Abstract
Causal inference problems are often tackled through the study of parameters of the distribution of a sequence of counterfactual variables, that represent the outcome in a hypothetical world where an intervention is enforced. The fundamental problem of causal inference is that, for a given individual, we only observe one such counterfactual outcome: the outcome under the treatment level actually assigned. Therefore, it is necessary to make certain untestable assumptions to identify the distribution of the missing counterfactual outcomes from the distribution of the observed data. One common such assumption is that the treatment mechanism does not depend on unmeasured factors that are causally related to the outcome. This assumption is often referred to as nonignorability of treatment assignment (Rubin 1976) or the (sequential) randomization assumption (van der Laan and Robins 2003).
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
T.J. Aragon, epitools: Epidemiology tools (2012). http://cran.r-project.org/package=epitools
I. Díaz, M.J. van der Laan, Sensitivity analysis for causal inference under unmeasured confounding and measurement error problems. Int. J. Biostat. 9(2), 149–160 (2013b)
P. Ding, T. VanderWeele, Sensitivity analysis without assumptions. Epidemiol. 27(3), 368–377 (2016)
J.L. Horowitz, C.F. Manski, Nonparametric analysis of randomized experiments with missing covariate and outcome data. J. Am. Stat. Assoc. 95(449), 77–84 (2000)
A.R. Luedtke, I. Díaz, M.J. van der Laan, The statistics of sensitivity analyses. Technical Report, Division of Biostatistics, University of California, Berkeley (2015b)
R.F. MacLehose, S. Kaufman, J.S. Kaufman, C. Poole, Bounding causal effects under uncontrolled confounding using counterfactuals. Epidemiology 16(4), 548–555 (2005)
C.F. Manski, Partial Identification of Probability Distributions (Springer, Berlin, Heidelberg, New York, 2003)
C.F. Manski, Nonparametric bounds on treatment effects. Am. Econ. Rev. 80, 319–323 (1990)
J.M. Robins, Association, causation and marginal structural models. Synthese 121, 151–179 (1999)
J.M. Robins, A. Rotnitzky, D.O. Scharfstein, Sensitivity analysis for selection bias and unmeasured confounding in missing data and causal inference models, in Statistical Models in Epidemiology, the Environment and Clinical Trials. IMA Volumes in Mathematics and Its Applications (Springer, Berlin, 1999)
P.R. Rosenbaum, D.B. Rubin, Assessing sensitivity to an unobserved binary covariate in an observational study with binary outcome. J. R. Stat. Soc. Ser. B 45, 212–218 (1983a)
R.H. Rosenman, M. Friedman, R. Straus, M. Wurm, R. Kositchek, W. Hahn, N.T. Werthessen, A predictive study of coronary heart disease: the western collaborative group study. J. Am. Med. Assoc. 189(1), 15–22 (1964)
R.H. Rosenman, R.J. Brand, C.D. Jenkins, M. Friedman, R. Straus, M. Wurm, Coronary heart disease in the western collaborative group study: final follow-up experience of 8 1/2 years. J. Am. Med. Assoc. 233(8), 872–877 (1975)
A. Rotnitzky, D. Scharfstein, S. Ting-Li Su, J. Robins, Methods for conducting sensitivity analysis of trials with potentially nonignorable competing causes of censoring. Biometrics 57(1), 103–113 (2001)
A. Rotnitzky, J.M. Robins, D.O. Scharfstein, Semiparametric regression for repeated outcomes with nonignorable nonresponse. J. Am. Med. Assoc. 93(444), 1321–1339 (1998)
D.B. Rubin, Multivariate matching methods that are equal percent bias reducing, II: maximums on bias reduction for fixed sample sizes. Biometrics 32(1), 121–132 (1976)
D.O. Scharfstein, J.M. Robins, Estimation of the failure time distribution in the presence of informative censoring. Biometrika 89(3), 617–634 (2002)
D.O. Scharfstein, A. Rotnitzky, J.M. Robins, Adjusting for nonignorable drop-out using semiparametric nonresponse models, (with discussion and rejoinder). J. Am. Stat. Assoc. 94, 1096–1120, 1121–1146 (1999)
R.J.C.M. Starmans, Models, inference, and truth: probabilistic reasoning in the information era, in Targeted Learning: Causal Inference for Observational and Experimental Data, ed. by M. van der Laan, S. Rose (Springer, Berlin, 2011)
M.J. van der Laan, Targeted estimation of nuisance parameters to obtain valid statistical inference. Int. J. Biostat. 10(1), 29–57 (2014b)
M.J. van der Laan, J.M. Robins, Unified Methods for Censored Longitudinal Data and Causality (Springer, Berlin Heidelberg New York, 2003)
M.J. van der Laan, S. Rose, Targeted Learning: Causal Inference for Observational and Experimental Data (Springer, Berlin, Heidelberg, New York, 2011)
T.J. VanderWeele, Bias formulas for sensitivity analysis for direct and indirect effects. Epidemiology 21(4), 540 (2010)
T.J. VanderWeele, O.A. Arah, Unmeasured confounding for general outcomes, treatments, and confounders: bias formulas for sensitivity analysis. Epidemiology 22(1), 42 (2011)
T.J. VanderWeele, B. Mukherjee, J. Chen, Sensitivity analysis for interactions under unmeasured confounding. Stat. Med. 31(22), 2552–2564 (2012a)
S. Vansteelandt, E. Goetghebeur, M.G. Kenward, G. Molenberghs, Ignorance and uncertainty regions as inferential tools in a sensitivity analysis. Stat. Sin. 16(3), 953–979 (2006)
T Woutersen, A simple way to calculate confidence intervals for partially identified parameters. Technical Report, Johns Hopkins University (2006)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG
About this chapter
Cite this chapter
Díaz, I., Luedtke, A.R., van der Laan, M.J. (2018). Sensitivity Analysis. In: Targeted Learning in Data Science. Springer Series in Statistics. Springer, Cham. https://doi.org/10.1007/978-3-319-65304-4_27
Download citation
DOI: https://doi.org/10.1007/978-3-319-65304-4_27
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-65303-7
Online ISBN: 978-3-319-65304-4
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)