Shelby L Ferrin and Dr. Gary M Burlingame, Psychology
Group psychotherapy is a growing field of both psychological research and clinical application. Credit for this growth has been ascribed to a variety of factors, among them the efficiency of seeing multiple clients at one time. Previous research has demonstrated that group therapy is as efficacious in the treatment of many disorders as individual therapy while also being more timeand cost-efficient. The efficacy of small group therapy is frequently attributed to the unique properties of group involvement (e.g., cohesion, avoidance, or engagement). However, researchers find it difficult to measure these properties. Even in research areas where many studies have been completed or more instruments are available, there is little evidence of crossvalidation or efforts to reduce overlap.
This study investigated the construct validity of common therapeutic relationships in group therapy using two of the most frequently used measures of small group climate. Specifically, a psychometric comparison of the Group Climate Questionnaire (GCQ-S: MacKenzie, 1983) and the Curative Climate Instrument (CCI: Fuhriman, et al., 1986) was conducted. While the instruments capture different parts of the therapeutic relationship, there is also considerable construct overlap between them. The GCQ, measuring members’ perceptions of group engagement, avoidance, and conflict, is the most commonly used group process instrument in the psychological literature. Likewise, the CCI measures conceptually similar properties (cohesion, catharsis, and insight). Although both instruments have been independently evaluated from a psychometric perspective, this is the first analysis to comparatively examine the construct validity of the two.
This investigation contained 168 group members in twenty small groups (5-15 members) with diagnoses ranging from anxiety, depression, post-traumatic stress disorder, addictions, personality disorders and other mental health issues common to a college population. Orientation of group leaders and intern co-leaders included a variety of theoretical orientations. Participants average age was 22. Following each of the 32 group therapy sessions, the two instruments were distributed to each group member. Each instrument was scored, and data were aggregated by construct, developmental stage of the group, patient demographics (e.g., gender, presenting complaint), and type of group. Construct validity (the degree to which each instrument measures what it purports to) and construct overlap were examined on the aforementioned groupings.
Collected data was explored using several steps. First, data from each instrument were analyzed for internal reliability. Data obtained using the two measures yielded coefficient alphas of .8745 for the GCQ and .9286 for the CCI.
Next, the relationships among scale items and latent variables were explored. Each instrument was first examined in terms of its internal structure and constructs. This was accomplished using exploratory factor analysis and the principle components method of extraction. Factor analysis revealed three latent constructs present in the GCQ that corresponded well to the construct subscales validated by previous research. Eleven of the twelve GCQ items loaded onto their respective subscales. The same process revealed three latent constructs in the CCI. However, these three factors did not correspond completely with the three identified by the authors of the CCI.
Analysis of correlations between the items and subscales of the GCQ and CCI revealed moderately high correlations among the three CCI subscales (r=.687 for cohesion-catharsis, .538 cohesion-insight, .and 713 catharsis-insight) and a high correlation between the theoreticallyrelated Engagement and Cohesion subscales (r=.866)
When the two scales were put together to reflect the overall relationship among group members, factor analysis revealed three latent constructs. The first corresponded strongly to the Engagement subscale of the GCQ while also containing the majority of items from the CCI relating to behaviors and skills considered helpful by group members. This factor was labeled Positive Factors, indicative of the fact that items within the construct involved active engagement by group members in activities they found helpful. The second factor contained five items from the GCQ and two items from the CCI relating to conflict, distrust, and rejection. None of the seven items in this construct correlated positively with items in the first factor, indicating that the two factors are either strongly independent or reversed correlations of the same underlying construct. This second factor was labeled Negative Factors. A the third factor contained four items (three from the GCQ and one from the CCI) related to avoidant behaviors such as blocking off feelings, looking to others for guidance, and seeking social approval. This factor was labeled Avoidance Factors. These three factors together account for the considerable construct overlap between the two instruments. Chi-square for the resulting model was 43, with 24 degrees of freedom.
Further research is called for to validate and replicate the findings of this study and to explore the larger question of what fundamental constructs are present in the therapeutic mechanisms of change in group psychotherapy. As part of a larger battery of group therapy measures, these two instruments may provide a useful comparison of what group members expect to find helpful and what they experience. A nomological network of intercorrelations between instruments would be a substantial step forward in determining which group therapy instruments are most useful in exploring and clarifying the therapeutic mechanisms in group.
References
Fuhriman, A., Drescher, S., Hanson, E., Henrie, R., & Rybicki, W. (1986). Refining the measurement of curativeness: An empirical approach. Small Group Behavior, 17, 186-201.
MacKenzie, K. R. (1983). The clinical application of group measure. In R. R. Dies & K. R. MacKenzie (Eds.), Advances in group psychotherapy: Integrating research and practice (pp. 159- 170). New York: International Universities Press.