Publication Bias in Road Safety Evaluation: How can It be Detected and how Common is It?

Abstract

Publication bias denotes the tendency not to publish a study if its findings are not statistically significant or are regarded as unwanted or difficult to explain. In road safety evaluation research, studies that find an increase in the number of accidents as a result of a safety treatment may be less likely to be published than studies that find an accident reduction. How publication bias can be detected and how its influence on summary estimates of effect in meta-analyses can be determined are demonstrated by applying the trim-and-fill technique. Examples are based on recent analyses that indicate the presence of publication bias in studies that evaluate several different road safety measures. Although there is no way to determine conclusively how widespread publication bias is, indications of bias have been found for a number of well-known and widely applied road safety measures. Factors that may influence the likelihood of publication bias are discussed.