The Role of Publication Selection Bias in Estimates of the Value of a Statistical Life

Abstract

Meta-regression estimates of the value of a statistical life (VSL) controlling for publication selection bias often yield bias-corrected estimates of VSL that are substantially below the mean VSL estimates. Labor market studies using the more recent Census of Fatal Occupational Injuries (CFOI) data are subject to less measurement error and also yield higher bias-corrected estimates than do studies based on earlier fatality rate measures. These results are borne out by the findings for a large sample of all VSL estimates based on labor market studies using CFOI data and for four meta-analysis data sets consisting of the authors’ best estimates of VSL. The confidence intervals of the publication bias-corrected estimates of VSL based on the CFOI data include the values that are currently used by government agencies, which are in line with the most precisely estimated values in the literature.