CoChat

Causal mediation analysis with double machine learning

Farbmacher, Helmut, Huber, Martin, Lafférs, Lukáš, Langen, Henrika, Spindler, Martin · 2022 · The Econometrics Journal
Cited by 198

The DOI matches its Crossref record; 13 of the 13 references it cites were found in a registry or an index.

Abstract

Summary: This paper combines causal mediation analysis with double machine learning for a data-driven control of observed confounders in a high-dimensional setting. The average indirect effect of a binary treatment and the unmediated direct effect are estimated based on efficient score functions, which are robust with respect to misspecifications of the outcome, mediator, and treatment models. This property is key for selecting these models by double machine learning, which is combined with data splitting to prevent overfitting. We demonstrate that the effect estimators are asymptotically normal and $n^{-1/2}$-consistent under specific regularity conditions and investigate the finite sample properties of the suggested methods in a simulation study when considering lasso as machine learner. We also provide an empirical application to the US National Longitudinal Survey of Youth, assessing the indirect effect of health insurance coverage on general health operating via routine checkups as mediator, as well as the direct effect.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1093/ectj/utac003

Cites

13 of 13 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 13 references

Research works printed but not linked: 51

  1. Albert, 2008, Mediation analysis via potential outcomes models, Statistics in Medicine, 27, 1282, 10.1002/sim.3016 · DOI · Paper page
  2. Albert, 2011, Generalized causal mediation analysis, Biometrics, 67, 1028, 10.1111/j.1541-0420.2010.01547.x · DOI · Paper page
  3. Bellani, 2018, The long-run effect of childhood poverty and the mediating role of education, Journal of the Royal Statistical Society: Series A (Statistics in Society), 182, 37, 10.1111/rssa.12388 · DOI
  4. Belloni, 2017, Program evaluation and causal inference with high-dimensional data, Econometrica, 85, 233, 10.3982/ECTA12723 · DOI · Paper page
  5. Bijwaard, 2018, An IPW estimator for mediation effects in hazard models: With an application to schooling, cognitive ability, and mortality, Empirical Economics, 57, 1
  6. Burstin, 1998, The effect of change of health insurance on access to care, Inquiry, 35, 389
  7. Cardella, 2014, The effect of health insurance coverage on the reported health of young adults, Economics Letters, 124, 406, 10.1016/j.econlet.2014.06.036 · DOI
  8. Chernozhukov, 2018, Double/debiased machine learning for treatment and structural parameters, Econometrics Journal, 21, C1, 10.1111/ectj.12097 · DOI · Paper page
  9. Cochran, 1957, Analysis of covariance: Its nature and uses, Biometrics, 13, 261, 10.2307/2527916 · DOI · Paper page
  10. Díaz, 2020, Machine learning in the estimation of causal effects: Targeted minimum loss-based estimation and double/debiased machine learning, Biostatistics, 21, 353
  11. Díaz, 2020, Causal mediation analysis for stochastic interventions, Journal of the Royal Statistical Society: Series B (Statistical Methodology), 82, 661, 10.1111/rssb.12362 · DOI
  12. Flores, 2009, Identification and estimation of causal mechanisms and net effects of a treatment under unconfoundedness, Institute of Labor Economics (IZA), 4237
  13. Fowler-Brown, 2007, Risk of cardiovascular events and death: Does insurance matter?, Journal of General Internal Medicine, 22, 502, 10.1007/s11606-007-0127-2 · DOI
  14. Hahn, 1998, On the role of the propensity score in efficient semiparametric estimation of average treatment effects, Econometrica, 66, 315, 10.2307/2998560 · DOI · Paper page
  15. Hong, 2010, Ratio of mediator probability weighting for estimating natural direct and indirect effects, 2401
  16. Huber, 2014, Identifying causal mechanisms (primarily) based on inverse probability weighting, Journal of Applied Econometrics, 29, 920, 10.1002/jae.2341 · DOI
  17. Huber, 2015, Causal pitfalls in the decomposition of wage gaps, Journal of Business and Economic Statistics, 33, 179, 10.1080/07350015.2014.937437 · DOI
  18. Huber, 2017, Why do tougher caseworkers increase employment? The role of program assignment as a causal mechanism, Review of Economics and Statistics, 99, 180, 10.1162/REST_a_00632 · DOI
  19. Huber, 2018, Direct and indirect effects of training vouchers for the unemployed, Journal of the Royal Statistical Society: Series A (Statistics in Society), 181, 441, 10.1111/rssa.12279 · DOI
  20. Imai, 2013, Identification and sensitivity analysis for multiple causal mechanisms: Revisiting evidence from framing experiments, Political Analysis, 21, 141, 10.1093/pan/mps040 · DOI · Paper page
  21. Kanamori, 2012, Statistical analysis of kernel-based least-squares density-ratio estimation, Machine Learning, 86(3), 335, 10.1007/s10994-011-5266-3 · DOI
  22. Kaufman, 2004, A further critique of the analytic strategy of adjusting for covariates to identify biologic mediation, Epidemiologic Perspectives & Innovations, 1, 4, 10.1186/1742-5573-1-4 · DOI · Paper page
  23. Keele, 2015, Identifying mechanisms behind policy interventions via causal mediation analysis, Journal of Policy Analysis and Management, 34, 937, 10.1002/pam.21853 · DOI · Paper page
  24. King, 2009, Public policy for the poor? A randomised assessment of the Mexican universal health insurance programme, Lancet, 373, 1447, 10.1016/S0140-6736(09)60239-7 · DOI · Paper page
  25. Luo, 2016, High-dimensional l2boosting: Rate of convergence, arXiv preprint, arXiv:1602.08927 · arXiv
  26. Maciosek, 2010, Greater use of preventive services in us health care could save lives at little or no cost, Health Affairs, 29, 1656, 10.1377/hlthaff.2008.0701 · DOI · Paper page
  27. Miles, 2020, On semiparametric estimation of a path-specific effect in the presence of mediator-outcome confounding, Biometrika, 107, 159
  28. Nakanishi, 1996, Do preventive health services reduce eventual demand for medical care?, Social Science & Medicine, 43, 999, 10.1016/0277-9536(96)00016-0 · DOI
  29. Neyman, 1959, Optimal Asymptotic Tests of Composite Statistical Hypotheses, 213
  30. Pagán, 2007, Health insurance coverage and the use of preventive services by Mexican adults, Health Economics, 16, 1359, 10.1002/hec.1226 · DOI
  31. Pearl, 2001, Direct and indirect effects, 411
  32. Press, 2014, Insurance coverage and preventive care among adults
  33. Rasmussen, 2007, Preventive health screenings and health consultations in primary care increase life expectancy without increasing costs, Scandinavian Journal of Public Health, 35, 365, 10.1080/14034940701219642 · DOI · Paper page
  34. R Core Team, 2020, R: A Language and Environment for Statistical Computing
  35. Robins, 2003, Semantics of causal DAG models and the identification of direct and indirect effects, 70, 10.1093/oso/9780198510550.003.0007 · DOI · Paper page
  36. Robins, 1992, Identifiability and exchangeability for direct and indirect effects, Epidemiology, 3, 143, 10.1097/00001648-199203000-00013 · DOI · Paper page
  37. Robins, 1995, Semiparametric efficiency in multivariate regression models with missing data, Journal of the American Statistical Association, 90, 122, 10.1080/01621459.1995.10476494 · DOI · Paper page
  38. Robins, 1994, Estimation of regression coefficients when some regressors are not always observed, Journal of the American Statistical Association, 90, 846, 10.1080/01621459.1994.10476818 · DOI · Paper page
  39. Simon, 2017, The impact of health insurance on preventive care and health behaviors: Evidence from the first two years of the aca medicaid expansions, Journal of Policy Analysis and Management, 36, 390, 10.1002/pam.21972 · DOI · Paper page
  40. Sommers, 2017, Three-year impacts of the affordable care act: Improved medical care and health among low-income adults, Health Affairs, 36, 1119, 10.1377/hlthaff.2017.0293 · DOI · Paper page
  41. Sugiyama, 2010, Dimensionality reduction for density ratio estimation in high-dimensional spaces, Neural Networks, 23(1), 44, 10.1016/j.neunet.2009.07.007 · DOI
  42. Tchetgen Tchetgen, 2013, Inverse odds ratio-weighted estimation for causal mediation analysis, Statistics in Medicine, 32, 4567, 10.1002/sim.5864 · DOI · Paper page
  43. Tchetgen Tchetgen, 2012, Semiparametric theory for causal mediation analysis: Efficiency bounds, multiple robustness, and sensitivity analysis, Annals of Statistics, 40, 1816
  44. Ten Have, 2007, Causal mediation analyses with rank preserving models, Biometrics, 63, 926, 10.1111/j.1541-0420.2007.00766.x · DOI · Paper page
  45. Tibshirani, 1996, Regresson shrinkage and selection via the lasso, Journal of the Royal Statistical Society, 58, 267, 10.1111/j.2517-6161.1996.tb02080.x · DOI · Paper page
  46. van der Laan, 2006, Targeted maximum likelihood learning, International Journal of Biostatistics, 2, 1
  47. van der Laan, 2007, Super learner, Statistical Applications in Genetics and Molecular Biology, 6, Article 25
  48. VanderWeele, 2013, A three-way decomposition of a total effect into direct, indirect, and interactive effects, Epidemiology, 24, 224, 10.1097/EDE.0b013e318281a64e · DOI · Paper page
  49. Vansteelandt, 2012, Imputation strategies for the estimation of natural direct and indirect effects, Epidemiologic Methods, 1, 129, 10.1515/2161-962X.1014 · DOI · Paper page
  50. Yörük, 2016, Health insurance coverage and self-reported health: New estimates from the NLSY97, International Journal of Health Economics and Management, 16, 285, 10.1007/s10754-016-9189-1 · DOI
  51. Zheng, 2012, Targeted maximum likelihood estimation of natural direct effects, International Journal of Biostatistics, 8, 1, 10.2202/1557-4679.1361 · DOI

Cited by

  • Artificial intelligence for dementia prevention

    Newby, D., Orgeta, V., Marshall, C. R., Lourida, I., Albertyn, C. P., Tamburin, S., Raymont, V., Veldsman, M., Koychev, I., Bauermeister, S., et al. (2023) · Alzheimer's & Dementia · doi:10.1002/alz.13463 · Cited by 42

    “95 An alternative approach to inferring causality involves analyzing the timing and trajectory of risk factors: those with a causal effect are likely to show increasing levels of dementia risk with greater duration and intensity of exposure.”
    Background

Metadata

Authors
Farbmacher, H., Huber, M., Lafférs, L., Langen, H., & Spindler, M.
Venue
The Econometrics Journal
Published
2022-01-31
Volume
25
Issue
2
Pages
277-300
ISSN
  • 1368-4221
  • 1368-423X
OpenAlex
W4210400620
Cited by
198
References
13
Type
journal article
Language
en
Open access
closed

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.