Parsimony in Designing a Drug Use Survey: A Methodology Study

Abstract

This paper makes a special study of how one question in an index frequently has as much relationship to students involvement in drug use as the entire attitudinal index developed from many questions. If one or two questions can cover an attitudinal area as well as many, this has significant implications for future surveying. Frequently, researchers want to develop much longer questionnaires than can be easily administered, and therefore have to leave out some crucial areas of interest. It was found that in each of the following attitudinal areas-family, school, friends, taking risks, and opinions about drugs-one item in each area had as much relationship to drug use as the entire indiex built on many items.