getting beyond…

IV. (June 11) Rejection Fallacies: Do P-values exaggerate evidence? Jeffreys-Lindley paradox or Bayes/Fisher disagreement:

Reading:

SIST: Excursion 4 Tour II

Recommended (if time): Excursion 4 Tour I: The Myth of “The Myth of Objectivity” 


Mayo Memos for Meeting 4

–Souvenirs  Meeting 4: Q: Have We Drifted From Testing Country? (Notes From an Intermission); R: The Severity Interpretation of Rejection (SIR)

FUN! Take a look at Richard Morey’s newly updated SEV app. It will display P-values, power and SEV (click display options). You can change the default by clicking the tab details and then using that link. Don’t forget to change the range of parameter values. If you change n to 25, you’ll get the answers to the example I gave in meeting #2.

  1. Solutions to problems given in Meeting #2: With X̅ =154 (PDF); with X̅ = 152 (PDF)
  2. Using the app for simple P-values: I wasn’t able to use the board to draw the curves for different P-values in meeting #2. Here’s how you can view them using Morey’s app for simple P-values. 

How do you interpret it? This just came out in NEJM (in defending policies based on antibody tests). “In the world of randomized clinical trials, statisticians test scientific hypotheses by requiring a probability of less than 5% that the observed result could have occurred by chance.” (Waiting for Certainty on Covid-19 Antibody Tests — At What Cost?)  https://www.nejm.org/doi/full/10.1056/NEJMp2017739?source=nejmtwitter&medium=organic-social

-See details on Bonus Meeting: June 25.


Slides & Video Links for Meeting

Slides: (PDF)