Selection bias in a controlled experiment: The case of moving to opportunity /

Saved in:
Bibliographic Details
Author / Creator:Pinto, Rodrigo, author.
Imprint:2015.
Ann Arbor : ProQuest Dissertations & Theses, 2015
Description:1 electronic resource (92 pages)
Language:English
Format: E-Resource Dissertations
Local Note:School code: 0330
URL for this record:http://pi.lib.uchicago.edu/1001/cat/bib/10773156
Hidden Bibliographic Details
Other authors / contributors:University of Chicago. degree granting institution.
ISBN:9781321898927
Notes:Advisors: James J. Heckman; Steven N. Durlauf Committee members: Dan A. Black.
This item must not be sold to any third party vendors.
This item must not be added to any third party search indexes.
Dissertation Abstracts International, Volume: 76-11(E), Section: A.
English
Summary:The Moving to Opportunity (MTO) is a social experiment designed to evaluate the effects of neighborhoods on the economic and social outcomes of disadvantaged families in the United States. It targeted over 4,000 families living in high poverty housing projects during the years of 1994-1997 across five U.S. cities. MTO randomly assigned voucher subsidies that incentivize families to relocate from high poverty housing projects to better neighborhoods. Nearly half of the families assigned to vouchers moved. Although the MTO randomization is well suited to evaluate the causal effects of offering vouchers to families, it is less clear how randomized vouchers might be used to assess the causal effects of neighborhoods on outcomes. I exploit the experimental design of the MTO to nonparametrically identify the causal effects of neighborhood relocation on socioeconomic outcomes. My identification strategy employs revealed preference analysis and combines it with tools of causal inference developed in the literature on Causal Bayesian Networks. I find that neighborhood relocation has statistically significant causal effects on labor market outcomes. I decompose the widely reported treatment-on-the-treated parameter - the voucher's effect divided by the compliance rate for the voucher - into components that are unambiguously interpreted in terms of neighborhood effects. The method that I develop is general and applies broadly to unordered choice models with categorical instrumental variables and multiple treatments.