Something for Everyone in May Issue of TAS
Joshua M. Tebbs, The American Statistician Editor
The May 2021 issue of The American Statistician (TAS) is available online and features 13 articles across different sections of the journal. One of the benefits of ASA membership is free access to the online issues of TAS.
The General section has seven articles. The first examines invariance properties and sufficiency for point estimation in general parametric models. The second provides a critical assessment of the use of the slugging percentage metric in baseball. The third article discusses confidence intervals for the exceedance probability as alternatives to confidence intervals for parameters and p-values. The fourth uses the LASSO to solve the problem of “distributive justice” among creditors in the estate division problem. The fifth treats the problem of estimating a large number of parallel effects in genome-wide association studies and related experiments. The sixth article proposes a Bayesian approach to test for ordering constraints among variances with dependent outcomes. Finally, the seventh article describes intuitive graphical methods to visualize influence functions.
The Interdisciplinary section has one article, which examines 30-day readmission rates in American hospitals, demonstrating patients who receive home health care actually have a higher probability of readmission. Using this as motivation, the authors describe a counterfactual probability metric that leads to necessary and sufficient conditions for sign reversals to occur.
The Statistical Computing and Graphics section also has one article. This article combines two existing recursive Bayesian inference methods to offer computational improvements to estimate any statistical model. This can streamline Bayesian data analysis in a variety of applications involving big data, streaming data, and optimal adaptive design.
There are two articles in the Statistical Practice section. The first revisits the long debate in experimental design between classic randomization and deterministic assignments based on notions of optimality. The authors then seek to provide practical advice about modified designs that harmonize both views. The second article considers the popular NCAA basketball tournament and investigates the infamous 5–12 seeded matchup in which “upsets” commonly occur. The authors offer an analysis of this using simple statistical learning models and existing tests for orderings among multinomial probabilities.
Finally, the Teacher’s Corner section has two articles. The first shows how calculus and probability techniques can be used to obtain Gini’s index and other statistical measures associated with the Lorenz curve, which is commonly used to describe income disparities. The second article examines rank statistics and presents new technical results on the correlations between variables and ranks.
For more information , visit The American Statistician online.