About

Donald B. Rubin is an Emeritus Professor of Statistics at Harvard University. His research interests include causal inference in experiments and observational studies, inference in sample surveys with nonresponse and in missing data problems, and the application of Bayesian and empirical Bayesian techniques. Rubin has developed and applied statistical models across various scientific disciplines. He earned his Ph.D. in Statistics from Harvard University in 1970, his M.A. in Computer Science from Harvard in 1966, and his A.B. in Psychology from Princeton University in 1965. Rubin served as a Professor in the Department of Statistics at Harvard from 1984 until his retirement in 2018, and he held the position of Chairman of the department during two separate periods. His professional experience also includes being a Fellow in Theoretical and Applied Statistics at the National Bureau of Economic Research and a Research Associate at NORC in Chicago.

Research topics

• Computer Science
• Political Science
• Data Mining
• Computer Security
• Mathematical analysis
• World Wide Web
• Mathematics
• Applied mathematics
• Data science
• Medicine
• Statistics

Selected publications

• A Bayesian Criterion for Rerandomization

  Journal of the American Statistical Association · 2025-06-02 · 1 citations

  article
  PublisherDOI

• Rubin Causal Model

  International Encyclopedia of Statistical Science · 2025-01-01

  book-chapter1st authorCorresponding
  PublisherDOI

• On a conjecture of Gill on Bell experiments

  Examples and Counterexamples · 2025-10-10

  articleOpen access1st authorCorresponding

  This note gives a counterexample to a conjecture of Richard Gill that was stated as open problem related to the statistical analysis of Bell experiments in quantum mechanics, and which can formulated as an elementary probability problem.

  PublisherDOI

• Catalytic Priors: Using Synthetic Data to Specify Prior Distributions in Bayesian Analysis

  Statistical Science · 2025-05-01

  article

  Catalytic prior distributions provide general, easy-to-use, and interpretable specifications of prior distributions for Bayesian analysis. They are particularly beneficial when the observed data are inadequate to stably estimate a complex target model. A catalytic prior distribution is constructed by augmenting the observed data with synthetic data that are sampled from the predictive distribution of a simpler model estimated from the observed data. We illustrate the usefulness of the catalytic prior approach using an example from labor economics. In the example, the resulting Bayesian inference reflects many important aspects of the observed data, and the estimation accuracy and predictive performance of the inference based on the catalytic prior are superior to, or comparable to, that of other commonly used prior distributions. We further explore the connection between the catalytic prior approach and a few popular regularization methods. We expect the catalytic prior approach to be useful in many applications.

  PublisherDOI

• Conditionally Affinely Invariant Rerandomization and its Admissible Complete Class

  arXiv (Cornell University) · 2024-02-17

  preprintOpen accessSenior author

  Rerandomization utilizes modern computing ability to improve covariate balance while adhering to the randomization principle originally advocated by RA Fisher. Affinely invariant rerandomization has the ``Equal Percent Variance Reducing'' (EPVR) property. When dealing with covariates of varying importance and/or mixed types, the conditionally EPVR property is often more desired. We discuss a general class of conditionally affinely invariant rerandomization methods and obtain their conditionally EPVR property. In addition, we set up a decision-theoretical framework to evaluate balance criteria for rerandomization. Popular rerandomization methods, such as the covariate balance table check, are found to be inadmissible. We suggest an admissible complete class of conditionally affinely invariant balance criteria, which can be applied to experimental designs involving tiers of covariates, stratification, and multiple treatment arms.

  PublisherOA PDFDOI

• Counternull Sets in Randomized Experiments

  The American Statistician · 2024-12-09 · 1 citations

  articleOpen accessSenior author

  of values. We explore advantages to reporting a counternull set in addition to the p-value associated with a null value; a first advantage is pedagogical, in that reporting it avoids the mistake of implicitly accepting a not-rejected null hypothesis; a second advantage is that the effort to construct a counternull set can be scientifically helpful by encouraging thought about nonnull values of estimands. Two examples are used to illustrate these ideas.

  PublisherOA PDFDOI

• Contrast-specific propensity scores for causal inference with multiple interventions

  Statistical Methods in Medical Research · 2024-03-18

  articleSenior author

  Existing methods that use propensity scores for heterogeneous treatment effect estimation on non-experimental data do not readily extend to the case of more than two treatment options. In this work, we develop a new propensity score-based method for heterogeneous treatment effect estimation when there are three or more treatment options, and prove that it generates unbiased estimates. We demonstrate our method on a real patient registry of patients in Singapore with diabetic dyslipidemia. On this dataset, our method generates heterogeneous treatment recommendations for patients among three options: Statins, fibrates, and non-pharmacological treatment to control patients' lipid ratios (total cholesterol divided by high-density lipoprotein level). In our numerical study, our proposed method generated more stable estimates compared to a benchmark method based on a multi-dimensional propensity score.

  PublisherDOI

• A201 EFFECT OF MIRIKIZUMAB ON BOWEL URGENCY CLINICALLY MEANINGFUL IMPROVEMENT AND REMISSION: RESULTS FROM THE PHASE 3 LUCENT INDUCTION AND MAINTENANCE STUDIES

  Journal of the Canadian Association of Gastroenterology · 2023-03-01

  articleOpen access

  Abstract Background Bowel urgency (BU) was assessed in mirikizumab (miri) Phase 3 LUCENT studies in moderately-to-severely active UC using the validated Urgency Numeric Rating Scale (UNRS). UNRS measures BU severity in the past 24 hours from 0 (no urgency) to 10 (worst possible urgency). Psychometric evaluation of the UNRS showed Clinically Meaningful Improvement (CMI) is >3 point change; Remission is a score of 0 or 1. Purpose This analysis evaluated the proportions of patients in LUCENT studies achieving BU CMI and BU remission. Method The modified intent-to-treat (mITT) population (patients receiving ≥1 dose of miri or placebo (PBO); N= 1281) was randomized at induction study baseline in a 3:1 ratio to IV doses of 300mg miri or PBO every 4 weeks (Q4W) during induction (W0, 4, and 8). Patients achieving Clinical Response, measured by Modified Mayo Score (MMS), to miri during induction were re-randomized at W0 of the maintenance study in a 2:1 ratio to subcutaneous (SC) 200mg miri or PBO Q4W through W40 (52 weeks of treatment). Patients recorded their UNRS score daily in an e-diary. Mean weekly UNRS scores were calculated from diary data if ≥4 days of data were available. Rates of BU CMI and BU remission in the miri v PBO groups were compared at W12 (induction) in the mITT population with a baseline UNRS score ≥3, and W52 (maintenance) among miri clinical responders at W12 with a baseline UNRS score ≥3. Cochran-Mantel-Haenszel tests with non-responder imputation for missing values were used for all treatment comparisons. Result(s) Patient population: mean age 43 years, 60% male, disease duration 7 years; 63.0% left-sided colitis; 36.3% pancolitis; 46.7% moderate disease (MMS 4-6); 53.2% severe disease (MMS 7-9). Significantly higher proportions of miri versus PBO patients achieved BU CMI (48.7% v 32.2%) and BU remission (22.1% v 12.3%) at W12 (both p<0.001; Table) in the induction study. Similarly, at W40 of maintenance, significantly greater proportion of miri patients achieved BU CMI (65.2% v 41.9%) and BU remission (42.9% v 25.0%) compared to PBO among miri induction responders (both p<0.001; Table). Image Conclusion(s) Miri had a highly significant and clinically meaningful benefit on reducing bowel urgency, one of the most disruptive UC symptoms. The Urgency Numeric Rating Scale usefully quantified the baseline level and change in bowel urgency after treatment across a spectrum of severity. Please acknowledge all funding agencies by checking the applicable boxes below Other Please indicate your source of funding; Eli Lilly and Company Disclosure of Interest None Declared

  PublisherOA PDFDOI

• Bayesian Criterion for Re-randomization

  arXiv (Cornell University) · 2023-03-14

  preprintOpen access

  Re-randomization has gained popularity as a tool for experiment-based causal inference due to its superior covariate balance and statistical efficiency compared to classic randomized experiments. However, the basic re-randomization method, known as ReM, and many of its extensions have been deemed sub-optimal as they fail to prioritize covariates that are more strongly associated with potential outcomes. To address this limitation and design more efficient re-randomization procedures, a more precise quantification of covariate heterogeneity and its impact on the causal effect estimator is in a great appeal. This work fills in this gap with a Bayesian criterion for re-randomization and a series of novel re-randomization procedures derived under such a criterion. Both theoretical analyses and numerical studies show that the proposed re-randomization procedures under the Bayesian criterion outperform existing ReM-based procedures significantly in effectively balancing covariates and precisely estimating the unknown causal effect.

  PublisherOA PDFDOI

• A216 BOWEL URGENCY COMMUNICATION GAP BETWEEN HEALTH CARE PROFESSIONALS AND PATIENTS WITH ULCERATIVE COLITIS IN THE US AND EUROPE: COMMUNICATING NEEDS AND FEATURES OF IBD EXPERIENCES (CONFIDE) SURVEY

  Journal of the Canadian Association of Gastroenterology · 2023-03-01 · 2 citations

  articleOpen access

  Abstract Background The Communicating Needs and Features of IBD Experiences (CONFIDE) study aims to increase understanding of the impact of symptoms on patients with moderate to severe UC and Crohn’s disease and to investigate gaps in communication with healthcare professionals (HCPs) in the United States (US), Europe (EUR), and Japan. Purpose This report focuses on patients with moderate to severe UC and HCPs from the US and EUR. Method Online, quantitative, cross-sectional surveys of patients with UC and HCPs were conducted in the US and EUR (France, Germany, Italy, Spain, and UK). HCP surveys included physicians and non-physician HCPs responsible for making prescribing decisions. Moderate to severe UC was defined based on treatment, steroid use, and/or hospitalization history. Data collected included perspectives on the experience of patients with UC. Result(s) A total of 200 US (62% male, mean age 40.4 years) and 556 EUR patients (57% male, mean age 38.9 years), and 200 US and 503 EUR HCPs completed the survey. According to US and EUR patients, the top 3 symptoms currently (past month) experienced were diarrhoea (63% and 50%), bowel urgency (47% and 30%) and increased stool frequency (39% and 30%). Blood in stool was reported as currently experienced by 27% and 24% of US and EUR patients, respectively. Among patients currently experiencing bowel urgency, 47% of US and 27% of EUR patients discuss this symptom at every appointment. Among those who do not discuss bowel urgency at every appointment, 74% and 75% of US and EUR patients would like to discuss this symptom more frequently with their HCP. A total of 30% and 43% of US and EUR patients that ever experienced bowel urgency were not comfortable reporting it to their HCP, with 62% and 58% of these US and EUR patients feeling embarrassed talking about this symptom (Table). HCPs in both the US and EUR ranked diarrhoea (74% and 65%), blood in stool (69% and 65%) and increased stool frequency (38% and 34%) as the top 3 symptoms most reported by patients. According to US and EUR HCPs, the top 4 symptoms proactively discussed in routine appointments were blood in stool (93% and 94%), diarrhoea (90% and 91%), increased stool frequency (82% and 82%) and bowel urgency (76% and 82%). Among HCPs who did not proactively discuss bowel urgency, 47% of US and 40% of EUR HCPs expect patients to bring this up if it is an issue. Image Conclusion(s) Communication gaps were similar between US and EUR patients and HCPs. Bowel urgency is the second-most reported symptom by patients with moderate to severe UC. However, this symptom is not among the HCP-perceived top 3 most reported symptoms. Although a substantial proportion of patients reported a desire to discuss bowel urgency more frequently with their HCP, some patients reported feeling embarrassed talking about it. Many HCPs who do not proactively discuss this symptom expect patients to bring this up. A communication gap was identified and highlights the under-appreciation of bowel urgency as an important symptom of UC. Please acknowledge all funding agencies by checking the applicable boxes below Other Please indicate your source of funding; Eli Lilly and Company Disclosure of Interest S. Travis Grant / Research support from: AbbVie, BUHLMANN Diagnostics, ECCO, Eli Lilly and Company, Ferring Pharmaceuticals, International Organization for the Study of Inflammatory Bowel Disease, Janssen, Merck Sharp & Dohme, Normal Collision Foundation, Pfizer, Procter & Gamble, Schering-Plough, Takeda, UCB Pharma, Vifor Pharma, and Warner Chilcott, A. Bleakman Employee of: Eli Lilly and Company, D. Rubin Grant / Research support from: Takeda, Consultant of: AbbVie, Allergan, AltruBio, American College of Gastroenterology, Arena Pharmaceuticals, Athos Therapeutics, Bellatrix Pharmaceuticals, Boehringer Ingelheim, Bristol Myers Squibb, Celgene/Syneos Health, Cornerstones Health (non-profit), Eli Lilly and Company, Galen/Atlantica, Genentech/Roche, Gilead Sciences, GoDuRn, InDex Pharmaceuticals, Ironwood Pharmaceuticals, Iterative Scopes, Janssen, Materia Prima, Pfizer, Prometheus Therapeutics and Diagnostics, Reistone Biopharma, Takeda, and TechLab, M. Dubinsky Shareholder of: Trellus Health, Grant / Research support from: AbbVie, Janssen, Pfizer, and Prometheus Biosciences, Consultant of: AbbVie, Arena Pharmaceuticals, Boehringer Ingelheim, Bristol Myers Squibb, Celgene, Eli Lilly and Company, F. Hoffmann-La Roche, Genentech, Gilead Sciences, Janssen, Pfizer, Prometheus Therapeutics and Diagnostics, Takeda, and UCB Pharma, R. Panaccione Grant / Research support from: AbbVie, Ferring Pharmaceuticals, Janssen, Pfizer, and Takeda, Consultant of: Abbott, AbbVie, Alimentiv, Amgen, Arena Pharmaceuticals, AstraZeneca, Biogen, Boehringer Ingelheim, Bristol Myers Squibb, Celgene, Celltrion, Cosmo Pharmaceuticals, Eisai, Elan Pharma, Eli Lilly and Company, Ferring Pharmaceuticals, Galapagos NV, Genentech, Gilead Sciences, GlaxoSmithKline, Janssen, Merck, Mylan, Oppilan Pharma, Pandion Therapeutics, Pfizer, Progenity, Protagonist Therapeutics, Roche, Sandoz, Satisfai Health, Shire, Sublimity Therapeutics, Takeda, Theravance Biopharma, and UCB Pharma, T. Hibi Grant / Research support from: AbbVie, Activaid, Alfresa Pharma, Bristol Myers Squibb, Eli Lilly Japan K.K., Ferring Pharmaceuticals, Gilead Sciences, Janssen Pharmaceutical K.K., JMDC, Nippon Kayaku, Mochida Pharmaceutical, Pfizer Japan, and Takeda, Consultant of: AbbVie, Apo Plus Station, Bristol Myers Squibb, Celltrion, EA Pharma, Eli Lilly and Company, Gilead Sciences, Janssen, Kyorin, Mitsubishi Tanabe Pharma, Nichi-Iko Pharmaceutical, Pfizer, Takeda, and Zeria Pharmaceutical, Speakers bureau of: AbbVie, Aspen Japan K.K., Ferring Pharmaceuticals, Gilead Sciences, Janssen, JIMRO, Mitsubishi Tanabe Pharma, Mochida Pharmaceutical, Pfizer, and Takeda, T. Gibble Employee of: Eli Lilly and Company, C. Kayhan Employee of: Eli Lilly and Company, E. Flynn Employee of: Eli Lilly and Company, C. Sapin Employee of: Eli Lilly and Company, C. Atkinson Consultant of: Eli Lilly and Company in connection with the development of this publication, Employee of: Adelphi Real World, S. Schreiber Grant / Research support from: personal fees and/or travel support from: AbbVie, Amgen, Arena Pharmaceuticals, Biogen, Bristol Myers Squibb, Celgene, Celltrion, Eli Lilly and Company, Dr. Falk Pharma, Ferring Pharmaceuticals, Fresenius Kabi, Galapagos NV, Gilead Sciences, I-MAB Biopharma, Janssen, Merck Sharp & Dohme, Mylan, Novartis, Pfizer, Protagonist Therapeutics, Provention Bio, Roche, Sandoz/Hexal, Shire, Takeda, Theravance Biopharma, and UCB Pharma, J. Jones: None Declared

  PublisherOA PDFDOI

Recent grants

• Collaborative Research: Generalized Propensity Score Methods

  NSF · $135k · 2006–2010

Frequent coauthors

• Guido W. Imbens

  Stanford University

  79 shared

• Robert Rosenthal

  57 shared

• Constantine Frangakis

  39 shared

• John Barnard

  32 shared

• Roderick J. A. Little

  University of Michigan–Ann Arbor

  31 shared

• Jennifer Hill

  New York University

  28 shared

• Xiao‐Hua Zhou

  Peking University International Hospital

  25 shared

• Martin McIntosh

  25 shared

Labs

• Sports Analytics Laboratory at Harvard UniversityPI

  The Sports Analytics Laboratory at Harvard University focuses on the application of statistical methods to sports data.

Education

• B.A., Mathematics

  Harvard University

  1963

• M.A., Statistics

  Harvard University

  1965

• Ph.D., Statistics

  Harvard University

  1968