Sarah F Ackley, Ryan M Andrews, Christopher Seaman, Michael Flanders, Ruijia Chen, Jingxuan Wang, Grissel Lopes, Kendra D Sims, Peter Buto, Erin Ferguson, Isabel Elaine Allen, M Maria Glymour
{"title":"流行病学期刊中P值分布的趋势:统计学、P曲线和模拟研究。","authors":"Sarah F Ackley, Ryan M Andrews, Christopher Seaman, Michael Flanders, Ruijia Chen, Jingxuan Wang, Grissel Lopes, Kendra D Sims, Peter Buto, Erin Ferguson, Isabel Elaine Allen, M Maria Glymour","doi":"10.1093/aje/kwaf184","DOIUrl":null,"url":null,"abstract":"<p><p>Epidemiologists have advocated for reporting confidence intervals and deemphasizing P values to address long-standing concerns about null-hypothesis statistical-significance testing, P hacking, and reproducibility. It is unknown if efforts to reduce reliance on P values have altered the distribution of P values. For 21,332 abstracts published 2000-2024 in 4 major epidemiology journals, two-sided P values (N=25,288) were calculated from estimates and confidence intervals scraped using ChatGPT's 4o model. We evaluated trends over time to determine whether the empirical distribution of P values changed. We fitted to expected P value distributions and simulated these distributions with and without assuming changes in statistical power over time. Average P values decreased from 2000 to 2024; the fraction of P values just below the 0.05 threshold also decreased. Fits to models indicate that statistical power increased. Increasing power would reduce average P value while also decreasing P values near the 0.05 threshold-precisely the trends observed in epidemiology journals. Although the frequency of P values near the 0.05 threshold has declined modestly, this likely reflects increases in statistical power rather than decreases in P hacking.</p>","PeriodicalId":7472,"journal":{"name":"American journal of epidemiology","volume":" ","pages":""},"PeriodicalIF":4.8000,"publicationDate":"2025-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Trends in the Distribution of P Values in Epidemiology Journals: A Statistical, P-Curve, and Simulation Study.\",\"authors\":\"Sarah F Ackley, Ryan M Andrews, Christopher Seaman, Michael Flanders, Ruijia Chen, Jingxuan Wang, Grissel Lopes, Kendra D Sims, Peter Buto, Erin Ferguson, Isabel Elaine Allen, M Maria Glymour\",\"doi\":\"10.1093/aje/kwaf184\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>Epidemiologists have advocated for reporting confidence intervals and deemphasizing P values to address long-standing concerns about null-hypothesis statistical-significance testing, P hacking, and reproducibility. It is unknown if efforts to reduce reliance on P values have altered the distribution of P values. For 21,332 abstracts published 2000-2024 in 4 major epidemiology journals, two-sided P values (N=25,288) were calculated from estimates and confidence intervals scraped using ChatGPT's 4o model. We evaluated trends over time to determine whether the empirical distribution of P values changed. We fitted to expected P value distributions and simulated these distributions with and without assuming changes in statistical power over time. Average P values decreased from 2000 to 2024; the fraction of P values just below the 0.05 threshold also decreased. Fits to models indicate that statistical power increased. Increasing power would reduce average P value while also decreasing P values near the 0.05 threshold-precisely the trends observed in epidemiology journals. Although the frequency of P values near the 0.05 threshold has declined modestly, this likely reflects increases in statistical power rather than decreases in P hacking.</p>\",\"PeriodicalId\":7472,\"journal\":{\"name\":\"American journal of epidemiology\",\"volume\":\" \",\"pages\":\"\"},\"PeriodicalIF\":4.8000,\"publicationDate\":\"2025-08-21\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"American journal of epidemiology\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1093/aje/kwaf184\",\"RegionNum\":2,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"American journal of epidemiology","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1093/aje/kwaf184","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH","Score":null,"Total":0}
Trends in the Distribution of P Values in Epidemiology Journals: A Statistical, P-Curve, and Simulation Study.
Epidemiologists have advocated for reporting confidence intervals and deemphasizing P values to address long-standing concerns about null-hypothesis statistical-significance testing, P hacking, and reproducibility. It is unknown if efforts to reduce reliance on P values have altered the distribution of P values. For 21,332 abstracts published 2000-2024 in 4 major epidemiology journals, two-sided P values (N=25,288) were calculated from estimates and confidence intervals scraped using ChatGPT's 4o model. We evaluated trends over time to determine whether the empirical distribution of P values changed. We fitted to expected P value distributions and simulated these distributions with and without assuming changes in statistical power over time. Average P values decreased from 2000 to 2024; the fraction of P values just below the 0.05 threshold also decreased. Fits to models indicate that statistical power increased. Increasing power would reduce average P value while also decreasing P values near the 0.05 threshold-precisely the trends observed in epidemiology journals. Although the frequency of P values near the 0.05 threshold has declined modestly, this likely reflects increases in statistical power rather than decreases in P hacking.
期刊介绍:
The American Journal of Epidemiology is the oldest and one of the premier epidemiologic journals devoted to the publication of empirical research findings, opinion pieces, and methodological developments in the field of epidemiologic research.
It is a peer-reviewed journal aimed at both fellow epidemiologists and those who use epidemiologic data, including public health workers and clinicians.