Page 62 - PhD GT

consequences showed marginal concordance with no significant difference between the two sets of reports for alcohol consumption. Collaterals reported fewer consequences than patients. Collateral reports only marginally improved the identification of alcohol dependence and the authors concluded that self-report was a valid method for detecting and assessing alcohol problems.
One might have predicted that the elderly would have more problems with recall or that social desirability might work differently for the elderly than it would for the young. However, Tucker et al. (1991) found “excellent agreement” for a group of adults aged over 60 years and their collaterals on self monitoring reports of number of drinking days and quantity of alcohol consumed per day, including temporal patterning of intake. No comparative data with young people were reported in this study.
In the review of 17 studies of self-reported drinking and related behaviour by Babor et al.(1987b) mentioned above, collateral reports were used as the criterion measure. Although correlations varied for measures used, moderate to good, statistically significant correlations were found in all of them. A further five studies were reviewed where self-reported alcohol related behaviour was compared with official records. Moderate to good agreement was consistently found and in the majority of disagreements, self-reports over-reported compared to official reports. The authors conclude that this is likely to reflect the inadequacy of official reports. Eleven further studies were reviewed where self-report was compared with a variety of other objective and subjective criteria, for example breath alcohol tests, blood alcohol tests, mental health professionals’ judgment and a variety of monitoring instruments. In these studies the finding of a positive correlation was much more likely to be affected by the situation, the state of intoxication in the respondent and the quality of judgments that professionals are able to make given the amount of contact and the effects, for example, of tolerance on the behaviour of the intoxicated individual.
In summary, studies of self-reported drinking appear to have shown that, in the great majority of cases, collateral report is consistent with self-report; where there is disagreement there appears to be no systematic direction to the disagreement (Babor et al. 1987b) and thus it cannot be compensated for.
3.3.1 Methodology of validating of self-report: collaterals
A number of the problems of using collateral report to validate self-report are due to the fact that collateral report is usually not itself validated, but presumed (possibly due to prejudice described above) to be the more accurate. Collateral report is likely to be impaired by problems of frequency of contact and by problems of assessment of the drinking level and consequences based
50