Donald B. Rubin
YOU?
Author Swipe
View article: Counternull Sets in Randomized Experiments
Counternull Sets in Randomized Experiments Open
Consider a study whose primary results are "not statistically significant". How often does it lead to the following published conclusion that "there is no effect of the treatment/exposure on the outcome"? We believe too often and that the …
View article: Conditionally Affinely Invariant Rerandomization and its Admissible Complete Class
Conditionally Affinely Invariant Rerandomization and its Admissible Complete Class Open
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'' …
View article: Towards more scientific meta-analyses
Towards more scientific meta-analyses Open
Meta-analysis can be a critical part of the research process, often serving as the primary analysis on which the practitioners, policymakers, and individuals base their decisions. However, current literature synthesis approaches to meta-an…
View article: Bayesian Criterion for Re-randomization
Bayesian Criterion for Re-randomization Open
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,…
View article: 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
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 Open
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…
View article: A201 EFFECT OF MIRIKIZUMAB ON BOWEL URGENCY CLINICALLY MEANINGFUL IMPROVEMENT AND REMISSION: RESULTS FROM THE PHASE 3 LUCENT INDUCTION AND MAINTENANCE STUDIES
A201 EFFECT OF MIRIKIZUMAB ON BOWEL URGENCY CLINICALLY MEANINGFUL IMPROVEMENT AND REMISSION: RESULTS FROM THE PHASE 3 LUCENT INDUCTION AND MAINTENANCE STUDIES Open
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 …
View article: PCA Rerandomization
PCA Rerandomization Open
Mahalanobis distance of covariate means between treatment and control groups is often adopted as a balance criterion when implementing a rerandomization strategy. However, this criterion may not work well for high‐dimensional cases because…
View article: High-dimensional randomization-based inference capitalizing on classical design and modern computing
High-dimensional randomization-based inference capitalizing on classical design and modern computing Open
A common complication that can arise with analyses of high-dimensional data is the repeated use of hypothesis tests. A second complication, especially with small samples, is the reliance on asymptotic p -values. Our proposed approach for a…
View article: Counternull sets in randomized experiments
Counternull sets in randomized experiments Open
Consider a statistical analysis for a randomized experiment that draws inferences based on hypothesis testing. In such settings, the plausibility of a null hypothesis is often examined using a p-value associated with a test statistic. In c…
View article: Catalytic Priors: Using Synthetic Data to Specify Prior Distributions in Bayesian Analysis
Catalytic Priors: Using Synthetic Data to Specify Prior Distributions in Bayesian Analysis Open
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 …
View article: Causal inference from treatment-control studies having an additional factor with unknown assignment mechanism
Causal inference from treatment-control studies having an additional factor with unknown assignment mechanism Open
Consider a situation with two treatments, the first of which is randomized but the second is not, and the multifactor version of this. Interest is in treatment effects, defined using standard factorial notation. We define estimators for th…
View article: Estimating adjusted risk differences by multiply‐imputing missing control binary potential outcomes following propensity score‐matching
Estimating adjusted risk differences by multiply‐imputing missing control binary potential outcomes following propensity score‐matching Open
We describe a new method to combine propensity‐score matching with regression adjustment in treatment‐control studies when outcomes are binary by multiply imputing potential outcomes under control for the matched treated subjects. This ena…
View article: On Optimal Rerandomization Designs
On Optimal Rerandomization Designs Open
Blocking is commonly used in randomized experiments to increase efficiency of estimation. A generalization of blocking removes allocations with imbalance in covariate distributions between treated and control units, and then randomizes wit…
View article: Automatic detection of influential actors in disinformation networks
Automatic detection of influential actors in disinformation networks Open
Significance Hostile influence operations (IOs) that weaponize digital communications and social media pose a rising threat to open democracies. This paper presents a system framework to automate detection of disinformation narratives, net…
View article: Contrast-specific propensity scores
Contrast-specific propensity scores Open
Basic propensity score methodology is designed to balance the distributions of multivariate pre-treatment covariates when comparing one active treatment with one control treatment. However, practical settings often involve comparing more t…
View article: The importance of having a conceptual stage when reporting non-randomized studies
The importance of having a conceptual stage when reporting non-randomized studies Open
Formal guidelines for statistical reporting of non-randomized studies are important for journals that publish results of such studies. Although it is gratifying to see some journals providing guidelines for statistical reporting, we feel t…
View article: Influence-Disinformation-Networks/PNAS-Narrative-Networks: PNAS Narrative Networks Release
Influence-Disinformation-Networks/PNAS-Narrative-Networks: PNAS Narrative Networks Release Open
This repository contains additional data used for the paper Automatic detection of influential actors in disinformation networks, Proc. Natl. Acad. Sci. U.S.A., to appear, doi:10.1073/pnas.2011216118.
View article: Influence-Disinformation-Networks/PNAS-Narrative-Networks: PNAS Narrative Networks Release
Influence-Disinformation-Networks/PNAS-Narrative-Networks: PNAS Narrative Networks Release Open
This repository contains additional data used for the paper Automatic detection of influential actors in disinformation networks, Proc. Natl. Acad. Sci. U.S.A., to appear, doi:10.1073/pnas.2011216118.
View article: Influence-Disinformation-Networks/PNAS-Narrative-Networks: PNAS Narrative Networks Initial Release
Influence-Disinformation-Networks/PNAS-Narrative-Networks: PNAS Narrative Networks Initial Release Open
This repository contains additional data used for the paper Automatic detection of influential actors in disinformation networks, Proc. Natl. Acad. Sci. U.S.A., to appear, doi:10.1073/pnas.2011216118.
View article: When possible, report exact p-values and display informative Fisherian null randomization distributions
When possible, report exact p-values and display informative Fisherian null randomization distributions Open
In randomized experiments, Fisherian exact p-values are available and should be used to help evaluate results rather than the more commonly reported asymptotic p-values. The Fisherian statistical framework, proposed in 1925, calculates a p…
View article: Heterogeneous ozone effects on the DNA methylome of bronchial cells observed in a crossover study
Heterogeneous ozone effects on the DNA methylome of bronchial cells observed in a crossover study Open
We used a randomized crossover experiment to estimate the effects of ozone (vs. clean air) exposure on genome-wide DNA methylation of target bronchial epithelial cells, using 17 volunteers, each randomly exposed on two separated occasions …
View article: Nonstandard conditionally specified models for nonignorable missing data
Nonstandard conditionally specified models for nonignorable missing data Open
Significance We consider data-analysis settings where data are missing not at random. In these cases, the two basic modeling approaches are 1) pattern-mixture models, with separate distributions for missing data and observed data, and 2) s…
View article: When possible, report a Fisher-exact<i>P</i>value and display its underlying null randomization distribution
When possible, report a Fisher-exact<i>P</i>value and display its underlying null randomization distribution Open
Significance Statistical analyses of randomized experiments often rely on asymptotic P values instead of using the actual randomization procedure that led to the observed data. Fisher-exact and asymptotic P values can differ dramatically: …
View article: Catalytic prior distributions with application to generalized linear models
Catalytic prior distributions with application to generalized linear models Open
Significance We propose a strategy for building prior distributions that stabilize the estimation of complex “working models” when sample sizes are too small for standard statistical analysis. The stabilization is achieved by supplementing…
View article: Rerandomization in $2^{K}$ factorial experiments
Rerandomization in $2^{K}$ factorial experiments Open
With many pretreatment covariates and treatment factors, the classical factorial experiment often fails to balance covariates across multiple factorial effects simultaneously. Therefore, it is intuitive to restrict the randomization of the…
View article: Diagnosing missing always at random in multivariate data
Diagnosing missing always at random in multivariate data Open
Summary Models for analysing multivariate datasets with missing values require strong, often unassessable, assumptions. The most common of these is that the mechanism that created the missing data is ignorable, which is a two-fold assumpti…
View article: Subject Index
Subject Index Open
No Abstract
View article: Author Index
Author Index Open
No Abstract