Surveys conducted since 1969 have focused on the prevalence of juvenile drug use. However, differences in sampling approaches and data collection methods have made comparisons of survey results problematic. Recent efforts have focused on correcting these difficulties by standardizing methodological approaches. The Peilstations-Survey, a national assessment of students ages 12-18 years, has been conducted 4 times since 1984. Results indicate an increase in both the lifetime and the last-month prevalence of marijuana use over the 1984-1996 period, although only a small percentage of the students reported using marijuana on a regular basis. The general characteristics of very regular marijuana users have changed somewhat over time and have become more like those of current nonusers of marijuana. Peer influence in particular has remained an important predictor of marijuana-using behavior. These strikingly different conclusions indicate that the actual impact of coffee shops on marijuana use rates in the Netherlands is unknown. Tables and 39 references (Author abstract modified)