Treatment evaluation with multiple outcome periods under endogeneity and attrition
Journal
Journal of the American Statistical Association
ISSN
0162-1459
ISSN-Digital
1537-274X
Type
journal article
Date Issued
2014-03
Author(s)
Abstract
This paper develops a nonparametric methodology for treatment evaluation with multiple outcome periods under treatment endogeneity and missing outcomes. We use instrumental variables, pre-treatment characteristics, and short-term (or intermediate) outcomes to identify the average treatment effect on the outcomes of compliers (the subpopulation whose treatment reacts on the instrument) in multiple periods based on inverse probability weighting. Treatment selection and attrition may depend on both observed characteristics and the unobservable compliance type, which is possibly related to unobserved factors. We also provide a simulation study and apply our methods to the evaluation of a policy intervention targeting college achievement, where we find that controlling for attrition considerably affects the effect estimates.
Language
English
Keywords
treatment effect
attrition
endogeneity
panel data
weighting
HSG Classification
contribution to scientific community
HSG Profile Area
SEPS - Quantitative Economic Methods
Refereed
Yes
Publisher
Taylor & Francis
Publisher place
Abingdon UK
Volume
2014
Number
Published online: 07 Mar 2014
Start page
1
End page
34
Pages
34
Subject(s)
Division(s)
Eprints ID
224619
File(s)![Thumbnail Image]()
![Thumbnail Image]()
Name
Frölich Huber 2014.pdf
Size
268.53 KB
Format
Adobe PDF
Checksum (MD5)
f0f751729212096de4f8382103ccd20b
open.access
Name
Supplementary Appendix.pdf
Size
346.52 KB
Format
Adobe PDF
Checksum (MD5)
1a501e86f63a89b5e5c14403843dad6a