Two-tailed test for R-square? PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Calvin D. Croy posted on Wednesday, May 20, 2009 - 1:42 pm
In Mplus 5.2 I regressed 1 continuous observed variable on 3 continuous observed predictors. No latent variables. The output lists R-square and the Two-Tailed P-value.

Is the reported p-value truly for a two-tailed test or is it really for a one-tailed test? I am trying to understand why one would ever want a two-tailed test for R-square, since the traditional R-square in OLS regression can only vary from 0 to 1. (A two-tailed test would seem to imply that one is considering the possibility that R-square is significantly less than 0.) If the R-square Mplus provides is something different from the traditional value (explained sum of squares divided by total sum of squares which is tested using 1 tail of the F distribution), could you please describe how your R-square value is calculated and why/when a 2-tailed test is appropriate? Thanks!
 Linda K. Muthen posted on Thursday, May 21, 2009 - 9:16 am
I think you are correct that the heading is not correct. The computations are as usual.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: