Adequacy of recall of drinking patterns and event occurrences associated with natural recovery from alcohol problems.

The adequacy of recall of drinking and events surrounding natural recovery from alcohol problems was investigated with 21 abstinent and 18 active untreated problem drinkers. During structured interviews, drinking was assessed for 6 years, and events were assessed for 1 year; both assessment periods ended about 6 years before data collection, which corresponded with the beginning of sobriety for abstinent subjects. The accuracy of subject reports of both variable classes was evaluated through comparisons with collateral reports, and the reliability of drinking reports was evaluated by reinterviewing a subset of subjects about 2 months after the first assessment. Reliability was generally high for subjects' drinking reports. Subject-collateral agreements were significant for drinking and event variables that collaterals could readily observe, but not for variables that were less observable. Current drinking status and the length of recall did not influence reporting accuracy. With some qualifications, these findings support the use of structured recall procedures to investigate drinking patterns and environmental circumstances related to natural recovery and highlight limitations in the use of collaterals as a verification method.