Conversation

Reading a paper from 2002... First F statistic is "approaching significance " at p=.06, second is "approaching significance" again at .09, and third is still "approached significance" at p = .11 Those F statistics aren't approaching significance you guys, they're running away.

528

This Tweet was deleted by the Tweet author. Learn more

Replying to

and

Thank you. People laugh at the wrong thing. They should be laughing at the arbitrary .05 threshold.

4:58 PM · Feb 18, 2019Twitter for Android

Replying to

.05 is definitely arbitrary, but this paper was interpreting p<.001 the same as p=. 11 and using both values as enough evidence to interpret everything as meaningfully different by this made up idea of “approaching significance”

Replying to

and

Exactly. The tyranny of 0.05 (and the binary concept of “significance”) is what creates this ridiculous torture of the language. Findings aren’t “trending towards” or “approaching” significance, but reviewers cling rigidly to 0.05 = “significant result” so this is what we get

Show replies