Propensity‐score‐based priors for Bayesian augmented control design

Drug developers are required to demonstrate substantial evidence of effectiveness through the conduct of adequate and well‐controlled (A&WC) studies to obtain marketing approval of their medicine. What constitutes A&WC is interpreted as the conduct of randomized controlled trials (RCTs). However, these trials are sometimes unfeasible because of their size, duration, and cost. One way to reduce sample size is to leverage information on the control through a prior. One consideration when forming data‐driven prior is the consistency of the external and the current data. It is essential to make this process less susceptible to choosing information that only helps improve the chances toward making an effectiveness claim. For this purpose, propensity score methods are employed for two reasons: (1) it gives the probability of a patient to be in the trial, and (2) it minimizes selection bias by pairing together treatment and control within the trial and control subjects in the external data that are similar in terms of their pretreatment characteristics. Two matching schemes based on propensity scores, estimated through generalized boosted methods, are applied to a real example with the objective of using external data to perform Bayesian augmented control in a trial where the allocation is disproportionate. The simulation results show that the data augmentation process prevents prior and data conflict and improves the precision of the estimator of the average treatment effect.     

Cross Section and Panel Data Estimators for Nonseparable Models with Endogenous Regressors

We propose two new methods for estimating models with nonseparable errors and endogenous regressors. The first method estimates a local average response. One estimates the response of the conditional mean of the dependent variable to a change in the explanatory variable while conditioning on an external variable and then undoes the conditioning. The second method estimates the nonseparable function and the joint distribution of the observable and unobservable explanatory variables. An external variable is used to impose an equality restriction, at two points of support, on the conditional distribution of the unobservable random term given the regressor and the external variable. Our methods apply to cross sections, but our lead examples involve panel data cases in which the choice of the external variable is guided by the assumption that the distribution of the unobservable variables is exchangeable in the values of the endogenous variable for members of a group.     

Extending permutation conditional inference to unconditional ones

In this presentation we discuss the extension of permutation conditional inferences to unconditional or population ones. Within the parametric approach this extension is possible when the data set is randomly selected by well-designed sampling procedures on well-defined population distributions, provided that their nuisance parameters have boundely complete statistics in the null hypothesis or are provided with invariant statistics. When these conditions fail, especially if selection-bias procedures are used for data collection processes, in general most of the parametric inferential extensions are wrong or misleading. We will see that, since they are provided with similarity and conditional unbiasedness properties and if correctly applicable, permutation tests may extend, at least in a weak sense, conditional to unconditional inferences.     

A note on the orthant probability of the elliptically contoured distribution

When a scale matrix satisfies certain conditions, the orthant probability of the elliptically contoured distribution with the scale matrix is expressed as the same probability of the equicorrelated normal distribution.     

Approximate exchangeability and de Finetti priors in 2022

This is a review paper, beginning with de Finetti's work on partial exchangeability, continuing with his approach to approximate exchangeability, and then his (surprising) approach to assigning informative priors in nonstandard situations. Recent progress on Markov chain Monte Carlo methods for drawing conclusions is supplemented by a review of work by Gerencsér and Ottolini on getting honest bounds for rates of convergence. The paper concludes with a speculative approach to combining classical asymptotics with Monte Carlo. This promises real speed-ups and makes a nice example of how theory and computation can interact.     

On statistical tests of functional connectome fingerprinting

Fingerprinting of functional connectomes is an increasingly standard measure of reproducibility in functional magnetic resonance imaging connectomics. In such studies, one attempts to match a subject's first session image with their second, in a blinded fashion, in a group of subjects measured twice. The number or percentage of correct matches is usually reported as a statistic, which is then used in permutation tests. Despite the simplicity and increasing popularity of such procedures, the soundness of the statistical tests, the power, and the factors impacting the test are unstudied. In this article, we investigate the statistical tests of matching based on exchangeability assumption in the fingerprinting analysis. We show that a nearly universal Poisson(1) approximation applies for different matching schemes. We theoretically investigate the permutation tests and explore the issue that the test is overly sensitive to uninteresting directions in the alternative hypothesis, such as clustering due to familial status or demographics. We perform a numerical study on two functional magnetic resonance imaging (fMRI) resting‐state datasets, the Human Connectome Project (HCP) and the Baltimore Longitudinal Study of Aging (BLSA). These datasets are instructive, as the HCP includes technical replications of long scans and includes monozygotic and dizygotic twins, as well as non‐twin siblings. In contrast, the BLSA study incorporates more typical length resting‐state scans in a longitudinal study. Finally, a study of single regional connections is performed on the HCP data.