49: War and p's

Episode 49 · July 31st, 2017 · 55 mins 59 secs

About this Episode

In this episode Dan and James discuss a forthcoming paper that's causing a bit of a stir by proposing that biobehavioral scientists should use a 0.005 p-value statistical significance threshold instead of 0.05.

Stuff they cover:

  • A summary of the paper and how they decided on 0.005.
  • Whether raising the threshold the best way to improve reproducibility?
  • Is 0.005 too stringent?
  • Would this new threshold unfairly favour “super” labs?
  • If we keep shifting the number does any threshold really matter?
  • Dan and James’ first impressions of the paper
  • A crash course on Mediterranean taxation systems
  • What would a 0.005 threshold practically mean for researchers?

Links
The paper https://osf.io/mky9j/
ENIGMA consortium http://enigma.ini.usc.edu

Music credits: Lee Rosevere freemusicarchive.org/music/Lee_Rosevere/

Support Everything Hertz

Episode Comments