Contrast effects in perceived risk of substance use

Adolescent perceptions of the risks associated with the use of licit and illicit substances have important implications for policy and research. However, the methodological properties of the most popular risk measures in school surveys in Europe and the United States are not well understood. This study examines the potential contrast effects of risk measures of "heavy" and frequent substance use on perceived risks of occasional, moderate, and "experimental" use. Responses to 11 measures of the perceived risk of occasional smoking, moderate drinking, and experimental use of illicit substances were compared between two question forms administered to a split-half sample of all Icelandic ninth (14-15 years of age) and tenth (15-16 years of age) grade students present in class on the day of administration in March 2003 (N = 7099). In one form, only these 11 questions were used, while the other form also contained 13 questions on the perceived risk of heavy smoking, heavy drinking, and regular use of illicit substances. The longer form is found to decrease response rates and suppress estimates of perceived risk of experimental illicit substance use. Question form and perceived risks of heavy and regular use generally do not affect the multivariate effects of perceived risks of occasional, moderate, and experimental substance use on lifetime abstinence from each substance. It is argued that measures of perceived risks of heavy and regular substance use are less useful for prevention policy and research than are corresponding measures of occasional, moderate, and experimental substance use, and including the former in the same instrument may adversely affect the measurement of the latter constructs.