Minimum Legal Drinking Ages and Highway Safety: A Methodological Critique (From Preventing Automobile Injury: New Findings From Evaluation Research, P 116-162, 1988, John D Graham, ed. -- See NCJ-118577)

• NCJ Number: 118580
• Author(s): S Garber
• Date Published: 1988
• Length: 47 pages
• Sale Source: Auburn House
• Publication Type: Legislation/Policy Analysis
• Language: English
• Country: United States of America

Annotation

This paper describes and critiques four general approaches to holding other factors constant in measuring the impact of minimum legal drinking age (MLDA) on highway accidents.

Abstract

The paper first discusses the decision problem of an idealized policymaker to facilitate the exposition of various general issues pertinent to enhancing the usefulness of future studies on the impact of the MLDA on traffic accidents. The discussion notes that controlling for other factors is fundamental in such studies. The methods described and critiqued are statistical analysis explicitly controlling for other theoretical causes, statistical analysis with "dummy-variable controls," simple paired comparisons, and paired comparisons using statistical projection. The critique of each approach focuses on the development of the conditions under which each method would succeed in isolating the policy-relevant component of an observed highway safety outcome. In concluding that each of the approaches is flawed, the paper suggests that future research consist of sensitivity analysis across methods. Two commentaries accompany the paper. One comment disagrees with the paper's view that statistical control methods rather than quasi-experimental methods are best suited to theory development. The second comment addresses the issues of data accuracy, economic factors, and policymaking. 42 references, 5 notes.