There is a lack of research examining patients' and other group members' agreement about their therapeutic alliance. In the present study, the person-group (P-G) fit model was adopted to predict that the group member symptom reduction will be greater when the group member's and the other group members' perceptions of their alliance to the group-as-a-whole fit and are high. In addition, the effect of disagreement between the group member and the other group members in their perceptions of the group alliance on the individual's outcome was investigated by using response surface analysis. Thirty-two patients from 5 long-term (greater than 1 year) psychodynamic group treatments were studied. Participants filled out the California Psychotherapy Alliance Scale-Group (CALPAS-G; Gaston & Marmar, 1993) and the Outcome Questionnaire-45 (OQ-45; Lambert et al., 1996) monthly over the course of group treatment until termination. As hypothesized, patient's symptom reduction was greater when there was agreement between the group member and the other group members that their alliance to the group as a whole was strong. Contrary to the authors' hypothesis, a lack agreement on alliance to the group as a whole between the patient and the other group members was not related to less symptom reduction, as lack of fit increased, symptom reduction decreased. Also contrary to the authors' hypothesis, when other group members saw their alliance to the group as a whole as stronger than did the patient, there was increased symptom reduction. The findings suggest that the P-G fit model may be a useful strategy to examine group process variables in the psychotherapy context. (Contains 2 tables and 1 figure.)