497.2001.016009606.x
Kuster, F., Orth, U., and Meier, L. L. (2012). Rumination mediates the prospective effect of low self-esteem on depression: A five-wave longitudinal study. Personality and Social Psychology Bulletin, 38(6), 747–759. https://doi.org/10.1177/0146167212437250
National Institute for Health and Care Excellence. (2022; reviewed 2026). Depression in adults: Treatment and management (NG222). https://www.nice.org.uk/guidance/ng222
Orth, U., Erol, R. Y., and Luciano, E. C. (2018). Development of self-esteem from age 4 to 94 years: A meta-analysis of longitudinal studies. Psychological Bulletin, 144(10), 1045–1080. https://pubmed.ncbi.nlm.nih.gov/30010349/
Orth, U., and Robins, R. W. (2013). Understanding the link between low self-esteem and depression. Current Directions in Psychological Science, 22(6), 455–460. https://doi.org/10.1177/0963721413492763
Orth, U., Robins, R. W., Meier, L. L., and Conger, R. D. (2016). Refining the vulnerability model of low self-esteem and depression: Disentangling the effects of genuine self-esteem and narcissism. Journal of Personality and Social Psychology, 110(1), 133–149. https://doi.org/10.1037/pspp0000038
Orth, U., Robins, R. W., and Roberts, B. W. (2008). Low self-esteem prospectively predicts depression in adolescence and young adulthood. Journal of Personality and Social Psychology, 95(3), 695–708. https://doi.org/10.1037/0022-3514.95.3.695
Orth, U., Robins, R. W., and Widaman, K. F. (2012). Life-span development of self-esteem and its effects on important life outcomes. Journal of Personality and Social Psychology, 102(6), 1271–1288. https://doi.org/10.1037/a0025558
Robins, R. W., Hendin, H. M., and Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin, 27(2), 151–161. https://doi.org/10.1177/0146167201272002
Sowislo, J. F., and Orth, U. (2013). Does low self-esteem predict depression and anxiety? A meta-analysis of longitudinal studies. Psychological Bulletin, 139(1), 213–240. https://doi.org/10.1037/a0028931
Sowislo, J. F., Orth, U., and Meier, L. L. (2014). What constitutes vulnerable self-esteem? Comparing the prospective effects of low, unstable, and contingent self-esteem on depressive symptoms. Journal of Abnormal Psychology, 123(4), 737–753. https://doi.org/10.1037/a0037770
U.S. Department of Veterans Affairs and U.S. Department of Defense. (2022). VA/DoD clinical practice guideline for the management of major depressive disorder. https://healthquality.va.gov/HEALTHQUALITY/guidelines/mh/mdd/index.asp
World Health Organization. (2026). Depressive disorder (depression). https://www.who.int/news-room/fact-sheets/detail/depression