{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,4]],"date-time":"2026-03-04T16:24:28Z","timestamp":1772641468228,"version":"3.50.1"},"reference-count":55,"publisher":"Oxford University Press (OUP)","issue":"5","license":[{"start":{"date-parts":[[2024,5,9]],"date-time":"2024-05-09T00:00:00Z","timestamp":1715212800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100000266","name":"EPSRC","doi-asserted-by":"publisher","award":["Grant EP\/N031938\/1"],"award-info":[{"award-number":["Grant EP\/N031938\/1"]}],"id":[{"id":"10.13039\/501100000266","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,11,13]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>We study partially linear models in settings where observations are arranged in independent groups but may exhibit within-group dependence. Existing approaches estimate linear model parameters through weighted least squares, with optimal weights (given by the inverse covariance of the response, conditional on the covariates) typically estimated by maximizing a (restricted) likelihood from random effects modelling or by using generalized estimating equations. We introduce a new \u2018sandwich loss\u2019 whose population minimizer coincides with the weights of these approaches when the parametric forms for the conditional covariance are well-specified, but can yield arbitrarily large improvements in linear parameter estimation accuracy when they are not. Under relatively mild conditions, our estimated coefficients are asymptotically Gaussian and enjoy minimal variance among estimators with weights restricted to a given class of functions, when user-chosen regression methods are used to estimate nuisance functions. We further expand the class of functional forms for the weights that may be fitted beyond parametric models by leveraging the flexibility of modern machine learning methods within a new gradient boosting scheme for minimizing the sandwich loss. We demonstrate the effectiveness of both the sandwich loss and what we call \u2018sandwich boosting\u2019 in a variety of settings with simulated and real-world data.<\/jats:p>","DOI":"10.1093\/jrsssb\/qkae032","type":"journal-article","created":{"date-parts":[[2024,5,9]],"date-time":"2024-05-09T13:51:58Z","timestamp":1715262718000},"page":"1286-1311","source":"Crossref","is-referenced-by-count":1,"title":["Sandwich boosting for accurate estimation in partially linear models for grouped data"],"prefix":"10.1093","volume":"86","author":[{"given":"Elliot H","family":"Young","sequence":"first","affiliation":[{"name":"Statistical Laboratory, University of Cambridge , Cambridge ,","place":["UK"]}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9073-3782","authenticated-orcid":false,"given":"Rajen D","family":"Shah","sequence":"additional","affiliation":[{"name":"Statistical Laboratory, University of Cambridge , Cambridge ,","place":["UK"]}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"286","published-online":{"date-parts":[[2024,5,9]]},"reference":[{"issue":"1","key":"2024111314115154700_qkae032-B1","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18637\/jss.v067.i01","article-title":"Fitting linear mixed-effects models using lme4","volume":"67","author":"Bates","year":"2015","journal-title":"Journal of Statistical Software"},{"issue":"2","key":"2024111314115154700_qkae032-B2","doi-asserted-by":"publisher","first-page":"587","DOI":"10.1016\/j.ijforecast.2020.07.007","article-title":"Kaggle forecasting competitions: An overlooked learning opportunity","volume":"37","author":"Bojer","year":"2021","journal-title":"International Journal of Forecasting"},{"issue":"356","key":"2024111314115154700_qkae032-B3","doi-asserted-by":"publisher","first-page":"791","DOI":"10.1080\/01621459.1976.10480949","article-title":"Science and statistics","volume":"71","author":"Box","year":"1976","journal-title":"Journal of the American Statistical Association"},{"issue":"7","key":"2024111314115154700_qkae032-B4","doi-asserted-by":"publisher","first-page":"1493","DOI":"10.1162\/089976699300016106","article-title":"Prediction games and arcing algorithms","volume":"11","author":"Breiman","year":"1999","journal-title":"Neural Computation"},{"issue":"4","key":"2024111314115154700_qkae032-B5","doi-asserted-by":"publisher","first-page":"477","DOI":"10.1214\/07-STS242","article-title":"Boosting algorithms: Regularization, prediction and model fitting","volume":"22","author":"B\u00fchlmann","year":"2007","journal-title":"Statistical Science"},{"issue":"462","key":"2024111314115154700_qkae032-B6","doi-asserted-by":"publisher","first-page":"324","DOI":"10.1198\/016214503000125","article-title":"Boosting with the L2 loss","volume":"98","author":"B\u00fchlmann","year":"2003","journal-title":"Journal of the American Statistical Association"},{"issue":"4","key":"2024111314115154700_qkae032-B7","doi-asserted-by":"publisher","first-page":"1224","DOI":"10.1214\/aos\/1176345987","article-title":"Adapting for heteroscedasticity in linear models","volume":"10","author":"Carroll","year":"1982","journal-title":"The Annals of Statistics"},{"key":"2024111314115154700_qkae032-B8","doi-asserted-by":"publisher","author":"Chen","year":"2016","DOI":"10.1145\/2939672.2939785"},{"issue":"1","key":"2024111314115154700_qkae032-B9","doi-asserted-by":"publisher","first-page":"C1","DOI":"10.1111\/ectj.12097","article-title":"Double\/debiased machine learning for treatment and structural parameters","volume":"21","author":"Chernozhukov","year":"2018","journal-title":"The Econometrics Journal"},{"issue":"1","key":"2024111314115154700_qkae032-B10","doi-asserted-by":"publisher","first-page":"31","DOI":"10.2307\/1267913","article-title":"Restricted maximum likelihood (REML) estimation of variance components in the mixed model","volume":"18","author":"Corbeil","year":"1976","journal-title":"Technometrics"},{"issue":"2","key":"2024111314115154700_qkae032-B11","doi-asserted-by":"publisher","first-page":"407","DOI":"10.1093\/biomet\/82.2.407","article-title":"On the use of a working correlation matrix in using generalised linear models for repeated measures","volume":"82","author":"Crowder","year":"1995","journal-title":"Biometrika"},{"key":"2024111314115154700_qkae032-B12","volume-title":"Analysis of longitudinal data","author":"Diggle","year":"2013","edition":"2nd ed."},{"key":"2024111314115154700_qkae032-B13","author":"Emmenegger","year":"2021"},{"issue":"2","key":"2024111314115154700_qkae032-B14","doi-asserted-by":"publisher","first-page":"6461","DOI":"10.1214\/21-EJS1931","article-title":"Regularizing double machine learning in partially linear endogenous models","volume":"15","author":"Emmenegger","year":"2021","journal-title":"Electronic Journal of Statistics"},{"issue":"4","key":"2024111314115154700_qkae032-B15","doi-asserted-by":"publisher","first-page":"1553","DOI":"10.1111\/sjos.12639","article-title":"Plug-in machine learning for partially linear mixed-effects models with repeated measurements","volume":"50","author":"Emmenegger","year":"2023","journal-title":"Scandinavian Journal of Statistics"},{"key":"2024111314115154700_qkae032-B16","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4757-3454-6","volume-title":"Multivariate statistical modelling based on generalized linear models","author":"Fahrmeir","year":"2001","edition":"2nd ed."},{"key":"2024111314115154700_qkae032-B17","author":"Freund","year":"1996"},{"issue":"2","key":"2024111314115154700_qkae032-B18","doi-asserted-by":"publisher","first-page":"337","DOI":"10.1214\/aos\/1016218223","article-title":"Additive logistic regression: A statistical view of boosting (with discussion and a rejoinder by the authors)","volume":"28","author":"Friedman","year":"2000","journal-title":"The Annals of Statistics"},{"issue":"1","key":"2024111314115154700_qkae032-B19","doi-asserted-by":"publisher","first-page":"43","DOI":"10.1093\/biomet\/73.1.43","article-title":"Multilevel mixed linear model analysis using iterative generalized least squares","volume":"73","author":"Goldstein","year":"1986","journal-title":"Biometrika"},{"issue":"3","key":"2024111314115154700_qkae032-B20","doi-asserted-by":"publisher","first-page":"622","DOI":"10.1093\/biomet\/76.3.622","article-title":"Restricted unbiased iterative generalized least-squares estimation","volume":"76","author":"Goldstein","year":"1989","journal-title":"Biometrika"},{"key":"2024111314115154700_qkae032-B21","author":"Gourieroux","year":"1993"},{"issue":"3","key":"2024111314115154700_qkae032-B22","doi-asserted-by":"publisher","first-page":"681","DOI":"10.2307\/1913471","article-title":"Pseudo maximum likelihood methods: Theory","volume":"52","author":"Gourieroux","year":"1984","journal-title":"Econometrica"},{"key":"2024111314115154700_qkae032-B23","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18637\/jss.v015.i02","article-title":"The R package geepack for generalized estimating equations","volume":"15\/2","author":"Halekoh","year":"2006","journal-title":"Journal of Statistical Software"},{"key":"2024111314115154700_qkae032-B24","volume-title":"Generalized estimating equations","author":"Hardin","year":"2003"},{"issue":"1-2","key":"2024111314115154700_qkae032-B25","doi-asserted-by":"publisher","first-page":"93","DOI":"10.1093\/biomet\/54.1-2.93","article-title":"Maximum-likelihood estimation for the mixed analysis of variance model","volume":"54","author":"Hartley","year":"1967","journal-title":"Biometrika"},{"issue":"1","key":"2024111314115154700_qkae032-B26","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1214\/ss\/1009212671","article-title":"Marginalized multilevel models and likelihood inference","volume":"15","author":"Heagerty","year":"2000","journal-title":"Statistical Science"},{"issue":"3","key":"2024111314115154700_qkae032-B27","doi-asserted-by":"publisher","first-page":"451","DOI":"10.1111\/sjos.2007.34.issue-3","article-title":"Efficient estimation in marginal partially linear models for longitudinal\/clustered data using splines","volume":"34","author":"Huang","year":"2007","journal-title":"Scandinavian Journal of Statistics"},{"key":"2024111314115154700_qkae032-B28","author":"Huber","year":"1967"},{"key":"2024111314115154700_qkae032-B29","author":"Kennedy","year":"2022"},{"issue":"540","key":"2024111314115154700_qkae032-B30","doi-asserted-by":"publisher","first-page":"1835","DOI":"10.1080\/01621459.2021.1888740","article-title":"Inference for high-dimensional linear mixed-effects models: A quasi-likelihood approach","volume":"117","author":"Li","year":"2022","journal-title":"Journal of the American Statistical Association"},{"issue":"4","key":"2024111314115154700_qkae032-B31","doi-asserted-by":"publisher","first-page":"721","DOI":"10.4310\/SII.2018.v11.n4.a15","article-title":"Doubly regularized estimation and selection in linear mixed-effects models for high-dimensional longitudinal data","volume":"11","author":"Li","year":"2018","journal-title":"Statistics and its Interface"},{"issue":"1","key":"2024111314115154700_qkae032-B32","doi-asserted-by":"publisher","first-page":"13","DOI":"10.1093\/biomet\/73.1.13","article-title":"Longitudinal data analysis using generalized linear models","volume":"73","author":"Liang","year":"1986","journal-title":"Biometrika"},{"issue":"1","key":"2024111314115154700_qkae032-B33","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1111\/rssb.1992.54.issue-1","article-title":"Multivariate regression analyses for categorical data","volume":"54","author":"Liang","year":"1992","journal-title":"Journal of the Royal Statistical Society. Series B (Methodological)"},{"issue":"1","key":"2024111314115154700_qkae032-B34","doi-asserted-by":"publisher","first-page":"354","DOI":"10.2307\/2533173","article-title":"Generalized estimating equations for ordinal data: A note on working correlation structures","volume":"52","author":"Lumley","year":"1996","journal-title":"Biometrics"},{"key":"2024111314115154700_qkae032-B35","author":"Mason","year":"1999"},{"key":"2024111314115154700_qkae032-B36","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4899-3242-6","volume-title":"Generalized linear models","author":"McCullagh","year":"1989","edition":"2nd ed."},{"key":"2024111314115154700_qkae032-B37","author":"Park","year":"2021"},{"key":"2024111314115154700_qkae032-B38","author":"Pinheiro","year":"2022"},{"key":"2024111314115154700_qkae032-B39","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4419-0318-1","volume-title":"Mixed-effects models in S and S-PLUS","author":"Pinheiro","year":"2000"},{"issue":"3","key":"2024111314115154700_qkae032-B40","doi-asserted-by":"publisher","first-page":"825","DOI":"10.2307\/2532642","article-title":"Estimating equations for parameters in means and covariances of multivariate discrete and continuous responses","volume":"47","author":"Prentice","year":"1991","journal-title":"Biometrics"},{"issue":"429","key":"2024111314115154700_qkae032-B41","doi-asserted-by":"publisher","first-page":"122","DOI":"10.1080\/01621459.1995.10476494","article-title":"Semiparametric efficiency in multivariate regression models with missing data","volume":"90","author":"Robins","year":"1995","journal-title":"Journal of the American Statistical Association"},{"issue":"427","key":"2024111314115154700_qkae032-B42","doi-asserted-by":"publisher","first-page":"846","DOI":"10.1080\/01621459.1994.10476818","article-title":"Estimation of regression coefficients when some regressors are not always observed","volume":"89","author":"Robins","year":"1994","journal-title":"Journal of the American Statistical Association"},{"issue":"4","key":"2024111314115154700_qkae032-B43","doi-asserted-by":"publisher","first-page":"875","DOI":"10.2307\/1911033","article-title":"Asymptotically efficient estimation in the presence of heteroskedasticity of unknown form","volume":"55","author":"Robinson","year":"1987","journal-title":"Econometrica"},{"issue":"2","key":"2024111314115154700_qkae032-B44","doi-asserted-by":"publisher","first-page":"221","DOI":"10.2307\/1403146","article-title":"Model robust confidence intervals using maximum likelihood estimators","volume":"54","author":"Royall","year":"1986","journal-title":"International Statistical Review \/ Revue Internationale de Statistique"},{"issue":"2","key":"2024111314115154700_qkae032-B45","doi-asserted-by":"publisher","first-page":"197","DOI":"10.1007\/BF00116037","article-title":"The strength of weak learnability","volume":"5","author":"Schapire","year":"1990","journal-title":"Machine Learning"},{"key":"2024111314115154700_qkae032-B46","volume-title":"Semiparametric theory and missing data","author":"Tsiatis","year":"2006"},{"key":"2024111314115154700_qkae032-B47","author":"U. D. o. L. Bureau of Labor Statistics","year":"2004"},{"key":"2024111314115154700_qkae032-B48","author":"U. o. C. B. S. o. B. James M. Kilts Center","year":"2022"},{"key":"2024111314115154700_qkae032-B49","doi-asserted-by":"crossref","DOI":"10.1017\/CBO9780511802256","volume-title":"Asymptotic statistics","author":"van der Vaart","year":"1998"},{"issue":"3","key":"2024111314115154700_qkae032-B50","doi-asserted-by":"publisher","first-page":"657","DOI":"10.1111\/rssb.12504","article-title":"Assumption-lean inference for generalised linear model parameters","volume":"84","author":"Vansteelandt","year":"2022","journal-title":"Journal of the Royal Statistical Society Series B: Statistical Methodology"},{"key":"2024111314115154700_qkae032-B51","doi-asserted-by":"crossref","DOI":"10.1201\/9781315370279","volume-title":"Generalized additive models: An introduction with R","author":"Wood","year":"2017","edition":"2nd ed."},{"issue":"8","key":"2024111314115154700_qkae032-B52","doi-asserted-by":"publisher","first-page":"1539","DOI":"10.1016\/j.jmva.2007.06.011","article-title":"Statistical inference of partially linear regression models with heteroscedastic errors","volume":"98","author":"You","year":"2007","journal-title":"Journal of Multivariate Analysis"},{"issue":"3","key":"2024111314115154700_qkae032-B53","doi-asserted-by":"publisher","first-page":"689","DOI":"10.2307\/2532783","article-title":"Semiparametric models for longitudinal data with application to CD4 cell numbers in HIV seroconverters","volume":"50","author":"Zeger","year":"1994","journal-title":"Biometrics"},{"issue":"1","key":"2024111314115154700_qkae032-B54","doi-asserted-by":"publisher","first-page":"121","DOI":"10.2307\/2531248","article-title":"Longitudinal data analysis for discrete and continuous outcomes","volume":"42","author":"Zeger","year":"1986","journal-title":"Biometrics"},{"key":"2024111314115154700_qkae032-B55","doi-asserted-by":"crossref","DOI":"10.1007\/978-1-4614-0499-6","volume-title":"Generalized estimating equations","author":"Ziegler","year":"2011"}],"container-title":["Journal of the Royal Statistical Society Series B: Statistical Methodology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/86\/5\/1286\/60648638\/qkae032.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/jrsssb\/article-pdf\/86\/5\/1286\/60648638\/qkae032.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,11,13]],"date-time":"2024-11-13T14:12:14Z","timestamp":1731507134000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/jrsssb\/article\/86\/5\/1286\/7667645"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,5,9]]},"references-count":55,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2024,5,9]]},"published-print":{"date-parts":[[2024,11,13]]}},"URL":"https:\/\/doi.org\/10.1093\/jrsssb\/qkae032","relation":{},"ISSN":["1369-7412","1467-9868"],"issn-type":[{"value":"1369-7412","type":"print"},{"value":"1467-9868","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2024,11]]},"published":{"date-parts":[[2024,5,9]]}}}