The effect of self-esteem on depressive symptoms among adolescents: the mediating roles of hope and anxiety

Adolescent depression is prevalent in China, with reports indicating 24.6% experiencing depression and 7.4% severe depression. Depression in adolescence is linked to self-harm and suicidal behavior, underscoring the need to clarify risk pathways. Guided by vulnerability models, low self-esteem is considered a risk factor for adolescent depression and may operate through cognitive and emotional vulnerabilities. The study investigates whether hope (a positive cognitive resource concerning goal-directed agency and pathways) and anxiety (a negative affective state commonly comorbid with depression) mediate the link between self-esteem and depressive symptoms. The authors hypothesize: (1) lower self-esteem predicts higher depressive symptoms via lower hope; (2) lower self-esteem predicts higher depressive symptoms via higher anxiety; and (3) these mediating pathways may differ by gender.

Prior work shows adolescence is a period of declining self-esteem and elevated risk for internalizing problems. Cross-sectional and longitudinal studies indicate low self-esteem predicts depressive symptoms. Cognitive vulnerability-transactional stress theory posits that maladaptive cognitions and coping styles heighten depression risk. Low hope, reflecting reduced belief in goal attainment and pathways thinking, is associated with lower self-esteem and higher depression. Anxiety is negatively associated with self-esteem and frequently co-occurs with depression; genetic liability for anxiety relates to depressive symptom trajectories. Gender is a known predictor of adolescent depression, suggesting potential gender-specific mechanisms in the self-esteem–depression link.

Design and participants: A cross-sectional online survey was conducted in a public middle school. Head teachers distributed an online link (WeChat/QQ). Participation was voluntary and unrelated to grades; guardians provided written informed consent. N=431 adolescents aged 13–18 years (M=15.73, SD=0.89); 52% female. All participants received compensation. Measures: (1) Self-esteem: Rosenberg Self-Esteem Scale (10 items, 4-point Likert, higher scores=higher self-esteem), α=0.85. (2) Depressive symptoms: CES-D (20 items, 0–3 scale, higher scores=more frequent symptoms), α=0.91. (3) Hope: Children’s Hope Scale (6 items, 1–6 scale, total 6–36; higher=more hope), α=0.86. (4) Anxiety: Zung Self-Rating Anxiety Scale (20 items, 1–4 scale, higher=greater anxiety), α=0.67. Analysis: Pearson correlations tested associations among gender, age, self-esteem, hope, anxiety, and depression. Structural equation modeling (SEM) examined parallel mediation by hope and anxiety separately for females and males using Mplus v8.3. Model fit indices indicated excellent fit (Females: χ²=324.807, df=6, p<0.001; RMSEA=0.000; CFI=1.000; TLI=1.012; SRMR=0.008. Males: χ²=306.754, df=6, p<0.001; RMSEA=0.000; CFI=1.000; TLI=1.019; SRMR=0.002). Indirect effects were evaluated with bootstrapping (5000 resamples); 95% CIs excluding zero indicated significant mediation (p<0.05).

Correlations: Gender correlated with self-esteem (r=0.12, p<0.05), hope (r=0.17, p<0.01), anxiety (r=−0.10, p<0.05), and depressive symptoms (r=−0.11, p<0.05). Age correlated with depressive symptoms (r=0.12, p<0.05). Self-esteem correlated positively with hope (r=0.69, p<0.001) and negatively with depressive symptoms (r=−0.64, p<0.001) and anxiety (r=−0.33, p<0.001). Hope correlated negatively with depressive symptoms (r=−0.50, p<0.001). Anxiety correlated positively with depressive symptoms (r=0.49, p<0.001). Mediation (females): Self-esteem predicted higher hope (β=1.56, p<0.001) and lower anxiety (β=−0.24, p<0.001). Hope predicted lower depressive symptoms (β=−0.10, p<0.01); anxiety predicted higher depressive symptoms (β=0.59, p<0.001). Self-esteem retained a direct effect on depressive symptoms (β=−0.34, p<0.001). Indirect effects: via hope=−0.16 (25.00% of total effect; 95% CI [−0.28, −0.05]); via anxiety=−0.14 (21.88%; 95% CI [−0.22, −0.08]); total indirect=−0.30 (46.88%; 95% CI [−0.46, −0.17]); total effect=−0.64. Mediation (males): Self-esteem predicted higher hope (β=1.15, p<0.001) and lower anxiety (β=−0.15, p<0.01). Anxiety predicted higher depressive symptoms (β=0.48, p<0.001); hope did not predict depressive symptoms (β=−0.04, p>0.05). Self-esteem retained a direct effect on depressive symptoms (β=−0.54, p<0.001). Indirect effect of anxiety was significant (95% CI [−0.14, −0.03]) and reported as 0.07 (11.48% of total effect); indirect effect of hope was not significant (95% CI [−0.14, 0.03]); total effect=−0.61.

Findings support vulnerability models in which low self-esteem contributes to adolescent depressive symptoms. The study clarifies mechanisms by showing gender-specific mediation pathways: for females, both higher hope and lower anxiety partially explain the protective effect of higher self-esteem on depression; for males, lower anxiety, but not hope, mediates this relationship. Results suggest female adolescents may leverage cognitive resources like hope to mitigate depression, whereas male adolescents may rely less on internal hopeful cognitions and be more affected by anxiety. These differences inform tailored prevention, indicating that interventions enhancing hope and cognitive resilience may be particularly effective for females with low self-esteem, while anxiety-reduction strategies may be more impactful for males.

The study demonstrates that self-esteem is inversely related to depressive symptoms among adolescents and that the mechanisms differ by gender: hope and anxiety partially mediate the association in females, while anxiety alone mediates in males. These insights guide precision mental health strategies that align with gender-specific pathways. Future research should employ longitudinal designs to establish causality, recruit more diverse and representative samples to improve generalizability, and examine potential moderators that may influence or condition the mediating effects of hope and anxiety.

The cross-sectional design precludes causal inference. Convenience sampling from a single public middle school limits generalizability. The study focused on hope and anxiety as mediators; other potential moderators or contextual factors that might condition these mediations were not tested.