Interpersonal conract has been found to play a role in changing impressions of participants toward certain target groups. A review of recent research (Rabkin, 1972) has indicated that aaitudes of college students, student nurses, hospital aides, and other selected occupational groups toward the mentally ill have been altered as a result of interpersonal conract with patients. It was of interest to the present authors to determine whether or not similar changes in attitudes would occur toward incarcerated delinquents as a result of interpersonal contact. The participants, in this case, were students from Rider College who enrolled in an interim study program during January, 1972. One group of 6 males and 9 females worked in training schools within the Trenton, New Jersey area, spending three to four days per week at an instimtion for four consecutive weeks. Males and females both worked at training schools for boys or girls. Prior to the start of the work experience, but after selection into the program, Ss received a booklet containing five stimulus persons: blind child, mentally retarded child, emotionally disturbed child, and foster home child. Ss were asked to rate the stimulus persons (in the above, or reversed order) on the three dimensions of the Semantic Differential (Osgood, Tannenbaum, & Suci, 1957), Evaluation, Potency, and Activity. The following bi-polar adjectives were used in the same order for each stimulus person : active-passive (Activity), unpleasant-pleasant (Evaluation), hard-soft (Potency), valuable-worthless (Evaluation), cold-hot (Activity), and thick-thin (Potency). Ss in the comparison group were 3 male and 21 female college students who worked with emorionally dismrbed or mentally retarded children in the same interim study course. They received the same booklets at the same time. An analysis of the pretest for the stimulus person juvenile indicated no significant differences between those working with delinquents and the comparison group. At the beginning of the fourth week all Ss were retested. Results of the preversus posttest for the stimulus person juvenile delinquent, using t test for repeated measures, indicated a significant increase in Evaluation (t = 4.23, df = 14, p < .01) and a significant decrease in Potency (t = 4.07, df = 14, fi < .01) for those Ss who worked with delinquents. The only significant change found in the comparison group was an increase in Activity (t = 2.49, df = 23, 9 < .05). We feel that the reduction in Potency scores coupled with the increase in Evaluation scores is important. The delinquent may well have evoked a stereotyped power impression. Under conditions permitting interpersonal contact Ss may have found their fears of physical danger not justified, thus minimizing one of the ncgarlve characteristics of the delinquents. Further research is necessary to show that Evaluat~on changes did not precede Potency changes or that the changes did not occur at the same time. Also, more research should be carried out on the stimulus characteristics of delinquents, primarily race and socia1 class, in order to determine the effect they may have on Potency or Potency-type scales.