Making Sense of Sensitivity: Extending Omitted Variable Bias

Journal of the Royal Statistical Society. Series B: Statistical Methodology - Tập 82 Số 1 - Trang 39-67 - 2020
Carlos Cinelli1, Chad Hazlett1
1University of California, Los Angeles, USA

Tóm tắt

SummaryWe extend the omitted variable bias framework with a suite of tools for sensitivity analysis in regression models that does not require assumptions on the functional form of the treatment assignment mechanism nor on the distribution of the unobserved confounders, naturally handles multiple confounders, possibly acting non-linearly, exploits expert knowledge to bound sensitivity parameters and can be easily computed by using only standard regression results. In particular, we introduce two novel sensitivity measures suited for routine reporting. The robustness value describes the minimum strength of association that unobserved confounding would need to have, both with the treatment and with the outcome, to change the research conclusions. The partial R2 of the treatment with the outcome shows how strongly confounders explaining all the residual outcome variation would have to be associated with the treatment to eliminate the estimated effect. Next, we offer graphical tools for elaborating on problematic confounders, examining the sensitivity of point estimates and t-values, as well as ‘extreme scenarios’. Finally, we describe problems with a common ‘benchmarking’ practice and introduce a novel procedure to bound the strength of confounders formally on the basis of a comparison with observed covariates. We apply these methods to a running example that estimates the effect of exposure to violence on attitudes toward peace.

Từ khóa


Tài liệu tham khảo

Altonji, 2005, An evaluation of instrumental variable strategies for estimating the effects of catholic schooling, J. Hum. Resour., 40, 791, 10.3368/jhr.XL.4.791

Angrist, 2008, Mostly Harmless Econometrics: an Empiricist’s Companion, 10.2307/j.ctvcm4j72

Angrist, 2017, Undergraduate econometrics instruction: through our classes, darkly. Technical Report., 10.3386/w23144

Blackwell, 2013, A selection bias approach to sensitivity analysis for causal effects, Polit. Anal., 22, 169, 10.1093/pan/mpt006

Brumback, 2004, Sensitivity analyses for unmeasured confounding assuming a marginal structural model for repeated measures, Statist. Med., 23, 749, 10.1002/sim.1657

Carnegie, 2016, Assessing sensitivity to unmeasured confounding using a simulated potential confounder, J. Res. Educ. Effect., 9, 395

Carnegie, 2016, treatsens: a package to assess sensitivity of causal analyses to unmeasured confounding

Cinelli, 2019, sensemakr: sensitivity analysis tools for OLS

Cinelli, 2019, Sensitivity analysis of linear structural causal models, Proc. Mach. Learn. Res., 97, 1252

Cornfield, 1959, Smoking and lung cancer: recent evidence and a discussion of some questions, J. Natn. Cancer Inst., 22, 173

Ding, 2015, To adjust or not to adjust?: Sensitivity analysis of M-bias and butterfly-bias, J. Causl Inf., 3, 41, 10.1515/jci-2013-0021

Dorie, 2016, A flexible, interpretable framework for assessing sensitivity to unmeasured confounding, Statist. Med., 35, 3453, 10.1002/sim.6973

Dunning, 2012, Natural Experiments in the Social Sciences: a Design-based Approach, 10.1017/CBO9781139084444

Flint, 2008, Darfur: a New History of a Long War, 10.5040/9781350219489

Frank, 2000, Impact of a confounding variable on a regression coefficient, Sociol. Meth. Res., 29, 147, 10.1177/0049124100029002001

Frank, 2013, What would it take to change an inference?: Using Rubin’s causal model to interpret the robustness of causal inferences, Educ. Evaln Poly Anal., 35, 437, 10.3102/0162373713493129

Frank, 2007, Indices of robustness for sample representation, Sociol. Methodol., 37, 349, 10.1111/j.1467-9531.2007.00186.x

Frank, 2008, Does NBPTS certification affect the number of colleagues a teacher helps with instructional matters?, Educ. Evaln Poly Anal., 30, 3, 10.3102/0162373707313781

Franks, 2019, Flexible sensitivity analysis for observational studies without observable implications, J. Am. Statist. Ass.

Frisch, 1933, Partial time regressions as compared with individual trends, Econometrica, 1, 387, 10.2307/1907330

Hazlett, 2019, Angry or weary?: The effect of personal violence on attitudes towards peace in Darfur, J. Conflct Resoln

Hong, 2018, Weighting-based sensitivity analysis in causal mediation studies, J. Educ. Behav. Statist., 43, 32, 10.3102/1076998617749561

Hosman, 2010, The sensitivity of linear regression coefficients’ confidence limits to the omission of a confounder, Ann. Appl. Statist., 4, 849, 10.1214/09-AOAS315

Imai, 2010, Identification, inference and sensitivity analysis for causal mediation effects, Statist. Sci., 25, 51, 10.1214/10-STS321

Imbens, 2003, Sensitivity to exogeneity assumptions in program evaluation, Am. Econ. Rev., 93, 126, 10.1257/000282803321946921

Imbens, 2015, Causal Inference in Statistics, Social, and Biomedical Sciences, 10.1017/CBO9781139025751

Jakiela, 2018, Gendered language. Policy Research Working Paper.

Kruskal, 1989, Concepts of relative importance in recent scientific literature, Am. Statistn, 43, 2, 10.1080/00031305.1989.10475596

Leamer, 1983, Let’s take the con out of econometrics, Am. Econ. Rev., 73, 31

Leamer, 2016, S-values: conventional context-minimal measures of the sturdiness of regression coefficients, J. Econmetr., 193, 147, 10.1016/j.jeconom.2015.10.013

Lovell, 1963, Seasonal adjustment of economic time series and multiple regression analysis, J. Am. Statist. Ass., 58, 993, 10.1080/01621459.1963.10480682

Lovell, 2008, A simple proof of the FWL theorem, J. Econ. Educ., 39, 88, 10.3200/JECE.39.1.88-91

Middleton, 2016, Bias amplification and bias unmasking, Polit. Anal., 24, 307, 10.1093/pan/mpw015

Oster, 2014, Unobservable selection and coefficient stability: theory and evidence. Working Paper.

Oster, 2019, Unobservable selection and coefficient stability: theory and evidence, J. Bus. Econ. Statist., 37, 187, 10.1080/07350015.2016.1227711

Pearl, 2009, Causality, 10.1017/CBO9780511803161

Pearl, 2011, Invited commentary: understanding bias amplification, Am. J. Epidem., 174, 1223, 10.1093/aje/kwr352

Robins, 1999, Association, causation, and marginal structural models, Synthese, 121, 151, 10.1023/A:1005285815569

Rosenbaum, 1984, The consequences of adjustment for a concomitant variable that has been affected by the treatment, J. R. Statist. Soc., 147, 656

Rosenbaum, 2002, Observational Studies, 1, 10.1007/978-1-4757-3692-2

Rosenbaum, 2005, Encyclopedia of Statistics in Behavioral Science, 1809

Rosenbaum, 2010, Design of Observational Studies, 10.1007/978-1-4419-1213-8

Rosenbaum, 2017, Observation and Experiment: an Introduction to Causal Inference, 10.4159/9780674982697

Rosenbaum, 1983, The central role of the propensity score in observational studies for causal effects, Biometrika, 70, 41, 10.1093/biomet/70.1.41

Rosenbaum, 1983, Assessing sensitivity to an unobserved binary covariate in an observational study with binary outcome, J. R. Statist. Soc., 45, 212, 10.1111/j.2517-6161.1983.tb01242.x

Steiner, 2016, The mechanics of omitted variable bias: bias amplification and cancellation of offsetting biases, J. Causl Inf., 4

Vanderweele, 2011, Bias formulas for sensitivity analysis of unmeasured confounding for general outcomes, treatments, and confounders, Epidemiology, 22, 42, 10.1097/EDE.0b013e3181f74493

VanderWeele, 2017, Sensitivity analysis in observational research: introducing the E-value, Ann. Intern. Med., 167, 268, 10.7326/M16-2607

Thông tin
Thông tin xuất bản

Journal of the Royal Statistical Society. Series B: Statistical Methodology

Tập 82 Số 1

39-67

Thông tin tác giả