H. Laurence Ross

METHODS FOR STUDYING THE IMPACT OF DRUNK DRIVING LAWS

This paper discusses methodological principles for the study of legal interventions based on the theory of deterrence, especially in the area of drunk driving. It reviews the validity of commonly available measures of drunk driving and identifies strengths and weaknesses of correlational, quasi-experimental, and experimental study designs for demonstrating causal connections between drunk driving and legal interventions. The time series is the most powerful of the quasi-experimental designs. Because it uses high-quality data that are generally available cheaply from official sources, and because it can usually overcome the common threats to inferential validity that plague many other designs, it is recommended as a model for studies of deterrence of drunk driving.

Reflections on Doing Policy-Relevant Sociology: How to Cope with MADD Mothers.

The author’s experiences in studying the ability of law to deter drunk driving lead him to conclude that research is selectively used by various politically-involved groups. These groups frequently are committed to specific definitions of a policy problem and to specific approaches to its resolution. The researcher finds that his product is not used for the purpose of selecting among alternative policy options, but instead it serves to supply rationalizations for positions to which the involved groups are previously committed.