Re: Negative Adjusted R Square is a "good" thing?

Posted by David Marso on Feb 11, 2014; 5:35pm
URL: http://spssx-discussion.165.s1.nabble.com/Negative-Adjusted-R-Square-is-a-good-thing-tp5724399p5724447.html

Maybe you should review your data and the method(s) you used to form composites?
Inspect the correlation matrix and go from there...
Maybe there is a data error.  Maybe one that can be corrected?
---

cynicalflyer wrote

Thanks to you all. I reduced the 37 indicators down to 5 based on literature + my judgment and ran the regressions; I still get all negative Adjusted R Squares vs. composite scores and vs. salient items. My methodologist has told me that the committee will likely rescind it prior approval of the proposal and that the only reason for getting a negative adjusted R square is data error (read: I screwed up). A proquest search of all dissertations finds only like 12 that had negative adjusted R squares. Have to start over.