Abstract
Selection of the best estimates of economic parameters frequently relies on the “best estimates” or a meta-analysis of the “best set” of parameter estimates from the literature. Using an all-set dataset consisting of all reported estimates of the value of a statistical life (VSL) as well as a best-set sample of the best estimates from these studies, this article estimates statistically significant publication selection biases in each case. Biases are much greater for the best-set sample, as one might expect, given the subjective nature of the best-set selection process. For the all-set sample, the mean bias-corrected estimate of the VSL for the preferred specification is $8.1 million for the whole sample and $11.4 million based on the CFOI data, while for the best-set results, the whole sample value is $3.5 million, and the CFOI data estimate is $4.4 million. Previous estimates of huge publication selection biases in the VSL estimates are attributable to these studies’ reliance on best-set samples.
Highlights
There may be fundamental policy ramifications if there is an impact of such biases on estimates of parameters that are influential components of policy assessment, such as the value of a statistical life (VSL), which is the focus of this article
The results reported in Viscusi (2015) updated their database and found bias-corrected VSL estimates on the order of $1–$2 million for best-set samples ranging from 39 to 60 studies
Despite the evidence that standard statistical procedures for adjusting for publication selection biases document the presence of substantial selection effects, the mean and median estimates of both the best-set estimates and the all-set estimates are in a reasonable range, but the distributions differ
Summary
Researchers selected the single best estimate from particular articles, as well as an “all-set” dataset that includes all VSL estimates reported in those studies.6 Each of these sets of results is subject to the same sources of biases, but to a different degree. The results reported in Viscusi (2015) updated their database and found bias-corrected VSL estimates on the order of $1–$2 million (in 2013 dollars) for best-set samples ranging from 39 to 60 studies. Despite the evidence that standard statistical procedures for adjusting for publication selection biases document the presence of substantial selection effects, the mean and median estimates of both the best-set estimates and the all-set estimates are in a reasonable range, but the distributions differ. The findings presented here demonstrate that best estimate selection bias may induce serious additional selection effects in addition to those that result from publication selection effects
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
