Lies, Damned Lies, and Statistics

I wonder if Alex Tabarrok is aware of just how prevalent small sample size statistics is in medical research? Or that it was originally developed for use in analyzing beer?

I think there’s a great book to be written on all of the things that Guinness has given the world. I’d certainly be willing to do the research.

2 comments… add one
  • Probably, note how he mentions that there is a bias towards publishing statistically significant results and that it is very easy to subset samples…..

    Ideally, even an insignificant result that is for something interesting should be published since it lets others know of the negative finding and work is not needlessly duplicated.

  • Dear Mr. Schuler, I think you’d be ill-advised to undertake that research alone. I would be willing to offer my assistance for a couple of hours a day, say between 5:30 and 7:30 pm. Best regards,

Leave a Comment