Tuesday, February 21, 2017

Pet Statistical Peeve

Just a brief bitching about a line I just read in a published study, about a difference being "nominally statistically significant."

Sorry, a difference or relationship is either significant or it's not. There's is no in-between. That's like saying: "I'm nominally pregnant." You are, or aren't.

Even better, the difference they're promoting? .40 compared to .42. Really? Sheesh. Not only is it not statistically significant, more importantly it's not substantively different.


No comments: