Study subjects were 56 male seniors at one Texas high school. A semantic differential device was used to measure subjects' self-concepts. The 56 responses resulted from 3 mailouts plus followup telephone calls to all 148 males in the class. Seventeen teachers used a four-point rating scale to assess the subjects' chances of coming into contact with the police. Teacher ratings were used to divide the sample into 'good' boys (those with little or no chance of becoming involved with the police) and 'bad' boys. Data were analyzed in two stages. Results did not support the study's hypothesis that bad boys would differ significantly from good boys on self-concept measures. Teachers' stated lack of knowledge about subjects' chances of coming into contact with the police was probably responsible for this lack of variance. Three-fifths of the teachers' rating indicated no knowledge of students' potential. Results also suggested that the study sample had few really bad boys, possibly because many of the worst boys had already dropped out of school. Several modifications of the research design are suggested. Discussions of the study's theoretical background and the high school service area's demographic characteristics are included. Footnotes, tables, a bibliography listing 28 references, and appendixes presenting study instruments and additional results are included.