Semiparametric efficient G-estimation with invalid instrumental variables

被引:4
作者
Sun, B. [1 ]
Liu, Z. [2 ]
Tchetgen, E. J. Tchetgen [3 ]
机构
[1] Natl Univ Singapore, Dept Stat & Data Sci, 6 Sci Dr 2, Singapore 117546, Singapore
[2] Columbia Univ, Dept Biostat, 722 West 168th St, New York, NY 10032 USA
[3] Univ Penn, Wharton Sch, Dept Stat & Data Sci, 265 South 37th St, Philadelphia, PA 19104 USA
基金
美国国家卫生研究院;
关键词
Causal inference; G-estimation; Instrumental variable; Multiple robustness property; Semiparametric theory; Unmeasured confounding; MENDELIAN RANDOMIZATION; GENERALIZED-METHOD; CONSISTENT ESTIMATION; MULTIPLE ROBUSTNESS; CAUSAL INFERENCE; REGRESSION; IDENTIFICATION; MODELS; SELECTION; MOMENTS;
D O I
10.1093/biomet/asad011
中图分类号
Q [生物科学];
学科分类号
07 ; 0710 ; 09 ;
摘要
The instrumental variable method is widely used in the health and social sciences for identification and estimation of causal effects in the presence of potential unmeasured confounding. To improve efficiency, multiple instruments are routinely used, raising concerns about bias due to possible violation of the instrumental variable assumptions. To address such concerns, we introduce a new class of G-estimators that are guaranteed to remain consistent and asymptotically normal for the causal effect of interest provided that a set of at least ? out of K candidate instruments are valid, for ? = K set by the analyst ex ante without necessarily knowing the identities of the valid and invalid instruments. We provide formal semiparametric efficiency theory supporting our results. Simulation studies and applications to UK Biobank data demonstrate the superior empirical performance of the proposed estimators compared with competing methods.
引用
收藏
页码:953 / 971
页数:20
相关论文
共 83 条
[1]   Semiparametric instrumental variable estimation of treatment response models [J].
Abadie, A .
JOURNAL OF ECONOMETRICS, 2003, 113 (02) :231-263
[2]   Asymptotic Efficiency of Semiparametric Two-step GMM [J].
Ackerberg, Daniel ;
Chen, Xiaohong ;
Hahn, Jinyong ;
Liao, Zhipeng .
REVIEW OF ECONOMIC STUDIES, 2014, 81 (03) :919-943
[3]   ESTIMATION OF THE PARAMETERS OF A SINGLE EQUATION IN A COMPLETE SYSTEM OF STOCHASTIC EQUATIONS [J].
ANDERSON, TW ;
RUBIN, H .
ANNALS OF MATHEMATICAL STATISTICS, 1949, 20 (01) :46-63
[4]  
Angrist JD, 1996, J AM STAT ASSOC, V91, P444, DOI 10.2307/2291629
[5]  
Angrist JD, 1999, J APPL ECONOM, V14, P57, DOI 10.1002/(SICI)1099-1255(199901/02)14:1<57::AID-JAE501>3.0.CO
[6]  
2-G
[7]   2-STAGE LEAST-SQUARES ESTIMATION OF AVERAGE CAUSAL EFFECTS IN MODELS WITH VARIABLE TREATMENT INTENSITY [J].
ANGRIST, JD ;
IMBENS, GW .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1995, 90 (430) :431-442
[8]   SPLIT-SAMPLE INSTRUMENTAL VARIABLES ESTIMATES OF THE RETURN TO SCHOOLING [J].
ANGRIST, JD ;
KRUEGER, AB .
JOURNAL OF BUSINESS & ECONOMIC STATISTICS, 1995, 13 (02) :225-235
[9]  
[Anonymous], SOCIOLOGICAL METHODO, DOI DOI 10.2307/271055
[10]  
[Anonymous], 2019, NAT COMMUN