As reports of recalled medications and contaminated consumer products continue to mount, people worldwide are increasingly taking issue with outcomes that rely on “statistical significance” testing.
Few understand what is wrong with statistical significance, nor are as well versed in what can best replace the flawed practice, as Roosevelt University Economist and Professor Stephen T. Ziliak.
One of Roosevelt’s leading faculty experts, and among the most accomplished in the nation on the timely topic of p-values and other tests of statistical significance, Ziliak is available for interviews at email@example.com or 312-341-3763.
A pioneer who has questioned the scientific and policy relevance of statistical testing since 1996, Ziliak is co-editor of a new 43 article volume (one of them, his own) published as a special issue of The American Statistician on “Statistical Inference in the 21st Century: A World Beyond p<0.05.”
Lead author of the critically acclaimed book, The Cult of Statistical Significance (2008), Ziliak is credited with starting the anti-statistical significance testing movement in the first place.
In recent years, this movement has been growing in voice, including at the U.S. Supreme Court in Matrixx v. Siracusano (2011), regarding the role of statistical significance to securities law and adverse event reports on biomedical supplies.
In his new article, Ziliak outlines for the first time “10 principles” of good statistical inference, which he discovered in archives related to the production and quality control of Guinness beer. The principles were used by William Sealy Gosset (1876-1937), aka “Student,” in Gosset’s job as Head Experimental Brewer at the Guinness Brewery in Dublin.
Ziliak’s article is entitled: “How Large Are Your G-Values? Try Gosset’s Guinnessometrics When a Little “p” is Not Enough.”