P-values can be tricky things, and if we’re not careful, we can report false positives much more frequently than 5% of the time. Here, I show how that can happen, and then I show to avoid it by doing power calculations - with an emphasis on the concepts behind the equations.

Subscribe: https://www.youtube.com/c/joshstarmer/featured

#p-hacking #machine-learning

P-Hacking and Power Calculations
2.50 GEEK