Title: Advancing statistics reform: Counting and causation before probability and inference. 

Abstract: Despite decades of reform efforts by hundreds of scientists and statisticians, routine teaching and practice of statistics remains mired in damaging yet editorially enforced norms of “significance”, “confidence”, and excessive focus on null hypotheses. Some of this abuse can be addressed by end users if they are allowed to replace misleading traditional jargon with accurate ordinary-language descriptions of conventional inferential statistics, and by shifting emphasis to compatibility and information measurement. Nonetheless, there is strong resistance to reform at high levels, which arguably stems from the poor foundations in human cognition and judgment that characterize even doctoral level statistical training. As has often been lamented, that training overemphasizes mathematical idealizations and computation, while omitting or trivializing the logical connections between algorithms and the ordinary-language descriptions of reality that all applications must employ. Much of “philosophy of statistics” has only aggravated this problem, focusing on abstract principles as if they were laws of thought rather than mundane guides for data analysis. Yet, essential bridges to reality can be systematized and taught to good effect. Doing so illustrates how Bayesian and frequentist methods are complementary tools, not conflicting philosophies, and how both can be built on basics of counting (as in data description), causal logic (as in mechanism description), and valuations. Combination of these elements is essential for sound scientific (as opposed to merely statistical) inference and reveals that many statistical truisms treated as deep facts (such as that a causal null hypothesis is “parsimonious”) are limited heuristics that become fallacious and misleading in many contexts. 

Biography: Dr Sander Greenland is Emeritus Professor of Epidemiology and Statistics at the UCLA Fielding School. He is a leading contributor to epidemiologic statistics, theory, and methods. His research focuses on the limitations and misuse of statistical methods in observational studies. He has authored or co-authored more than 400 articles and book chapters in epidemiology and statistics and co-authored the textbook Modern Epidemiology. Dr Greenland has served on the editorial boards of statistics and epidemiology journals, and as an advisor to numerous agencies including the Food and Drug Administration, the Environmental Protection Agency, the Centers for Disease Control and Prevention, the State of California, the National Academy of Sciences, and the World Health Organization. He is a Fellow of the American Statistical Association and the Royal Statistical Society and has been an invited speaker at universities and conferences in statistics and epidemiology throughout the world. He has received honors from the Albert Einstein Medical College, the University of Aarhus, and the Food and Drug Administration for his work on statistical and epidemiologic methods for drug and device safety evaluation. Dr Greenland received his MPH and DrPH in epidemiology from the UCLA Fielding School. 

All welcome Zoom link