NCJ Number
167873
Journal
Addiction Volume: 91 Issue: 4 Dated: (1996) Pages: 557-564
Date Published
1996
Length
8 pages
Annotation
Data from a regional general population survey were used to examine the selection bias introduced into national alcohol surveys as a result of the use of screening to exclude certain cases.
Abstract
The study was prompted by recognition that national alcohol surveys that use personal interviews often, for reasons of efficiently, use screener items to identify individuals who are likely to have experienced alcohol-related problems. Only these individuals are chosen to respond to a list of questions on alcohol-related problems. The present research included the screening questions used in the National Health Interview Survey (NHIS) and the National Alcohol Survey (NAS). However, it did not screen on those items. The analysis used three subsamples: (1) all current drinkers (without screening), (2) those who passed the screener items for the NHIS, and (3) those who passed the screener items for the NAS. Data were analyzed by gender by means of risk curves and estimated probabilities from logistic regressions of three items on alcohol dependence symptoms and two indexes of problems. Results indicated a minimal effect of the screener items on the estimated prevalence of the measures. Thus, findings supported the judicious use of screener items when greater survey efficiency is required. Figures and 12 references (Author abstract modified)