The studies included in this review had control groups and a repeated-measures design, and training lasted not less than 20 hours. Of 100 studies using measurement immediately after training, 78 detected changes significantly greater than those shown by controls. Of 31 studies using measures completed one month or more after training, 21 also found significant change. The changes most frequently found included more fvaorable self-concept, reduced prejudice, modified scores on Schutz's Fundamental Interpersonal Relationship Orientation-Behavior Test (FIRO-B) and Shostrom's Personal Orientation Inventory (POI), changed behavior as perceived by others not present during training, and a variety of modifications of organizational behavior in which groups are conducted within an organization. Methodological problems requiring further attention include the use of potentially biased observers, inappropriate tests, possible test-sensitization effects, and the use of poorly matched controls.