The ABCs of p’s and q’s: Potential Application in Clinical Safety Evaluation
主 题: The ABCs of p’s and q’s: Potential Application in Clinical Safety Evaluation
报告人: Joseph F. Heyse （Merck Research Laboratories）
时 间: 2018-06-13 14:00 - 2018-06-13 15:30
地 点: Room 77201, Jingchunyuan 78, BICMR
\n\tThe\nconcept of “statistical significance” as assessed using the p-value has a long\nhistory and is a principle mainstay of scientific experimentation.? Computations of p-values date back to the\n1770s when they were calculated by Pierre-Simon Laplace who compared birth rates\nof girls and boys.? The p-value was first\nformally introduced by Karl Pearson in 1900 with the formulation of his\nwell-known Pearson’s Chi-squared test.?\nThe use of p-values in experimental settings was popularized by Sir\nRonald A. Fisher who proposed the commonly used p < 0.05 as a threshold for\njudging statistical significance.?\nWasserstein and Lazar (2016) published a highly visible position paper\non the appropriate uses and frequent misuses of the p-value.? The misuses relate to a misunderstanding of\nthe interpretation of p-values, and issues in multiple testing.? These issues have heightened with the recent\ntechnological advances and have introduced the need for large scale testing and\nthe q-value. <\/o:p>\n<\/p>\n
\n\t?\n<\/p>\nThis\ntalk will provide a brief historical development of p-values and q-values\nincluding the breakthrough method of False Discovery Rate (FDR) control and the\nBenjamini-Hochberg procedure.? The\nq-value is defined and developed as providing a quantitative measure using FDR\ncontrol.? Two examples are discussed\nincluding HIV data and a clinical safety evaluation.