Regularization and Confounding in Linear Regression for Treatment Effect Estimation

被引：46

作者：

Hahn, P. Richard ^{[1
]}

Carvalho, Carlos M. ^{[2
]}

Puelz, David ^{[2
]}

He, Jingyu ^{[1
]}

机构：

[1] Univ Chicago, Booth Sch Business, 5807 South Woodlawn Ave, Chicago, IL 60637 USA

[2] Univ Texas Austin, McCombs Sch Business, 2110 Speedway Stop, Austin, TX 78712 USA

来源：

BAYESIAN ANALYSIS | 2018年 / 13卷 / 01期

关键词：

causal inference; observational data; shrinkage estimation; PROPENSITY SCORE; VARIABLE SELECTION; UNCERTAINTY; OUTCOMES;

D O I：

10.1214/16-BA1044

中图分类号：

O1 [数学];

学科分类号：

0701 ; 070101 ;

摘要：

This paper investigates the use of regularization priors in the context of treatment effect estimation using observational data where the number of control variables is large relative to the number of observations. First, the phenomenon of "regularization-induced confounding" is introduced, which refers to the tendency of regularization priors to adversely bias treatment effect estimates by over-shrinking control variable regression coefficients. Then, a simultaneous regression model is presented which permits regularization priors to be specified in a way that avoids this unintentional "re-confounding". The new model is illustrated on synthetic and empirical data.

引用

页码：163 / 182

页数：20

共 27 条

[11] Bayesian treatment effects models with variable selection for panel outcomes with an application to earnings effects of maternity leave [J].

Jacobi, Liana ;

Wagner, Helga ;

Fruehwirth-Schnatter, Sylvia .

JOURNAL OF ECONOMETRICS, 2016, 193 (01) :234-250

[12]

Leamer E. E., 1978, SPECIFICATION SEARCH, V53

[13]

LEAMER EE, 1983, AM ECON REV, V73, P31

[14] The effect of the prior distribution in the Bayesian Adjustment for Confounding algorithm [J].

Lefebvre, Genevieve ;

Atherton, Juli ;

Talbot, Denis .

COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2014, 70 :227-240

[15]

Levitt StevenD., 2010, Freakonomics, V61

[16]

Li ML, 2014, BAYESIAN INFERENCE IN THE SOCIAL SCIENCES, P63

[17] Mixtures of g priors for Bayesian variable selection [J].

Liang, Feng ;

Paulo, Rui ;

Molina, German ;

Clyde, Merlise A. ;

Berger, Jim O. .

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 2008, 103 (481) :410-423

[18] Bayesian propensity score analysis for observational data [J].

McCandless, Lawrence C. ;

Gustafson, Paul ;

Austin, Peter C. .

STATISTICS IN MEDICINE, 2009, 28 (01) :94-112

[19]

Morgan SL, 2015, ANAL METHOD SOC RES, P1

[20] ESTIMATING EXPOSURE EFFECTS BY MODELING THE EXPECTATION OF EXPOSURE CONDITIONAL ON CONFOUNDERS [J].

ROBINS, JM ;

MARK, SD ;

NEWEY, WK .

BIOMETRICS, 1992, 48 (02) :479-495

← 1 2 3 →