The majority of substance use IRT models have been applied to modeling DSM criteria corresponding to specific substances (Gillespie et al., 2007; Harford et al., 2009; Langenbucher et al., 2004; Martin et al., 2006; Saha et al., 2006) or other measures of specific substance-related problems (Alterman et al., 2007; Kahler et al., 2003; Kahler et al., 2004; Kirisci et al., 1994; Krueger et al., 2004; Neal et al., 2006). These studies find that items tend to indicate greater levels of substance misuse severity in females than in males (Kahler et al., 2004; Martin et al., 2006) and greater severity in younger samples compared to older samples (Neal et al., 2006). That is, in females and younger persons, endorsement of items such as DSM criteria indicated a more severe level of substance misuse. These single-substance studies represent an important starting point for conceptualizing the potential effects of age and sex on poly-substance use.
