A comparison of results from an alcohol survey of a prerecruited Internet panel and the national epidemiologic survey on alcohol and related conditions

Background: Given today's telecommunications environment, random digit dial (RDD) telephone surveys face declining response rates and coverage, and increasing costs. As an alternative to RDD, we surveyed participants in a randomly recruited standing Internet panel supplemented with a randomly sampled telephone survey of nonpanel members for a study of associations between onset of alcohol use and later alcohol-related problems. The purpose of this paper was to compare results from our survey with results from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), a face-to-face probability sample survey of 43,093 adults, with a focus on associations between demographics, age of drinking onset, and alcohol dependence. Methods: Demographic and drinking characteristics from our survey of 4,021 ever-drinkers between the ages of 18 and 39 years were compared with the characteristics of 11,549 similarly aged ever-drinkers from the NESARC. Weighted analyses accounting for sampling design compared these 2 samples on drinking characteristics over the past year and during a respondent's heaviest period of drinking, and in multivariate models examining associations between demographics, age of drinking onset, and lifetime alcohol dependence. Results: Participants in the supplemented Internet panel were similar to the national population of 18- to 39-year-old ever drinkers on gender, education, and race/ethnicity, while adults who were aged 18 to 25 years were under-represented in the Internet panel. The supplemented Internet panel reported higher rates of moderate risk drinking over the past 12 months, lifetime high-risk drinking, and lifetime (ever) alcohol dependence. Estimates of the associations between alcohol dependence and age of drinking onset, risky drinking, and family history of alcohol problems did not significantly differ between the supplemented Internet sample and the NESARC survey. Conclusions: Randomly recruited Internet-based panels may provide an alternative to random digit dial telephone surveys and in-person surveys for some studies of factors associated with alcohol-related problems.