{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T21:50:51Z","timestamp":1791323451775,"version":"4.3.1"},"reference-count":59,"publisher":"Institute of Mathematical Statistics","issue":"5","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Ann. Statist."],"published-print":{"date-parts":[[2026,10,1]]},"abstract":"Crossover designs randomly assign each unit to receive a sequence of treatments. By comparing outcomes within the same unit, these designs can effectively eliminate between-unit variation and facilitate the identification of both instantaneous effects of current treatments and carryover effects from past treatments. They are widely used in traditional biomedical studies and are increasingly adopted in modern digital platforms. However, standard analyses of crossover designs often rely on strong parametric models, making inference vulnerable to model misspecification. This paper adopts the design-based framework to analyze general crossover designs. We make two main contributions. First, we use potential outcomes to formally define the causal estimands and assumptions on the data-generating process. For any given type of crossover design and assumptions on potential outcomes, we outline a procedure for identification and estimation, emphasizing the central role of the treatment assignment mechanism in the design-based inference. Second, we unify the analysis of crossover designs using least squares, with restrictions on the coefficients and weights on the units. Based on the theory, we recommend the specification of the regression function, weighting scheme, and coefficient restrictions to assess identifiability, construct efficient estimators, and conduct inference in a unified fashion. Crucially, the least squares procedure is simple to implement, and yields not only consistent and efficient point estimates but also valid inference even when the working regression model is misspecified.<\/jats:p>","DOI":"10.1214\/26-aos2663","type":"journal-article","created":{"date-parts":[[2026,10,2]],"date-time":"2026-10-02T12:55:28Z","timestamp":1790945728000},"page":"2614-2636","source":"Crossref","is-referenced-by-count":0,"title":["Principled analysis of crossover designs: Causal effects, efficient estimation and robust inference"],"prefix":"10.1214","volume":"54","author":[{"given":"Zhichao","family":"Jiang","sequence":"first","affiliation":[{"name":"School of Mathematics, Sun Yat-sen University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2704-2353","authenticated-orcid":false,"given":"Peng","family":"Ding","sequence":"additional","affiliation":[{"name":"Department of Statistics, University of California, Berkeley"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"108","reference":[{"key":"14","doi-asserted-by":"publisher","unstructured":"FOGARTY, C. B.<\/small> (2018). Regression-assisted inference for the average treatment effect in paired experiments. Biometrika<\/i> 105<\/b> 994\u20131000.","DOI":"10.1093\/biomet\/asy034"},{"key":"16","doi-asserted-by":"publisher","unstructured":"FREEDMAN, D. A.<\/small> (2008). Randomization does not justify logistic regression. Statist. Sci.<\/i> 23<\/b> 237\u2013249.","DOI":"10.1214\/08-STS262"},{"key":"32","doi-asserted-by":"crossref","unstructured":"LI, X.<\/small> and DING, P.<\/small> (2020). Rerandomization and regression adjustment. J. R. Stat. Soc. Ser. B. Stat. Methodol.<\/i> 82<\/b> 241\u2013268.","DOI":"10.1111\/rssb.12353"},{"key":"33","doi-asserted-by":"crossref","unstructured":"Liang, K.-Y.<\/small> and Zeger, S. L.<\/small> (1986). Longitudinal data analysis using generalized linear models. Biometrika<\/i> 73<\/b> 13\u201322.","DOI":"10.1093\/biomet\/73.1.13"},{"key":"5","doi-asserted-by":"publisher","unstructured":"BLONIARZ, A.<\/small>, LIU, H.<\/small>, ZHANG, C.-H.<\/small>, SEKHON, J. S.<\/small> and YU, B.<\/small> (2016). Lasso adjustments of treatment effect estimates in randomized experiments. Proc. Natl. Acad. Sci. USA<\/i> 113<\/b> 7383\u20137390.","DOI":"10.1073\/pnas.1510506113"},{"key":"7","doi-asserted-by":"publisher","unstructured":"BOJINOV, I.<\/small> and SHEPHARD, N.<\/small> (2019). Time series experiments and causal estimands: Exact randomization tests and trading. J. Amer. Statist. Assoc.<\/i> 114<\/b> 1665\u20131682.","DOI":"10.1080\/01621459.2018.1527225"},{"key":"15","doi-asserted-by":"publisher","unstructured":"FREEDMAN, D. A.<\/small> (2008). On regression adjustments to experimental data. Adv. in Appl. Math.<\/i> 40<\/b> 180\u2013193.","DOI":"10.1016\/j.aam.2006.12.003"},{"key":"21","doi-asserted-by":"publisher","unstructured":"IMAI, K.<\/small>, JIANG, Z.<\/small> and MALANI, A.<\/small> (2021). Causal inference with interference and noncompliance in two-stage randomized experiments. J. Amer. Statist. Assoc.<\/i> 116<\/b> 632\u2013644.","DOI":"10.1080\/01621459.2020.1775612"},{"key":"36","doi-asserted-by":"publisher","unstructured":"LIU, H.<\/small> and YANG, Y.<\/small> (2020). Regression-adjusted average treatment effect estimates in stratified randomized experiments. Biometrika<\/i> 107<\/b> 935\u2013948.","DOI":"10.1093\/biomet\/asaa038"},{"key":"45","doi-asserted-by":"crossref","unstructured":"PASHLEY, N. E.<\/small> and MIRATRIX, L. W.<\/small> (2021). Insights on variance estimation for blocked and matched pairs designs. J. Educ. Behav. Stat.<\/i> 46<\/b> 271\u2013296.","DOI":"10.3102\/1076998620946272"},{"key":"22","doi-asserted-by":"crossref","unstructured":"IMBENS, G. W.<\/small> and RUBIN, D. B.<\/small> (2015). Causal Inference in Statistics, Social, and Biomedical Sciences<\/i>. Cambridge Univ. Press, New York.","DOI":"10.1017\/CBO9781139025751"},{"key":"2","doi-asserted-by":"publisher","unstructured":"ATHEY, S.<\/small> and IMBENS, G. W.<\/small> (2022). Design-based analysis in difference-in-differences settings with staggered adoption. J. Econometrics<\/i> 226<\/b> 62\u201379.","DOI":"10.1016\/j.jeconom.2020.10.012"},{"key":"10","doi-asserted-by":"crossref","unstructured":"COCHRAN, W.<\/small> (1939). Long-term agricultural experiments. Suppl. J. R. Stat. Soc.<\/i> 6<\/b> 104\u2013148.","DOI":"10.2307\/2983686"},{"key":"34","doi-asserted-by":"publisher","unstructured":"Lin, W.<\/small> (2013). Agnostic notes on regression adjustments to experimental data: Reexamining Freedman\u2019s critique. Ann. Appl. Stat.<\/i> 7<\/b> 295\u2013318.","DOI":"10.1214\/12-AOAS583"},{"key":"56","doi-asserted-by":"crossref","unstructured":"Zellner, A.<\/small> (1962). An efficient method of estimating seemingly unrelated regressions and tests for aggregation bias. J. Amer. Statist. Assoc.<\/i> 57<\/b> 348\u2013368.","DOI":"10.1080\/01621459.1962.10480664"},{"key":"31","doi-asserted-by":"publisher","unstructured":"Li, X.<\/small> and Ding, P.<\/small> (2017). General forms of finite population central limit theorems with applications to causal inference. J. Amer. Statist. Assoc.<\/i> 112<\/b> 1759\u20131769.","DOI":"10.1080\/01621459.2017.1295865"},{"key":"13","doi-asserted-by":"publisher","unstructured":"DASGUPTA, T.<\/small>, PILLAI, N. S.<\/small> and RUBIN, D. B.<\/small> (2015). Causal inference from 2<\/mn><\/mrow>K<\/mi><\/mrow><\/msup><\/math> factorial designs by using potential outcomes. J. R. Stat. Soc. Ser. B. Stat. Methodol.<\/i> 77<\/b> 727\u2013753.","DOI":"10.1111\/rssb.12085"},{"key":"18","doi-asserted-by":"publisher","unstructured":"GUO, K.<\/small> and BASSE, G.<\/small> (2023). The generalized Oaxaca\u2013Blinder estimator. J. Amer. Statist. Assoc.<\/i> 118<\/b> 524\u2013536.","DOI":"10.1080\/01621459.2021.1941053"},{"key":"58","doi-asserted-by":"publisher","unstructured":"ZHAO, A.<\/small> and DING, P.<\/small> (2022). Regression-based causal inference with factorial experiments: Estimands, model specifications and design-based properties. Biometrika<\/i> 109<\/b> 799\u2013815.","DOI":"10.1093\/biomet\/asab051"},{"key":"59","doi-asserted-by":"publisher","unstructured":"ZHAO, A.<\/small> and DING, P.<\/small> (2023). Covariate adjustment in multiarmed, possibly factorial experiments. J. R. Stat. Soc. Ser. B. Stat. Methodol.<\/i> 85<\/b> 1\u201323.","DOI":"10.1093\/jrsssb\/qkac003"},{"key":"6","doi-asserted-by":"publisher","unstructured":"BOJINOV, I.<\/small>, RAMBACHAN, A.<\/small> and SHEPHARD, N.<\/small> (2021). Panel experiments and dynamic causal effects: A finite population perspective. Quant. Econ.<\/i> 12<\/b> 1171\u20131196.","DOI":"10.3982\/qe1744"},{"key":"24","doi-asserted-by":"publisher","unstructured":"JIANG, Z.<\/small>, CHEN, S.<\/small> and DING, P.<\/small> (2023). An instrumental variable method for point processes: Generalized Wald estimation based on deconvolution. Biometrika<\/i> 110<\/b> 989\u20131008.","DOI":"10.1093\/biomet\/asad005"},{"key":"1","doi-asserted-by":"crossref","unstructured":"ARKHANGELSKY, D.<\/small> and IMBENS, G.<\/small> (2024). Causal models for longitudinal and panel data: A survey. Econom. J.<\/i> 27<\/b> C1\u2013C61.","DOI":"10.1093\/ectj\/utae014"},{"key":"3","unstructured":"BALAAM, L.<\/small> (1968). A two-period design with t<\/mi><\/mrow>2<\/mn><\/mrow><\/msup><\/math> experimental units. Biometrics<\/i> 24<\/b> 61\u201373."},{"key":"4","doi-asserted-by":"crossref","unstructured":"BASSE, G. W.<\/small>, DING, Y.<\/small> and TOULIS, P.<\/small> (2023). Minimax designs for causal effects in temporal experiments with treatment habituation. Biometrika<\/i> 110<\/b> 155\u2013168.","DOI":"10.1093\/biomet\/asac024"},{"key":"8","doi-asserted-by":"crossref","unstructured":"BOJINOV, I.<\/small>, SIMCHI-LEVI, D.<\/small> and ZHAO, J.<\/small> (2023). Design and analysis of switchback experiments. Manag. Sci.<\/i> 69<\/b> 3759\u20133777.","DOI":"10.1287\/mnsc.2022.4583"},{"key":"9","doi-asserted-by":"crossref","unstructured":"BROWN, B. W. JR<\/small> (1980). The crossover experiment for clinical trials. Biometrics<\/i> 36<\/b> 69\u201379.","DOI":"10.2307\/2530496"},{"key":"11","doi-asserted-by":"crossref","unstructured":"COCHRAN, W. G.<\/small> and COX, G. M.<\/small> (1950). Experimental Designs<\/i>. Wiley, NJ.","DOI":"10.1097\/00010694-195008000-00014"},{"key":"12","doi-asserted-by":"crossref","unstructured":"COOK, T. D.<\/small> and DEMETS, D. L.<\/small> (2007). Introduction to Statistical Methods for Clinical Trials<\/i>. CRC Press, New York.","DOI":"10.1201\/9781420009965"},{"key":"17","doi-asserted-by":"crossref","unstructured":"GREENE, W. H.<\/small> and SEAKS, T. G.<\/small> (1991). The restricted least squares estimator: A pedagogical note. Rev. Econ. Stat.<\/i> 73<\/b> 563\u2013567.","DOI":"10.2307\/2109587"},{"key":"19","doi-asserted-by":"crossref","unstructured":"HEDAYAT, A.<\/small>, STUFKEN, J.<\/small> and YANG, M.<\/small> (2006). Optimal and efficient crossover designs when subject effects are random. J. Amer. Statist. Assoc.<\/i> 101<\/b> 1031\u20131038.","DOI":"10.1198\/016214505000001384"},{"key":"20","unstructured":"HINKELMANN, K.<\/small> and KEMPTHORNE, O.<\/small> (2007). Design and Analysis of Experiments, Volume<\/i> 2: Advanced Experimental Design<\/i>. John Wiley & Sons, NJ."},{"key":"23","unstructured":"JANKAR, J.<\/small> and MANDAL, A.<\/small> (2021). Optimal crossover designs for generalized linear models: An application to work environment experiment. Statist. Appl.<\/i> 19<\/b> 319\u2013336."},{"key":"25","doi-asserted-by":"crossref","unstructured":"JIANG, Z.<\/small> and DING, P.<\/small> (2026). Supplement to \u201cPrincipled analysis of crossover designs: causal effects, efficient estimation and robust inference.\u201d https:\/\/doi.org\/10.1214\/26-AOS2663SUPP<\/a>","DOI":"10.1214\/26-AOS2663SUPP"},{"key":"26","doi-asserted-by":"crossref","unstructured":"JONES, B.<\/small> and KENWARD, M. G.<\/small> (1987). Modelling binary data from a three-period cross-over trial. Stat. Med.<\/i> 6<\/b> 555\u2013564.","DOI":"10.1002\/sim.4780060504"},{"key":"27","doi-asserted-by":"crossref","unstructured":"JONES, B.<\/small> and KENWARD, M. G.<\/small> (2003). Design and Analysis of Cross-over Trials<\/i>. Chapman and Hall, New York.","DOI":"10.1201\/9781420036091"},{"key":"28","doi-asserted-by":"crossref","unstructured":"KEMPTHORNE, O.<\/small> (1952). The Design and Analysis of Experiments<\/i>. Williams & Wilkins, Philadelphia.","DOI":"10.1097\/00010694-195205000-00012"},{"key":"29","doi-asserted-by":"crossref","unstructured":"KERSHNER, R. P.<\/small> and FEDERER, W. T.<\/small> (1981). Two-treatment crossover designs for estimating a variety of effects. J. Amer. Statist. Assoc.<\/i> 76<\/b> 612\u2013619.","DOI":"10.1080\/01621459.1981.10477693"},{"key":"30","doi-asserted-by":"crossref","unstructured":"KUNERT, J.<\/small> and STUFKEN, J.<\/small> (2008). Optimal crossover designs for two treatments in the presence of mixed and self-carryover effects. J. Amer. Statist. Assoc.<\/i> 103<\/b> 1641\u20131647.","DOI":"10.1198\/016214508000000760"},{"key":"35","unstructured":"LIN, Z.<\/small> and DING, P.<\/small> (2025). Unifying regression-based and design-based causal inference in time-series experiments. arXiv preprint. Available at arXiv:2510.22864<\/a>."},{"key":"37","doi-asserted-by":"crossref","unstructured":"MADSEN, J. E. H.<\/small>, SCHEIKE, T.<\/small> and PIPPER, C.<\/small> (2023). Unbiased and efficient estimation of causal treatment effects in crossover trials. Biom. J.<\/i> 65<\/b> 2200170.","DOI":"10.1002\/bimj.202200170"},{"key":"38","doi-asserted-by":"crossref","unstructured":"MILLIKEN, G. A.<\/small> and JOHNSON, D. E.<\/small> (2009). Analysis of Messy Data, Volume I<\/i>: Designed Experiments<\/i>. CRC Press, New York.","DOI":"10.1201\/EBK1584883340"},{"key":"39","unstructured":"MILLS, E. J.<\/small>, CHAN, A.-W.<\/small>, WU, P.<\/small>, VAIL, A.<\/small>, GUYATT, G. H.<\/small> and ALTMAN, D. G.<\/small> (2009). Design, analysis, and presentation of crossover trials. Trials<\/i> 10<\/b> 1\u20136."},{"key":"40","doi-asserted-by":"crossref","unstructured":"MIRATRIX, L. W.<\/small>, SEKHON, J. S.<\/small> and YU, B.<\/small> (2013). Adjusting treatment effect estimates by post-stratification in randomized experiments. J. R. Stat. Soc. Ser. B. Stat. Methodol.<\/i> 75<\/b> 369\u2013396.","DOI":"10.1111\/j.1467-9868.2012.01048.x"},{"key":"41","doi-asserted-by":"crossref","unstructured":"MURRAY, G. D.<\/small> (1994). Editorial: Special issue on crossover trials. Stat. Methods Med. Res.<\/i> 3<\/b> 301\u2013302.","DOI":"10.1177\/096228029400300401"},{"key":"42","doi-asserted-by":"crossref","unstructured":"NEYMAN, J.<\/small> (1990). On the application of probability theory to agricultural experiments. Essay on principles. Section 9. Statist. Sci.<\/i> 5<\/b> 465\u2013472. Translated from the 1923 Polish original and and edited by D. M. Dabrowska and T. P. Speed.","DOI":"10.1214\/ss\/1177012031"},{"key":"43","unstructured":"NI, T.<\/small> (2025). Decision analytics of switchback experiments: A robust optimization approach. Available at SSRN 5245482."},{"key":"44","unstructured":"NI, T.<\/small> and BOJINOV, I.<\/small> (2025). Enhancing efficiency and robustness for switchback experiments: A practical model-assisted framework. Available at SSRN 5229804."},{"key":"46","doi-asserted-by":"crossref","unstructured":"PASHLEY, N. E.<\/small> and MIRATRIX, L. W.<\/small> (2022). Block what you can, except when you shouldn\u2019t. J. Educ. Behav. Stat.<\/i> 47<\/b> 69\u2013100.","DOI":"10.3102\/10769986211027240"},{"key":"47","doi-asserted-by":"crossref","unstructured":"PATTERSON, S. D.<\/small> and JONES, B.<\/small> (2017). Bioequivalence and Statistics in Clinical Pharmacology<\/i>. CRC Press, New York.","DOI":"10.1201\/9781315374161"},{"key":"48","unstructured":"RAO, C. R.<\/small> (1973). Linear Statistical Inference and Its Applications<\/i>. Wiley, New York."},{"key":"49","doi-asserted-by":"crossref","unstructured":"SENN, S. S.<\/small> (2002). Cross-over Trials in Clinical Research<\/i>, 5th ed. Wiley, NJ.","DOI":"10.1002\/0470854596"},{"key":"50","doi-asserted-by":"crossref","unstructured":"SHI, D.<\/small> and YE, T.<\/small> (2024). Behavioral carry-over effect and power consideration in crossover trials. Biometrics<\/i> 80<\/b> ujae023.","DOI":"10.1093\/biomtc\/ujae023"},{"key":"51","unstructured":"SUN, J.<\/small>, GUO, P.<\/small>, CHEN, X.<\/small> and TAN, X.<\/small> (2025). Dynamic treatment effect analysis in crossover designs through repeated measures. Stat. Med.<\/i> 44<\/b> e70070."},{"key":"52","unstructured":"THEIL, H.<\/small> (1971). Principles of Econometrics<\/i>. Wiley, NJ."},{"key":"53","doi-asserted-by":"crossref","unstructured":"WANG, X.<\/small> and CHINCHILLI, V. M.<\/small> (2021). Analysis of crossover designs with nonignorable dropout. Stat. Med.<\/i> 40<\/b> 64\u201384.","DOI":"10.1002\/sim.8762"},{"key":"54","unstructured":"WU, C. J.<\/small> and HAMADA, M. S.<\/small> (2011). Experiments<\/i>: Planning, Analysis, and Optimization<\/i>. Wiley, NJ."},{"key":"55","doi-asserted-by":"crossref","unstructured":"XU, R.<\/small>, MEHROTRA, D. V.<\/small> and SHAW, P. A.<\/small> (2018). Incorporating baseline measurements into the analysis of crossover trials with time-to-event endpoints. Stat. Med.<\/i> 37<\/b> 3280\u20133292.","DOI":"10.1002\/sim.7834"},{"key":"57","doi-asserted-by":"crossref","unstructured":"ZHAO, A.<\/small> and DING, P.<\/small> (2022). Reconciling design-based and model-based causal inferences for split-plot experiments. Ann. Statist.<\/i> 50<\/b> 1170\u20131192.","DOI":"10.1214\/21-AOS2144"}],"container-title":["The Annals of Statistics"],"original-title":[],"deposited":{"date-parts":[[2026,10,6]],"date-time":"2026-10-06T21:05:56Z","timestamp":1791320756000},"score":1,"resource":{"primary":{"URL":"https:\/\/projecteuclid.org\/journals\/annals-of-statistics\/volume-54\/issue-5\/Principled-analysis-of-crossover-designs--Causal-effects-efficient-estimation\/10.1214\/26-AOS2663.full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,1]]},"references-count":59,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2026,10,1]]}},"URL":"https:\/\/doi.org\/10.1214\/26-aos2663","relation":{},"ISSN":["0090-5364","2168-8966"],"issn-type":[{"value":"0090-5364","type":"print"},{"value":"2168-8966","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,1]]}}}