Graphical Models for Inference Under Outcome-Dependent Sampling

被引:50
作者
Didelez, Vanessa [1 ]
Kreiner, Svend [2 ]
Keiding, Niels [2 ]
机构
[1] Univ Bristol, Dept Math, Bristol BS8 1TW, Avon, England
[2] Univ Copenhagen, Dept Biostat, DK-1014 Copenhagen, Denmark
关键词
Causal inference; collapsibility; odds ratios; selection bias; CONTINGENCY-TABLES; CONDITIONAL-INDEPENDENCE; CAUSAL INFERENCE; COLLAPSIBILITY; DIAGRAMS; ASSOCIATION; RISK; BIAS;
D O I
10.1214/10-STS340
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
We consider situations where data have been collected such that the sampling depends on the outcome of interest and possibly further covariates, as for instance in case-control studies. Graphical models represent assumptions about the conditional independencies among the variables. By including a node for the sampling indicator, assumptions about sampling processes can be made explicit. We demonstrate how to read off such graphs whether consistent estimation of the association between exposure and outcome is possible. Moreover, we give sufficient graphical conditions for testing and estimating the causal effect of exposure on outcome. The practical use is illustrated with a number of examples.
引用
收藏
页码:368 / 387
页数:20
相关论文
共 70 条
[1]  
ALTHAM PME, 1970, J ROY STAT SOC B, V32, P63
[2]  
Angrist JD, 1996, J AM STAT ASSOC, V91, P444, DOI 10.2307/2291629
[3]  
[Anonymous], EPIDEMIOLOGIC PERSPE
[4]  
[Anonymous], INT BIOM C FREIB
[5]  
[Anonymous], INT J BIOSTAT
[6]  
[Anonymous], INT STAT REV
[7]  
[Anonymous], STAT SURVEY IN PRESS
[8]  
[Anonymous], P 4 C UNC ART INT
[9]  
[Anonymous], 2010, JMLR P
[10]  
[Anonymous], EPIDEMIOLOGY