Are Telephone Surveys Suitable for Studying Substance Abuse? Cost, Administration, Coverage and Response Rate Issues

Article excerpt

Substance-abuse telephone surveys have become increasingly popular; however, the methodological literature has focused solely on potential disadvantages of the method when compared to face-to-face surveys. Striving for greater balance, this paper examines the potential advantages (lower cost, greater security and privacy, and easier administration) and disadvantages (noncoverage, lower response rates, and lower validity). Existing evidence suggests that the disadvantages typically have a small impact on prevalence estimates. When there is a risk of substantial bias, researchers should employ available techniques for minimizing it. On balance, the telephone survey 's advantages appear to outweigh the disadvantages for many research problems, but enhanced designs are needed to realize the method's full potential in this substantive area.


Telephone surveys are an increasingly popular method of estimating alcohol and drug use, abuse, and dependence in the general population (McAuliffe, Paletz, et al. 1996). At present, all states are regularly collecting alcohol-use data by telephone as part of the Centers for Disease Control (CDC) and Prevention's Behavioral Risk Factor Surveillance System (BRFSS) (Gentry et al. 1985; Remington et al. 1988; Bradstock et al. 1985; Gentry and CDC 1989). States are also collecting alcohol and drug use and use-disorder data as part of the State Needs Assessment Program sponsored by the Center for Substance Abuse Treatment (CSAT) (Becker et al. 1993). Canadians have used telephone administration in 1989 and 1994 to conduct their national substance-abuse survey (Single and Wortley 1994; Sobell et al. 1996; Statistics Canada 1994), and many individual scientists have conducted national, state, and local telephone surveys (see McAuliffe, Paletz, et al. 1996 for a review). We estimate that approximately 700 telephone surveys that collected substance use and abuse data have been completed since 1967 (McAuliffe, Paletz, et al. 1996). Thus, it appears that telephone surveys have quietly become a commonly used form of substance-abuse interview survey.

Surprisingly, the most widely cited literature on related methodological issues (e.g., Johnson et al. 1989; Gfroerer and Hughes 1991; Aquilino 1992, 1994) has focused mainly on the disadvantages and limitations of telephone surveys for collecting drug-use data rather than on the advantages that have led to the increasingly widespread adoption of the method for substance-abuse studies. Citing several deficiencies of telephone surveys, these authors concluded that telephone interview surveys produced drug-use data that were less valid than data produced in face-toface interview surveys, especially the National Household Survey of Drug Abuse (NHSDA). Two leading government officials concerned with drug abuse epidemiology, Gfroerer and Hughes ( 1991:391), concluded that, "...biases seem to be present in telephone surveys...which suggests that they may not be appropriate for collecting drug use data."

As a result, the growing use of telephone surveys for collecting drug-use data has been questioned. For example, in a comprehensive assessment of drug-abuse statistics for the federal government, the Rand Corporation expressed doubts about the CSAT decision to use telephone surveys in its needs assessment program (Ebener et al. 1994).

The most common criticism of telephone surveys of substance abuse is that their results are biased because they miss people without telephones, fewer people are willing to participate in telephone surveys than in face-to-face surveys, and people who participate in telephone surveys are more likely to underreport drug use than are people who participate in face-to-face surveys (e.g., LoSciuto et al. 1993). However, no publication that we could find has comprehensively examined the advantages and disadvantages of telephone surveys when compared to face-to-face substance-abuse surveys. …