Article
Statistics & Probability
Jared Wolf, Hong Zhou
Summary: The proposed data-driven fallback procedure combines the strengths of data-driven multiple comparison procedures and predetermined strategies, controlling the familywise error rate and demonstrating more power than traditional methods in many cases.
COMMUNICATIONS IN STATISTICS-THEORY AND METHODS
(2021)
Article
Statistics & Probability
Huajiang Li, Hong Zhou
Summary: The novel approach proposed in this paper, based on the covering principle, can strongly control the familywise error rate. Computer simulations indicate that this new method rejects more hypotheses on primary endpoints compared to traditional approaches in most scenarios.
COMMUNICATIONS IN STATISTICS-THEORY AND METHODS
(2021)
Article
Biochemistry & Molecular Biology
Shin June Kim, Youngjae Oh, Jaesik Jeong
Summary: With advancements in technology, analyzing complex and large-scale data requires more advanced techniques. In comparing different methods for analyzing omics data from two different groups, controlling error rates such as false discovery rate is crucial. Three methods were selected for comparison study, with Efron's approach being one-dimensional and Ploner's and Kim's approaches being two-dimensional. Variants of Ploner's approach were also considered in the performance comparison on simulated and real data.
Article
Statistics & Probability
Dennis Boos, Kaiyuan Duan
Summary: Friedman's rank test and the associated aligned rank test are standard rank alternatives for the RCBD, while the current practice lacks good rank alternatives for the Tukey-Kramer procedure in the normal linear model. However, the closed method applied to aligned ranks shows strong control and good power compared to the Tukey-Kramer method for long-tailed error distributions.
COMMUNICATIONS IN STATISTICS-THEORY AND METHODS
(2022)
Article
Statistics & Probability
Brice Ozenne, Esben Budtz-Jorgensen, Sebastian Elgaard Ebert
Summary: This article introduces the use of the max-test procedure in latent variable models for multiple testing. The effectiveness of the procedure for Wald and Score tests in LVMs is evaluated through simulation studies. The procedure is then applied to quantify the neuroinflammatory response to mild traumatic brain injury in multiple brain regions.
COMPUTATIONAL STATISTICS
(2023)
Article
Biology
Michael A. Proschan, Dean A. Follmann
Summary: This article discusses the problem of controlling the type I error rate in a clinical trial using a gatekeeping approach, and finds that the naive method is unable to effectively control the familywise error rate. It is a common misconception that the closure principle can be used to prove the effectiveness of the naive procedure.
Article
Biology
Huijuan Zhou, Xianyang Zhang, Jun Chen
Summary: A novel covariate-adaptive procedure is proposed to control the familywise error rate, incorporating external covariates, and numerical studies show its robustness across different settings.
Article
Statistics & Probability
Dennis D. Boos, Siyu Duan
Summary: The Wilcoxon rank sum test and the Kruskal-Wallis rank test are robust alternatives to t-tests and F-tests for non-normal data with long tails. The closed method of Marcus et al. applied to ranks is powerful for both small and large samples, and better than other nonparametric methods. Additionally, the closed method applied to means is even more powerful than the Tukey-Kramer method, especially for non-normal data with moderately long tails and small samples.
AMERICAN STATISTICIAN
(2021)
Article
Statistics & Probability
Jelle J. Goeman, Jesse Hemerik, Aldo Solari
Summary: The passage discusses the class of multiple testing methods controlling tail probabilities of the false discovery proportion, showing that all such methods are either equivalent to a closed testing procedure or can be uniformly improved by one. It also demonstrates the practical usefulness of focusing on closed testing methods when designing methods.
ANNALS OF STATISTICS
(2021)
Article
Mathematical & Computational Biology
Ajit C. Tamhane, Dong Xi
Summary: This article proposes a simulation-based method for computing the multiplicity adjusted p-values and critical constants for the Dunnett procedure under heteroskedasticity. The method is shown to accurately control the familywise error rate (FWER) through simulation experiments.
BIOMETRICAL JOURNAL
(2023)
Article
Mathematics
Narayanaswamy Balakrishnan, Ghobad Barmalzan, Sajad Kosari
Summary: This paper investigates the stochastic comparisons of parallel systems with PRHR distributed components and starting devices. By considering specific examples, the hazard rate and reversed hazard rate orders of the systems are proven.
Article
Mathematical & Computational Biology
Zheng Wang, Yu Cheng, Eric C. Seaberg, Leah H. Rubin, Andrew J. Levine, James T. Becker
Summary: Motivated by the Multicenter AIDS Cohort Study (MACS), this study developed classification procedures for cognitive impairment based on longitudinal measures, adapting the cross-sectional multivariate normative comparisons (MNC) method to control family-wise error. The study proposed longitudinal MNC procedures based on multivariate mixed effects models to effectively control family-wise error at a predetermined level, using a dataset from a neuropsychological substudy of the MACS to illustrate the applications of the proposed classification procedures.
STATISTICS IN MEDICINE
(2021)
Article
Biology
Ningning Xu, Aldo Solari, Jelle J. Goeman
Summary: In this paper, a multiple testing method based on closed testing is proposed for the Globaltest, which controls the familywise error rate and allows post hoc inference. Additionally, a novel shortcut is derived to circumvent the computation time issue of closed testing.
Article
Biology
Willi Maurer, Frank Bretz, Xiaolei Xun
Summary: This article discusses the problem of testing multiple null hypotheses in a real-life context where different losses may occur due to incorrect decisions. The authors argue that traditional methods for controlling Type I error rate may be too restrictive and suggest using loss functions and decision-theoretic approach to control familywise expected loss instead. They illustrate their methods using the problem of establishing efficacy of a new medicinal treatment in non-overlapping subgroups of patients.
Article
Economics
John A. List, Azeem M. Shaikh, Atom Vayalinkal
Summary: This paper offers a framework for simultaneously testing multiple null hypotheses using experimental data with simple random sampling for assigning treatment status. By utilizing general results from the multiple testing literature, the paper develops a procedure that asymptotically controls the familywise error rate and is asymptotically balanced, leading to gains in power. It also incorporates observed baseline covariates to further improve power, based on recent results from the statistics literature.
JOURNAL OF APPLIED ECONOMETRICS
(2023)