Experimentally estimating the value of a statistical life

How to cite

Killeen, Grady. 2022. "Experimentally estimating the value of a statistical life." AEA RCT Registry. AEARCTR-0010101. https://doi.org/10.1257/rct.10101-3.2

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Grady Killeen*

* Primary investigator

Registry ID
AEARCTR-0010101
DOI
https://doi.org/10.1257/rct.10101-3.2
First registered
Last updated
Status
Completed
Trial dates
2022-10-17 – 2023-05-15
Countries
Kenya (Nairobi)
JEL codes
O18, R49, J17
Registry record
http://www.socialscienceregistry.org/trials/10101

Abstract

This document proposes a randomized controlled trial (RCT) in Nairobi, Kenya to estimate the value of a statistical life (VSL). Previous work has relied on observational evidence from settings such as transportation or employment decisions to bound individuals' price elasticity of demand for reductions in mortality risk. These approaches assume that individuals have unbiased information about mortality risk, and then use objectively measured risks to estimate demand models. This study first aims to investigate the possibility that individuals have systematically biased beliefs about risk. Second, we aim to generate experimental variation to estimate VSL. We plan to implement this experiment in the context of demand for motorcycle helmets.

Experimental design

I plan to conduct a lab-in-the-field experiment that will consist of one survey lasting 15-45 minutes. Respondents will be given randomized information about motorcycle safety, and then all treatment groups will participate in a willingness to accept exercise to measure valuations for motorcycle helmets.

Design details

Randomization method
Randomization will be conducted in SurveyCTO using the random() function, which uses the Java randomization algorithm to take a pseudo-random draw from a standard uniform distribution. This study does not stratify randomization since the sample is not known en-ante.
Randomization unit
Individual
Clusters
We estimate 1,000-2,000 surveys.
Observations
We estimate 1,000-2,000 surveys.
Arms
The chance of being assigned to pure control is 10%, control 30%, information treatment 1 30% and information treatment 2 30%.
Primary outcomes
The primary outcome is the value of a statistical life (VSL).
Secondary outcomes
Semi-elasticity of demand for helmet with respect to information, price elasticity of demand for helmets, and the reduced form effect of each treatment assignment on helmet demand.

Cited by

No citation data available for this registration yet.

Keywords