2

The title says it all.

In what situation R squared (non adjusted) is more useful and should be used instead of the adjusted one? Why?

Marco Rudelli
  • 550
  • 1
  • 11
  • 1
    I disagree with closing this question as 'opinion based'. There's an element of opinion, to be sure, but no more than with many other questions about statistical methods that we entertain here. And I think this is a useful question too - confusion around the purpose of adjusted R2 is fairly common. – mkt Jun 15 '21 at 09:05
  • There are many related threads about this, though I haven't yet found a totally satisfactory answer. Some relevant threads: https://stats.stackexchange.com/questions/353807/which-is-better-r-squared-or-adjusted-r-squared?rq=1 (I disagree with the top answer there), https://stats.stackexchange.com/questions/408267/justification-for-and-optimality-of-r2-adj-as-a-model-selection-criterion , https://stats.stackexchange.com/questions/147688/why-to-report-r-squared?rq=1 – mkt Jun 15 '21 at 09:17

0 Answers0