The case for uniform controls in drug policy studies

Abstract

One of the greatest threats to the validity and reliability of drug policy research studies is the selective inclusion of controls. Policies are not formulated in a vacuum, and the inclusion of a rich, comprehensive array of demographic controls reduces the possibility that a correlation in the data is better explained by endogenous factors than the studied independent variable. Whether intentional or not, exclusion (or inclusion) of controls can make spurious relationships appear significant (or vice versa). This issue is important in studies of drug policy as demographics contribute significantly to usage rates and attitudes (Kandel, Chen, Warner, Kessler, & Grant, 1997).