• individual self-esteem;
  • group self-esteem;
  • relative self-esteem;
  • group cognitive-behavioural therapy;
  • social comparison


Despite a strong association between individual self-esteem and treatment outcome in group cognitive-behavioural therapy (GCBT), no study has investigated how patient outcomes might be influenced by an individual's self-esteem relative to other group members.


The study comprised a retrospective examination of patients' data and used a multiple regression analysis to identify predictors of treatment outcome. Patients' pre-treatment self-esteem scores were assessed on a continuum and assigned to be low, medium, or high. Therapy groups were assigned to be either low, balanced or high self-esteem groups based on averaged self-esteem scores of participants.


In this study, 3,878 patients who had completed a 10-day intensive cognitive behavioural group therapy programme at a private psychiatric facility were included in the study. The Rosenberg Self-Esteem measure was chosen to assess self-esteem. The three subscales of the Depression Anxiety Stress Scales were used as the outcome measures.


Patient outcomes were influenced by pre-treatment self-esteem scores, such that higher initial self-esteem was associated with better treatment outcomes. Low group self-esteem was predictive of significantly better outcomes for depression, relative to higher self-esteem groups. Additionally, the combined influence of high individual self-esteem and low group self-esteem was associated with significantly enhanced depression improvement.


High self-esteem patients perform better on outcome measures following completion of GCBT. Low self-esteem groups show greater improvement in depression symptoms. Similar results for depression are achieved when patients with high self-esteem complete treatment in low self-esteem groups.

Practitioner points

  • Patients with higher pre-treatment self-esteem perform better on outcome measures compared to individuals with lower self-esteem.
  • Low self-esteem groups perform better with respect to depression improvement. Similar results are found when high self-esteem patients complete treatment in low self-esteem groups.
  • The study was a retrospective examination of patients' data and not a manipulation study; therefore, cause and effect relationships were not able to be determined.
  • The generalizability of the finding to patients in other health settings remains to be determined.