{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,24]],"date-time":"2026-06-24T06:50:59Z","timestamp":1782283859024,"version":"3.54.5"},"reference-count":38,"publisher":"Springer Science and Business Media LLC","issue":"5","license":[{"start":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T00:00:00Z","timestamp":1777593600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,5,6]],"date-time":"2026-05-06T00:00:00Z","timestamp":1778025600000},"content-version":"vor","delay-in-days":5,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100016378","name":"Technische Universit\u00e4t Dortmund","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100016378","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Mach Learn"],"published-print":{"date-parts":[[2026,5]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    We consider the problem of estimating the expected causal effect\n                    <jats:inline-formula>\n                      <jats:tex-math>$$E[Y|do(X)]$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    for a target variable\n                    <jats:inline-formula>\n                      <jats:tex-math>$$Y$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    when treatment\n                    <jats:inline-formula>\n                      <jats:tex-math>$$X$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    is set by intervention, focusing on continuous random variables. In settings without selection bias or confounding,\n                    <jats:inline-formula>\n                      <jats:tex-math>$$E[Y|do(X)] = E[Y|X]$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    , which can be estimated using standard regression methods. However, regression fails when systematic missingness induced by selection bias, or confounding distorts the data. Proxy variables unaffected by the selection process can, under certain constraints, be used to correct for selection bias to recover\n                    <jats:inline-formula>\n                      <jats:tex-math>$$E[Y|X]$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    , and hence\n                    <jats:inline-formula>\n                      <jats:tex-math>$$E[Y|do(X)]$$<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    , reliably. When data is additionally affected by confounding, recovering the causal effect from selection-biased data is more challenging and requires access to proxies to both correct for confounding and for the selection mechanism. Assuming access to such proxies from external unbiased observational data, we derive theoretical conditions ensuring identifiability and recoverability of causal effects. We further introduce a linear two-step regression estimator (TSR), which can be extended through adding non-linear basis functions, capable of exploiting proxy variables to adjust for selection bias while accounting for confounding. We show that TSR is consistent with previous estimators when confounding is absent, but achieves a lower variance. Extensive simulation studies validate TSR\u2019s correctness for scenarios that include both selection bias and confounding with proxy variables.\n                  <\/jats:p>","DOI":"10.1007\/s10994-026-07054-6","type":"journal-article","created":{"date-parts":[[2026,5,6]],"date-time":"2026-05-06T09:04:48Z","timestamp":1778058288000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Regression-Based Estimation of Causal Effects in the Presence of Selection Bias and Confounding"],"prefix":"10.1007","volume":"115","author":[{"ORCID":"https:\/\/orcid.org\/0009-0006-0222-0025","authenticated-orcid":false,"given":"Marlies","family":"Hafer","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1575-824X","authenticated-orcid":false,"given":"Alexander","family":"Marx","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,5,6]]},"reference":[{"key":"7054_CR1","doi-asserted-by":"publisher","first-page":"399","DOI":"10.1080\/00273171.2011.568786","volume":"46","author":"PC Austin","year":"2011","unstructured":"Austin, P. C. (2011). An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behavioral Research, 46, 399\u2013424. https:\/\/doi.org\/10.1080\/00273171.2011.568786","journal-title":"Multivariate Behavioral Research"},{"issue":"11","key":"7054_CR2","doi-asserted-by":"publisher","first-page":"3661","DOI":"10.1002\/sim.6607","volume":"34","author":"PC Austin","year":"2015","unstructured":"Austin, P. C., & Stuart, E. A. (2015). Moving towards best practice when using inverse probability of treatment weighting (iptw) using the propensity score to estimate causal treatment effects in observational studies. Statistics in Medicine, 34(11), 3661\u20133679. https:\/\/doi.org\/10.1002\/sim.6607","journal-title":"Statistics in Medicine"},{"key":"7054_CR3","unstructured":"Bareinboim, E., & Pearl, J. (2012). Controlling selection bias in causal inference. In Proceedings of the fifteenth international conference on artificial intelligence and statistics, vol\u00a022. PMLR, pp. 100\u2013108"},{"key":"7054_CR4","doi-asserted-by":"publisher","unstructured":"Bareinboim, E., & Pearl, J. (2016). Causal inference and the data-fusion problem. In Proceedings of the national academy of sciences, pp. 7345\u20137352. https:\/\/doi.org\/10.1073\/pnas.1510507113","DOI":"10.1073\/pnas.1510507113"},{"key":"7054_CR5","doi-asserted-by":"publisher","unstructured":"Bareinboim, E., & Tian, J. (2015). Recovering causal effects from selection bias. In Proceedings of the AAAI conference on artificial intelligence, pp. 3475\u20133481, https:\/\/doi.org\/10.1609\/aaai.v29i1.9679","DOI":"10.1609\/aaai.v29i1.9679"},{"key":"7054_CR6","doi-asserted-by":"publisher","unstructured":"Bareinboim, E., Tian, J., & Pearl, J. (2014). Recovering from selection bias in causal and statistical inference. In Proceedings of the AAAI conference on artificial intelligence, pp. 2410\u20132416. https:\/\/doi.org\/10.1609\/aaai.v28i1.9074","DOI":"10.1609\/aaai.v28i1.9074"},{"issue":"3","key":"7054_CR7","doi-asserted-by":"publisher","first-page":"47","DOI":"10.2307\/3002000","volume":"2","author":"J Berkson","year":"1946","unstructured":"Berkson, J. (1946). Limitations of the application of fourfold table analysis to hospital data. Biometrics Bulletin, 2(3), 47\u201353. https:\/\/doi.org\/10.2307\/3002000","journal-title":"Biometrics Bulletin"},{"key":"7054_CR8","unstructured":"Boeken, P., de Kroon, N., de Jong, M., Mooij, J. M., & Zoeter, O. (2023). Correcting for selection bias and missing response in regression using privileged information. In Uncertainty in artificial intelligence (pp. 195-205). PMLR"},{"key":"7054_CR9","unstructured":"Chen, L., Zoeter, O., & Mooij, J. M. (2024). Modeling latent selection with structural causal models. arxiv:2401.06925"},{"key":"7054_CR10","doi-asserted-by":"publisher","unstructured":"Correa, J. D., & Bareinboim, E. (2017). Causal effect identification by adjustment under confounding and selection biases. In Proceedings of the AAAI conference on artificial intelligence, pp. 3740\u20133746. https:\/\/doi.org\/10.1609\/aaai.v31i1.11060","DOI":"10.1609\/aaai.v31i1.11060"},{"key":"7054_CR11","doi-asserted-by":"publisher","unstructured":"Correa, J.D., Tian, J., & Bareinboim, E. (2018). Generalized adjustment under confounding and selection biases. In Proceedings of the AAAI conference on artificial intelligence, pp. 6335\u20136342. https:\/\/doi.org\/10.1609\/aaai.v32i1.12125","DOI":"10.1609\/aaai.v32i1.12125"},{"key":"7054_CR12","doi-asserted-by":"publisher","unstructured":"Correa, J.D., Tian, J., & Bareinboim, E. (2019). Identification of causal effects in the presence of selection bias. In Proceedings of the AAAI conference on artificial intelligence, pp. 2744\u20132751. https:\/\/doi.org\/10.1609\/aaai.v33i01.33012744","DOI":"10.1609\/aaai.v33i01.33012744"},{"issue":"1","key":"7054_CR13","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1111\/j.2517-6161.1977.tb01600.x","volume":"39","author":"AP Dempster","year":"1977","unstructured":"Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from incomplete data via the em algorithm. Journal of the Royal Statistical Society: Series B (Methodological), 39(1), 1\u201322. https:\/\/doi.org\/10.1111\/j.2517-6161.1977.tb01600.x","journal-title":"Journal of the Royal Statistical Society: Series B (Methodological)"},{"key":"7054_CR14","doi-asserted-by":"publisher","first-page":"3807","DOI":"10.1002\/sim.7389","volume":"36","author":"I D\u00edaz","year":"2017","unstructured":"D\u00edaz, I., & van der Laan, M. J. (2017). Doubly robust inference for targeted minimum loss\u2013based estimation in randomized trials with missing outcome data. Statistics in Medicine, 36, 3807\u20133819. https:\/\/doi.org\/10.1002\/sim.7389","journal-title":"Statistics in Medicine"},{"key":"7054_CR15","doi-asserted-by":"crossref","unstructured":"Dost, K., Taskova, K., Riddle, P., & Wicker, J. (2020). Your best guess when you know nothing: Identification and mitigation of selection bias. In 2020 IEEE international conference on data mining (ICDM) (pp. 996-1001). IEEE.","DOI":"10.1109\/ICDM50108.2020.00115"},{"key":"7054_CR16","unstructured":"Forr\u00e9, P., & Mooij, J. M. (2020). Causal calculus in the presence of cycles, latent confounders and selection bias. In Proceedings of the 35th annual conference on uncertainty in artificial intelligence, pp. 71\u201380"},{"issue":"1","key":"7054_CR17","doi-asserted-by":"publisher","first-page":"153","DOI":"10.2307\/1912352","volume":"47","author":"JJ Heckman","year":"1979","unstructured":"Heckman, J. J. (1979). Sample selection bias as a specification error. Econometrica, 47(1), 153\u2013161. https:\/\/doi.org\/10.2307\/1912352","journal-title":"Econometrica"},{"issue":"4","key":"7054_CR18","doi-asserted-by":"publisher","first-page":"6","DOI":"10.1111\/1740-9713.01413","volume":"17","author":"A Herbert","year":"2020","unstructured":"Herbert, A., Griffith, G., Hemani, G., et al. (2020). The Spectre of Berkson\u2019s paradox: Collider bias in covid-19 research. Significance, 17(4), 6\u20137. https:\/\/doi.org\/10.1111\/1740-9713.01413","journal-title":"Significance"},{"key":"7054_CR19","doi-asserted-by":"publisher","unstructured":"Hill, J. L. (2011). Bayesian nonparametric modeling for causal inference. In Proceedings of the 35th annual conference on uncertainty in artificial intelligence, pp. 217\u2013240. https:\/\/doi.org\/10.1198\/jcgs.2010.08162","DOI":"10.1198\/jcgs.2010.08162"},{"issue":"2","key":"7054_CR20","doi-asserted-by":"publisher","first-page":"467","DOI":"10.2307\/2951620","volume":"62","author":"GW Imbens","year":"1994","unstructured":"Imbens, G. W., & Angrist, J. D. (1994). Identification and estimation of local average treatment effects. Econometrica, 62(2), 467\u2013475. https:\/\/doi.org\/10.2307\/2951620","journal-title":"Econometrica"},{"key":"7054_CR21","doi-asserted-by":"crossref","unstructured":"Jung, Y., Tian, J., & Bareinboim, E. (2024). Unified covariate adjustment for causal inference. In Proceedings of the 38th international conference on neural information processing systems, pp. 6448\u20136499","DOI":"10.52202\/079017-0208"},{"key":"7054_CR22","doi-asserted-by":"publisher","unstructured":"Kaltenpoth, D., & Vreeken, J. (2023). Identifying selection bias from observational data. In Proceedings of the AAAI conference on artificial intelligence, pp. 8177\u20138185. https:\/\/doi.org\/10.1609\/aaai.v37i7.25987","DOI":"10.1609\/aaai.v37i7.25987"},{"key":"7054_CR23","doi-asserted-by":"publisher","unstructured":"Kuang, K., Cui, P., Li, B., Jiang, M., Yang, S., & Wang, F. (2017). Treatment effect estimation with data-driven variable decomposition. In Proceedings of the AAAI conference on artificial intelligence (Vol. 31, No. 1). https:\/\/doi.org\/10.1609\/aaai.v31i1.10480","DOI":"10.1609\/aaai.v31i1.10480"},{"key":"7054_CR24","doi-asserted-by":"publisher","first-page":"697","DOI":"10.1111\/jep.12714","volume":"23","author":"A Linden","year":"2017","unstructured":"Linden, A. (2017). Improving causal inference with a doubly robust estimator that combines propensity score stratification and weighting. Journal of Evaluation in Clinical Practice, 23, 697\u2013702. https:\/\/doi.org\/10.1111\/jep.12714","journal-title":"Journal of Evaluation in Clinical Practice"},{"key":"7054_CR25","doi-asserted-by":"publisher","DOI":"10.1002\/9781119013563","volume-title":"Statistical analysis with missing data","author":"RJA Little","year":"2002","unstructured":"Little, R. J. A., & Rubin, D. B. (2002). Statistical analysis with missing data (2nd ed.). Wiley.","edition":"2"},{"key":"7054_CR26","doi-asserted-by":"publisher","DOI":"10.1093\/aje\/kwae370","author":"J Liu","year":"2025","unstructured":"Liu, J., Park, C., Li, K., et al. (2025). Regression-based proximal causal inference. American Journal of Epidemiology. https:\/\/doi.org\/10.1093\/aje\/kwae370","journal-title":"American Journal of Epidemiology"},{"key":"7054_CR27","unstructured":"Louizos, C., Shalit, U., Mooij, J. M., Sontag, D., Zemel, R., & Welling, M. (2017). Causal effect inference with deep latent-variable models. Advances in Neural Information Processing Systems, 30"},{"issue":"4","key":"7054_CR28","doi-asserted-by":"publisher","first-page":"987","DOI":"10.1093\/biomet\/asy038","volume":"105","author":"W Miao","year":"2018","unstructured":"Miao, W., Geng, Z., & Tchetgen Tchetgen, E. J. (2018). Identifying causal effects with proxy variables of an unmeasured confounder. Biometrika, 105(4), 987\u2013993. https:\/\/doi.org\/10.1093\/biomet\/asy038","journal-title":"Biometrika"},{"issue":"2","key":"7054_CR29","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3592389","volume":"1","author":"W Miao","year":"2024","unstructured":"Miao, W., Liu, L., Li, Y., et al. (2024). Identification and semiparametric efficiency theory of nonignorable missing data with a shadow variable. ACM\/JMS Journal of Data Science, 1(2), 1\u201323. https:\/\/doi.org\/10.1145\/3592389","journal-title":"ACM\/JMS Journal of Data Science"},{"issue":"534","key":"7054_CR30","doi-asserted-by":"publisher","first-page":"1023","DOI":"10.1080\/01621459.2021.1874961","volume":"116","author":"K Mohan","year":"2021","unstructured":"Mohan, K., & Pearl, J. (2021). Graphical models for processing missing data. Journal of the American Statistical Association, 116(534), 1023\u20131037. https:\/\/doi.org\/10.1080\/01621459.2021.1874961","journal-title":"Journal of the American Statistical Association"},{"key":"7054_CR31","doi-asserted-by":"publisher","first-page":"131","DOI":"10.1515\/ijb-2015-0028","volume":"12","author":"R Neugebauer","year":"2016","unstructured":"Neugebauer, R., Schmittdiel, J. A., & van der Laan, M. J. (2016). A case study of the impact of data-adaptive versus model-based estimation of the propensity scores on causal inferences from three inverse probability weighting estimators. The International Journal of Biostatistics, 12, 131\u2013155. https:\/\/doi.org\/10.1515\/ijb-2015-0028","journal-title":"The International Journal of Biostatistics"},{"key":"7054_CR32","doi-asserted-by":"publisher","unstructured":"Pearl, J. (2009). Causality, 2nd edn. Cambridge University Press, Cambridge. https:\/\/doi.org\/10.1017\/cbo9780511803161","DOI":"10.1017\/cbo9780511803161"},{"key":"7054_CR33","unstructured":"Pearl, J. (2012). A solution to a class of selection-bias problems. https:\/\/ftp.cs.ucla.edu\/pub\/stat_ser\/r405.pdf"},{"key":"7054_CR34","doi-asserted-by":"publisher","first-page":"41","DOI":"10.2307\/2335942","volume":"70","author":"PR Rosenbaum","year":"1983","unstructured":"Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70, 41\u201355. https:\/\/doi.org\/10.2307\/2335942","journal-title":"Biometrika"},{"issue":"387","key":"7054_CR35","doi-asserted-by":"publisher","first-page":"516","DOI":"10.2307\/2288398","volume":"79","author":"PR Rosenbaum","year":"1984","unstructured":"Rosenbaum, P. R., & Rubin, D. B. (1984). Reducing bias in observational studies using subclassification on the propensity score. Journal of the American Statistical Association, 79(387), 516\u2013524. https:\/\/doi.org\/10.2307\/2288398","journal-title":"Journal of the American Statistical Association"},{"issue":"10","key":"7054_CR36","doi-asserted-by":"publisher","first-page":"927","DOI":"10.1089\/cmb.2024.0666","volume":"31","author":"R Schill","year":"2024","unstructured":"Schill, R., Klever, M., L\u00f6sch, A., et al. (2024). Correcting for observation bias in cancer progression modeling. Journal of Computational Biology, 31(10), 927\u2013945. https:\/\/doi.org\/10.1089\/cmb.2024.0666","journal-title":"Journal of Computational Biology"},{"key":"7054_CR37","unstructured":"Singh, R., & Zhou, H. (2022). Kernel methods for long term dose response curves. arxiv:2201.05139"},{"issue":"2","key":"7054_CR38","doi-asserted-by":"publisher","first-page":"101","DOI":"10.1007\/s41060-021-00259-z","volume":"12","author":"Y Wang","year":"2021","unstructured":"Wang, Y., & Singh, L. (2021). Analyzing the impact of missing values and selection bias on fairness. International Journal of Data Science and Analytics, 12(2), 101\u2013119. https:\/\/doi.org\/10.1007\/s41060-021-00259-z","journal-title":"International Journal of Data Science and Analytics"}],"container-title":["Machine Learning"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10994-026-07054-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10994-026-07054-6","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10994-026-07054-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,24]],"date-time":"2026-06-24T06:31:50Z","timestamp":1782282710000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10994-026-07054-6"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5]]},"references-count":38,"journal-issue":{"issue":"5","published-print":{"date-parts":[[2026,5]]}},"alternative-id":["7054"],"URL":"https:\/\/doi.org\/10.1007\/s10994-026-07054-6","relation":{},"ISSN":["0885-6125","1573-0565"],"issn-type":[{"value":"0885-6125","type":"print"},{"value":"1573-0565","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5]]},"assertion":[{"value":"31 October 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"24 March 2026","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"14 April 2026","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"6 May 2026","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no conflict of interest.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of interest"}},{"value":"Not applicable.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethical Approval"}},{"value":"Not applicable.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent to Participate"}}],"article-number":"115"}}