The purpose of the present study is to investigate differential attrition rates in terms of both demographic characteristics and initial levels of child functioning of participants in North Carolina's system of care evaluation. Participants included 303 families (78 dropped out of the study, a 26% attrition rate). Families dropped out of the evaluation for three main reasons: refusal to participate, moving, or inability to contact the family. The majority of families who dropped out of the study did so after completing only the baseline interview. Differential attrition did not exist in terms of demographic characteristics, but there was some support for possible differential attrition indicated by initial levels of child functioning in some cases. Additionally, higher parental education level predicted longer participation in longitudinal evaluation. Recommendations and implications for applied research are offered. (Contains 3 tables.)