Selection for psychotherapy may be improved by identifying predictors of non-responding to treatment, but there are only few studies of non-responding in short-term psychodynamic group therapy. We analyzed potential socio-demographic and clinical predictors in a sample of 239 patients in 39 sessions of psychodynamic group psychotherapy, including self-reported symptoms, personality, and extra-therapeutic events. Non-responding was assessed by the Symptom Check List-90-Revised Global Severity Index (SCL-90-R GSI) according to Jacobson and Truax’s Reliable Change Index. Non-responders represented 44.8% of the sample and these patients were still significantly higher in symptom load at a one-year follow-up. SCL-90-R Interpersonal Sensitivity, Compulsive personality as assessed by Millon Clinical Multiaxial Inventory-II (MCMI-II), adverse life events during the course of treatment (e.g. somatic illness, loss of partner or job), lack of social support, social burden and occupation were all significantly associated with non-responding in the bivariate analyses. However, in the multivariate analysis only adverse life events reached significance, and Interpersonal Sensitivity marginal significance, explaining 12.0% of variance. When only pre-treatment variables were included, Interpersonal sensitivity and MCMI Compulsive personality predicted non-responding. Adverse life events may indicate a need for re-evaluation of the treatment plan. Exclusion of these patients (17.6%) revealed lack of social support, low MCMI Passive–Aggressive personality scores, and presence of SCL-90-R Somatization symptoms to predict non-responding. Social support and scores on symptom and personality disorder scales, which indicate awareness of problems related to the interpersonal domain, may be highly relevant to explore before selection of patients to short-term time-limited psychodynamic group therapy.