The reliability of alcohol abusers' self-reports of drinking and life events that occurred in the distant past

This study investigated the test-retest reliability of 69 alcohol abusers' current reports about their past (approximately 8 years prior to interview) drinking behavior and life events. Drinking behavior was assessed by the Lifetime Drinking History (LDH) questionnaire and life events were assessed using the Recent Life Changes Questionnaire (RLCQ). Reliability coefficients for LDH variables were generally moderate to high (r = .52 to .81). Using empirical criteria, the diagnostic power of the two LDH interviews to classify correctly subjects as either having had or not having had a drinking problem was quite high. The reliability coefficient for the RLCQ was r = .85 and 91.7% of the identified events were reported in both interviews. Similarly high test-retest reliabilities and individual event agreement rates were obtained for the six homogeneous subscales of the RLCQ. Subjects were also asked why they had given inconsistent answers to life events questions in the two interviews. Inconsistencies often resulted from errors in the temporal placement of events or from misunderstanding items, rather than from failure to recall an event; this suggests that some sources of error in recalling life events can be reduced. It is concluded that alcohol abusers' reports of drinking and life events occurring many years prior to the date of interview are generally reliable. This finding is consistent with previous studies showing high test-retest reliabilities for reports of recent drinking and related events.