that appear in the appendix show differences in the amount of research done on them beyond the original development studies. As research and clinical applications of the diagnosis measures increase, an empirical base will emerge for continued refinement and understanding of the data that the measures provide.

REFERENCES

American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: APA, 1994.

Beutler, L.E., and Clarkin, J.F. Systematic Treatment Selection. New York: Brunner/Mazel, 1990. Blashfield, R.K. Alternative taxonomic models of psychiatric classification. In: Robins, L.N., and Barrett, J.E., eds. The Validity of Psychiatric Diagnoses. New York: Raven, 1989. pp.19-34. Clark, L.A.; Watson, D.; and Reynolds, S. Diagnosis and classification of psychopathology: Challenges to the current system and future directions. Annu Rev Psychol 46:121-153, 1995. Edwards, G., and Gross, M. Alcohol dependence: Provisional description of a clinical syndrome. Br Med J 1:1058-1061, 1976. Frances, R.J.. and Miller, S.I., eds. The Clinical Textbook of Addictive Disorders. New York: Guilford, 1991.

Grant, B.F.; Chou, S.P.; Pickering, R.P.; and Hasin,

D.S. Empirical subtypes of DSM-III-R alcohol dependence: United States 1988. Drug Alcohol Depend 30:75-84, 1992.

Grant, B.F., and Towle, L.H. A comparison of

diagnostic criteria: DSM-III-R, proposed DSM-IV, and proposed ICD-10. Alcohol Health Res World 15:284-292, 1991. Jacobson, G.R. A comprehensive approach to pretreatment evaluation: I. Detection, assessment, and diagnosis of alcoholism. In: Hester, R.K., and Miller, W.R., eds. Handbook of Alcoholism Treatment Approaches. New York: Pergamon, 1989. pp. 17-53.

Keller, M., and Doria, J. On defining alcoholism. Alcohol Health Res World 15:253-259, 1991. Maisto, S.A., and Connors, G.J. Clinical diagnostic

techniques and assessment tools in alcohol research. Alcohol Heath Res World 14:232-238, 1990.

Maisto, S.A.; Galizio, M.; and Connors, G.J. Drug Use and Abuse. 2nd ed. Ft. Worth, TX: Harcourt Brace, 1995.

Marlatt, G.A., and Gordon, J.R. Relapse Prevention. New York: Guilford, 1985.

Martin, C.S.; Kaczynski, N.A.; Maisto, S.A.; Buckstein, O.M.; and Moss, H.B. Patterns of alcohol abuse and dependence symptoms in adolescent drinkers. J Stud Alcohol in press.

Miller, W.R., and Hester, R.K. Treating alcohol problems: Toward an informed eclecticism. In: Hester, R.K., and Miller, W.R., eds. Handbook of Alcoholism Treatment Approaches. New York: Pergamon, 1989. pp. 3-14.

Miller, W.R., and Rollnick, S. Motivational Interviewing. New York: Guilford, 1991. National Council on Alcoholism. Criteria for the diagnosis of alcoholism. Am J Psychiatr 129: 127-135, 1972.

Rinaldi, R.C.; Steindler, E.M.; Wilford, B.B.; and Goodwin, D. Clarification and standardization of substance abuse terminology. JAMA 259:555-557, 1988.

Robins, E., and Guze, S.B. Establishment of diagnostic validity in psychiatric illness: Its application to schizophrenia. Am J Psychiatr 126: 983-987, 1970.

Rosenberg, H. Prediction of controlled drinking by alcoholics and problem drinkers. Psychol Bull 113:129-139, 1993.

Schuckit, M.A.; Irwin, M.R.; Howard, T.; and

Smith, T. A structured diagnostic interview for identifying primary alcoholism: A preliminary evaluation. J Stud Alcohol 49:93-99, 1988. Spitzer, R.L. Psychiatric diagnosis: Are clinicians still necessary? Compr Psychiatry 24:399-411, 1983.

Sullivan, J.T.; Swift, R.M.; and Lewis, D.C. Benzodiazepine requirement during alcohol withdrawal syndrome: Clinical implications of using a standardized withdrawal scale. J Clin Psychopharmacol 11:291-295, 1991.

Todd, R.D., and Reich, T. Linkage markers and validation of psychiatric nosology: Toward an etiologic classification of psychiatric disorders. In: Robins, L.N., and Barrett, J.E., eds. The Validity of Psychiatric Diagnosis. New York: Raven, 1989. pp. 163-175. Wartenberg, A.A.; Nirenberg, T.D.; Liepman, M.R.; Silvia, L.Y.; Begin, A.M.; and Monti, P.M. Detoxification of alcoholics: Improving care by symptom-triggered sedation. Alcohol Clin Exp Res 14:71-75, 1990.

Woodruff, R.A.; Reich, T.; and Croughan, J.L. Strategies of patient management in the presence of diagnostic uncertainty. Compr Psychiatry 18:443-448, 1977.

World Health Organization. Nomenclature and classification of drugs and alcohol-related problems: A WHO memorandum. Bull WHO 59:225-242, 1981.

gate measures of consumption, have long dominated the alcohol field and are usually referred to as quantity-frequency (QF) measures or estimation formulas. Methods where respondents report all their drinking have been referred to as either daily drinking or daily diary methods. With one exception, the measures reviewed in this chapter fall into these two categories.

Although concerns about how to best measure drinking date back at least threequarters of a century (Pearl 1926), serious measurement of drinking began in the late 1960's with the survey work of Cahalan and his colleagues (see Cahalan 1987; Polich and Kaelber 1985; Room 1990). Early measurement of alcohol consumption in surveys was primitive compared to today's standards and focused on arbitrary classifications of drinkers into categories (e.g., abstainers, heavy, light, or moderate drinkers-Room 1990). In the early 1950's, Straus and Bacon (1953) introduced the first QF measure. It asked a series of short questions about the average amount consumed on a typical day by the number of days this occurred in a given period. The first widespread use of this QF measure occurred in the Iowa drinking surveys, although the assessment of drinking was limited to the past 30 days (Fitzgerald and Mulford 1982). Variants of Straus and Bacon's QF measure have been widely used in subsequent years (reviewed in Room 1990).

Over the years, QF measures have advanced considerably, such that second- and third-generation QF methods now exist. While QF measures were being refined, several alternatives to QF measures were being developed. Such alternatives (i.e., retrospective daily recall) involve the assessment of all drinking occurring in a given time period.

Self-Report Issues

Before reviewing specific drinking measures, it is important to address the general issue of the validity and reliability of selfreports, an issue that is ever present when

assessing drinking. The basic question is whether self-reports of drinking can be trusted. Despite the substantial literature on the validity and reliability of self-reports of alcohol use and abuse (reviewed in Babor et al. 1990; Maisto et al. 1990; Midanik 1982; Sobell and Sobell 1990), there appears to be a widespread belief among practitioners that self-reports of alcohol use are not totally accurate, particularly in the direction of underreports or 'denial.' Self-reports have generally been questioned on two levels: underreporting by alcohol abusers and underreporting of consumption in general population surveys. In the first regard, several major reviews of the scientific literature have concluded that alcohol abusers' self-reports are relatively accurate and can be used with confidence if the data are gathered under appropriate conditions (reviewed in Babor et al. 1990; Brown et al. 1992; Maisto et al. 1990; Sobell and Sobell 1990). Factors that enhance accurate self-reporting include: (1) individuals being alcohol free when interviewed, (2) giving individuals assurances of confidentiality, (3) conducting the interview in a clinical or research setting designed to encourage honest reporting (versus, e.g., an interview with a probation officer), and (4) clear, understandable wording of questions.

As reviewed elsewhere (Sobell and Sobell 1990; Sobell et al. 1994), there is one condition when alcohol abusers' self-reports tend to be invalid and underestimated-when clients are interviewed with any alcohol in their system. Interestingly, alcohol abusers have also stated that their reports would be most accurate when they were alcohol free and that their self-reports would likely be increasingly inaccurate as a function of the amount of alcohol they had consumed (Sobell et al. 1992). Lastly, therapists' judgments about clients' level of drinking appear to be no more accurate than the clients' reports (Sobell et al. 1979), probably due to the effect of tolerance. One way to ensure that individuals are alcohol free when interviewed is to use objective measures to assess alcohol use prior to the interview (Sobell et

al. 1994). Several inexpensive portable breath alcohol testers can be used for this purpose.

Questions about the accuracy of reports of drinking in population surveys have focused on bias in aggregate alcohol consumption. Several general population surveys have shown that, when projected to the total population, self-reports of consumption only reflect a portion of the total beverage sales (reviewed in Midanik 1982; Pernanen 1974; Poikolainen 1985). A variety of possible explanations for the discrepancy have been offered, including questionnaire defects, errors in sampling timeframe, and failure to include certain response categories for heavy drinkers and light drinkers (Alanko 1984; Midanik 1982).

A review of several studies shows that with minimal sampling problems and heavy drinking factored into aggregate consumption, the variability between reports of drinking and alcoholic beverage sales figures can be substantially reduced (Midanik 1982). A recent report describing two Swedish alcohol surveys sheds some light on the differences in drinking reports using two different survey methods (Kuhlhorn and Leifman 1993). The two methods were contrasted to account for the large differences in their retail sales coverage rates (i.e., registered alcoholic beverages sales), namely 75 and 28 percent. Both surveys were conducted in the late 1980's by respected research institutes using large numbers of Swedish respondents. In the survey with a high coverage rate, the interview technique considered the actual daily drinking patterns of the respondents (i.e., could reflect a concentration of drinking on weekends). By dividing a "normal week's" drinking into four periods (Monday-Thursday, Friday, Saturday, and Sunday), the questions allowed one to average periods with varying drinking habits. The survey with a low coverage rate used an undifferentiated QF measure and thus a normal week's drinking could not be divided in this manner. A test of internal validity of the survey with the

higher coverage supported the conclusion that such coverage was due to the refined technique in drinking questions.

How the nature of questionnaire options can affect estimated alcohol consumption is nicely highlighted in a study by Poikolainen and Kärkkäinen (1985). Estimated mean daily alcohol use increased more than twofold using the "heavy" (302 g) instead of the "light" (137 g) questionnaire. Similar results have been reported in other surveys using heavy or atypical drinking questions (e.g., Göransson and Hanson 1994; Polich and Orvis 1979).

In summary, the scientific literature suggests that, although the accuracy of a single individual's report may be difficult to determine, from a group perspective, self-reports of alcohol use from clinical and nonclinical samples are fairly accurate when people are interviewed under the conditions noted above.

Review of Drinking Measures

The remainder of this chapter focuses on the evaluation and selection of drinking measures. This section provides descriptive information for the seven drinking measures listed in table 1, which includes all of the measures that met the selection criteria. A fact sheet for each of these measures is in the appendix. The review of each measure includes its recommended uses, advantages, and limitations. Following these technical reviews is a review of studies that have compared different drinking measures and a discussion of future research directions.

Overview of Measures

Although a number of drinking measures have been used and reported in the literature, only seven satisfied the criteria for inclusion in this volume. These measures are divided into two types: primary and composite. Drinking is the only variable assessed by a primary measure. Composite measures

Examples of

Normed

Administration Target Training Computer Fee population groups used Norms groups Options Time needed; scoring for with

(min) Scoring (min)

use