The data used in this analysis were collected in Toronto in 1987-88. At that time, social service workers estimated that between 10,000 and 15,000 adolescents lived on the streets and shelters in Toronto. The study used a purposive sampling strategy to contact these youths, and over a 1-year period collected data from 390 homeless adolescents. Potential respondents answered seven screening questions and completed a reading assessment test before completing an anonymous self-report survey. Demographic data from the sample show that approximately two-thirds of the sample were male and one-third were female. The analysis focused on relationships between crime and three concepts central to differential association: deviant associations, symbolic elements that support offending, and tutelage in criminal activities. Effects of these variables on two types of crime, drug selling and theft since leaving home, were explored. The two measures used in this study both involve the frequency of exposure to deviant association; however, the first variable also captures elements of priority; whereas, the second measure may reflect intensity. The six questions that measured tutelage were crime- specific and focused on criminal instruction offered by tutors. The study found that models that include deviant associations, attitudes, and desires improve with the addition of a measure of tutelage. Failure to consider the role of tutelage in an analysis of the differential association process can lead to a misinterpretation of findings that correspond with Sutherland's theory. 2 tables and 37 references