Addictive Behaviors, Vol. 11, pp. 25-30, 1986 Printed in the USA. All rights reserved. 0306-4603/86 $3.00 + .OO Copyright e 1986 Pergamon Press Ltd REPORTS OF PATERNAL ALCOHOLISM: RELIABILITY ACROSS SIBLINGS KENNETH J. SHER and CAROL DESCUTNER University of Missouri-Columbia Abstract- Recent studies of the correlates of having an alcoholic relative have relied heavily upon study participants’ reports of the drinking behavior of their parents and other near relatives. These investigations have rarely attempted to validate the reports of familial alcohol problems or report on the reliability of information across other family members. The present study examined the reliability of reports of parental drinking behavior in a sample of 88 sibling pairs on versions of the Short Michigan Alcoholism Screening Test adapted so that subjects en- dorsed items pertaining to their father’s (F-SMAST) or mother’s (M-SMAST) drinking be- havior. Because of the relative infrequency of offspring reports of maternal problem drinking in this sample, only data from the F-SMAST were suited for analysis. Siblings’ reports of specific paternal drinking consequences showed relatively low agreement on items requiring in- ference (e.g., the presence or absence of guilt) and relatively high agreement on items referring to specific behavioral acts or consequences (e.g., seeking help, being arrested, driving under the influence). Global judgments of paternal alcoholism and overall scores on the F-SMAST showed adequate levels of reliability. It is concluded that, although it is always desirable to validate offsprings’ reports of relatives’ drinking, such reports tend to agree with those of other siblings, and the use of these reports is reasonable when validation is impossible or unfeasible. Recent studies into the correlates of having an alcoholic relative have relied heavily upon study participants’ reports of parents’ and other near relatives’ drinking behavior and consequences. These investigations into both the correlates of risk for alcoholism (i.e., high-risk studies) and the correlates of familial alcoholism have rarely attempted to validate the reports of familial alcohol problems or report on the reliability of infor- mation across family members. While at least two high-risk studies have sampled off- spring of alcoholics by identifying alcoholic parents and sampling their offspring (Behar et al., 1983; Gabrielli et al., 1982), most high-risk studies (e.g., Schuckit, 1980) have sampled offspring by screening young adult samples for self-reported drinking problems in biological near relatives. Studies of the familial incidence of alcoholism in clinical samples have similarly relied heavily upon the self-report of study participants. For example, Cotton (1979) found that of 39 studies of the familial incidence of alcoholism, 26 relied on data from the patient only. On the basis of the results of those studies that did attempt to validate patients’ reports, Cotton (1979) concluded that the reports of the presence of alcohol- ism in family members tend to be accurate (i.e., false positives are rare), but that pa- tients appear to underreport alcohol problems (i.e., false negatives occur with some frequency). The validity of reports of a near relative’s alcoholism was also assessed by McCauley, Longabaugh, and Gross (1978). These investigators found better agreement between a psychiatric patient’s score on the Michigan Alcoholism Screening Test (MAST; Selzer, 1971) and a family member’s rating of the patient on an adapted version of the MAST (Kappa = .52) than between the patient’s MAST score and the DSM-II dig.gnosis of al- coholism given by the patient’s physician (Kappa = .28). (These Kappas are calculated Preparation of this article was supported in part by a Summer Research Fellowship from the Graduate School, University of Missouri and a New Investigator Research Award AA6182 from NIAAA to the first author. 25