The Guilford-Zimmerman Temperament Survey (GZTS) has been described as one of the most widely used tests in marriage counseling. The GZTS was administered to 61 couples when they entered marriage counseling between 1980 and 1985. When counseling ended, couples were divided into three groups--married and together after counseling (MT); married but divorced after counseling (MD); and married but separated after counseling (MS). Husbands and wives were compared within groups. The pattern that emerged for the MT group was a wife who was more serious-minded, less impulsive, and more persistent than her husband. The MD pattern revealed a submissive wife who was friendlier than her husband. Husbands in the MS group scored significantly higher on the impulsive end of the Restraint scale and wives scored higher on the serious-minded side of the scale. Wives were then compared to wives in the three groups, and mean responses by male and female marital groups were compared to GZTS means. The wives in the three groups were found to differ significantly from the norms, while only the husbands in the MT group differed significantly from norms. The three groups differed significantly from one another in several personality traits. These findings could be useful for marriage counselors working with couples having marital problems and for predicting which couples have the best chance for successful marriage counseling. (NB)