The use of low-level theory to guide the interpretation of road safety evaluation studies

Abstract

There are two main interpretations of empirical research: methodological and substantive. A methodological interpretation usually rejects a study by arguing that it is based on poor data or methods. A substantive interpretation accepts results as showing real effects. This paper argues that by developing and testing hypotheses about systematic variation in the effects of road safety measures, it may be possible to defend a substantive interpretation of the results of studies that might otherwise be rejected on methodological grounds. Studies evaluating the road safety effects of road lighting are used to illustrate the approach. Ten hypotheses are proposed and tested by means of two meta-analyses. Most of the hypotheses are supported. Thus, although many studies evaluating the road safety effects of road lighting control poorly for potential confounding factors, the systematic pattern of results found in these studies indicates that they mainly show the effects of road lighting, not of confounding factors not controlled for.