A. Most cited papers:
A.1. 2020
An updated list of “Ranking of top articles for applied biostatisticians, 1985-2002” which appears as Table 3 in Schell MJ. Identifying key statistical papers from 1985 to 2002 using citation data for applied biostatisticians. The American Statistician 2010; 64(4):310-317
“The Most-Cited Statistical Papers” list will be updated in collaboration with Drs. Thomas Ryan and William Woodall, expanding upon the 25 articles identified in the paper “The Most-Cited Statistical Papers” by TP Ryan and WH Woodall in the Journal of Applied Statistics, 32(5):461-474, 2005. Aug 5, 2011.
Most Cited List in an Excel file ( 8/5/2011): Their list has been expanded to all statistical methods papers we know of with 2000 or more citations.
A.2. 2025
– M Schell and J-H Lee, Identification and Classification of Highly-Cited Statistical Papers, Statistical Science 2025 (Nov 2025 Accepted).
Abstract
Having a strong classification system for statistical papers would help applied statisticians to easily identify and stay updated on relevant research. We have developed a 7-level classification system, with the first four levels hierarchically describing the paper subject content, and three additional levels highlighting specific paper features. In developing this system, we identified 4,492 “highly cited” statistical papers from 147 peer-reviewed journals, including all 125 classified as “Statistics Probability” in the ISI Web of Science. For 58 “mainline” statistical journals, these papers have ≥ 100 citations in the ISI Web of Science and ≥ 40 citations in 2018-2020 (108 additional papers from a 2020-2022 update). For the other journals, additional criteria were applied to cull out the statistically-focused papers; these rules required some non-automated processing, and thus were not extended to every journal in ISI Web of Science.
Count-based measures are provided to assess each paper’s impact, trajectory, diffusion, and primary application area, including an estimated percentage of citation counts reflecting “applied” use. These measures aid in conducting targeted searches for papers of interest, and they are particularly valuable to applied statisticians seeking to enhance their statistical practice. Methodological statisticians can also benefit from this non-exhaustive database of highly impactful statistical papers, available online at statpapers.moffitt.org. We hope that this article provides a comprehensive overview of significant contributions in the field of statistics and data science, helping statisticians stay informed about the latest advancements and trends in their areas of expertise.
Keywords: Statistical papers; Paper classification; Highly cited papers
– Database and Its Utilization Guideline
B. Useful practices we believe
B.1. Ten Analytical Steps that Every Statistical Analyst Should Do! The two most critical preliminary data analysis tasks (Schell, 2023 Lecture Note): see a copy of the slides
B.2. Logistic regression model-Firth correction: Greenland, Mansournia, Altman (2017)
B.3. For a 2 x 2 contingency table, use Barnard’s Test (also known as the unconditional test): Draft for publication (2024)
B.4. For R x C ordinal-ordinal contingency tables, use the Mantel-Haenszel Test and Spearman’s R: Draft for publication.
B.5. For the 2-sample location problem, use t-test for small n and normal scores or Wilcoxon rank sum test for large n: Draft for publication.
C. Noteworthy opinion papers: philosophical and practical insights.
c.1. The scandal of poor medical research, D G Altman, 1994. Link