The authors discuss the complexities associated with classifying child maltreatment according to substantiation status. Substantiation status is often used as a proxy measure for maltreatment, and can in turn shape the research findings that appear in the literature. Certain jurisdictions and research projects utilize a three-tiered classification system (including substantiated, suspected, and unfounded maltreatment) and others utilize a two-tiered classification system (only including substantiated and unsubstantiated maltreatment). The authors examined data from the 2003 Canadian Incidence Study of Reported Child Abuse and Neglect (CIS-2003; Trocmé et al., 2005), utilizing unweighted data obtained from a national sample excluding Quebec. Bivariate and multivariate analyses were conducted. Certain factors were found to be associated with the decision to classify an investigation as suspected rather than unsubstantiated including referrals made by police, housing risks, uncooperative caregivers, child behavioural issues, caregiver risk factors, and presence of physical/emotional harm to the child. In analysing factors associated with substantiation, the study found that investigations were more likely to be classified as substantiated rather than suspected if the child was physically or emotionally harmed as a result of maltreatment, the referral was made by police, or if there had been a previous substantiated report of maltreatment. The results confirm that in Canada unsubstantiated cases involve situations with fewer risk factors, and they support the use of a three-tiered classification that does not force suspected cases into a substantiated/unsubstantiated dichotomy.