Respondent Characteristics

Assess when the subject is psychiatrically stable. Available evidence suggests that less accurate reports are likely to be obtained on admission interviews or in more acute care settings (e.g., Barbee et al. 1989; Shaner et al. 1993) than among stable outpatients. Even in the absence of an underreporting bias, impairment in reality testing may have unpredictable effects on self-reported substance use patterns. The conditions under which psychotic symptoms impair selfreport accuracy remain to be documented empirically.

Assess when the subject is not intoxicated or in withdrawal. Many studies of primary substance abusers have found less accurate reporting when subjects were intoxicated (e.g., Brown et al. 1992). This finding should also generalize to mentally ill substance abusers. Another lesson from the substance abuse literature suggests using objective measures of intoxication (i.e., breath or urine tests), because even trained interviewers are not always able to detect intoxicated individuals (Sobell et al. 1979).