A major purpose of marathon encounter groups is to enable participants to engage subsequently in more gratifying interpersonal relationships, but this effect has not been demonstrated convincingly in the available research literature. The present study attempted to assess such outcomes of marathon groups by utilizing an unusual approach to assessment, in which behavior description and behavior evaluation are distinct steps. This allowed observed behavior changes to be evaluated from more than the usual single viewpoint.

The primary outcome measure was the Scenario Instrument, a self- report instrument developed for this study, which consists of 16 brief, story-like depictions of an interpersonal situation in which the subject is stipulated to be interacting with another person. A point is reached where action by the subject is called for; he is then asked, "What would you do now?" This procedure yielded a set of behavior descriptions, which were then evaluated from each of three viewpoints (the "group movement," traditional psychotherapists, and a "community" viewpoint). Each evaluative viewpoint was the result of pooling the highly homogeneous ratings of six representative judges.

The study utilized a classic "pretest, posttest, control group" design, with two variations: (1) Posttest was split in half, with the first administration immediately following treatment, the second one month later. (2) A second control group (the "fake" group, N = 16) was utilized in addition to the usual test-retest control group (N = 15). This "fake" group was instructed to complete posttest measures as if they had been members of the experimental group; this was designed to test the hypothesis that experimental group results could be duplicated on the basis of expectations alone. The experimental group (N = 22) participated in a marathon encounter group (there were three marathons, led by three different leaders). All subjects were recruited from the population served by the group program of a university counseling center, utilizing leaders and procedures indigenous to the center. No subject had previous marathon group experience.

Results of the study confirmed that marathon participants showed significantly more behavior change than did non-participants. Further, it was found that the "changes" in the "fake" group were significantly different from changes in the experimental group, leading to the conclusion that the changes constitute a genuine treatment effect.

It was not possible, however, to conclude straightforwardly that the observed changes constituted improvement, because it was only from the "group movement" viewpoint that this was demonstrably the case. From the viewpoints of traditional psychotherapists and the community-at-large, the changes were evaluated as neither improvement nor deterioration, but rather as merely different. [176 pp.]