Reliability of retrospective estimates of alcohol consumption during pregnancy by recovering women alcoholics.

Twenty-six recovering women alcoholics were interviewed concerning their alcohol consumption during a pregnancy prior to their recovery. Several months later they were reinterviewed. Test-retest correlations were found to be high. Reliability was not related to the length of time since the pregnancy or to the test-retest interval. The frequency of consumption reported did not differ significantly from test to retest. However, the reported amount of consumption per occasion was significantly lower at the second interview. Reports of binge drinking were particularly unreliable. Results should be interpreted cautiously because the sample size was small. Although it may not be possible to estimate accurately actual alcohol consumption, it is possible to rank-order consumption reliably. It is not known whether these results may be generalized to other populations of pregnant women alcoholics, or to nonpregnant women and men.