The validity of a province-wide student drug use survey: lessons in design

This study assesses the validity and reliability of a province-wide survey on drug use by high school students in Nova Scotia in 1991. The Nova Scotia instrument was derived from the prototype provided in the Canadian guidelines for self-reported adolescent drug use surveys. To our knowledge, this is the first study on the validity and reliability of the Canadian instrument, as applied in a province, since 1978. Two new methods are proposed as validity and reliability checks: a statistical method to assess possible exaggeration by the group claiming to have used a fictitious drug, and a method to review the entire instrument based on responding error distributions. Overall, this study provides strong evidence of the validity and reliability of the items specifically about drug use. The major source of error uncovered using the responding error distributions, was related to instrument design and occurred in items considered to be explanatory variables for drug use.