Criterion intervals for pretreatment drinking measures in treatment evaluation

SVMMartY. Drinking during a 30-day pretreatment period was found not to be representative of longer pretreatment intervals, especially in a population of se-riously impaired inpatient alcoholics..AT!ENTS ' SELF-REPORTS of drinking and related behaviors are usually the primary source of data in alcoholism treat- ment evaluation studies. Skepticism about such reports abounds, however, because popular beliefs portray alcoholics as minimizing and denying their drinking problems. If self-reports about drinking are inaccurate, then we should rely less on such reports and judiciously qualify any conclusions based on them. Despite the importance ofdetermining the validity of alcoholics ' self-reports, this area has only recently received attention. Data from several studies (1-5) show that when alcoholics are interviewed in a treatment context, their self-reports of verifiable events (e.g., arrests and hospitalizations) are highly valid. The ac-curacy of alcoholics ' elf-reports of drinking behavior has also re