(1) Interpreting R-square (2) control... PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Gavin Elder posted on Saturday, November 13, 2010 - 7:01 pm
Hi there, I have two questions regarding a path model we are getting set to finish up. The model consists of 1 DV, 1 IV, and 4 mediators (through three indirect paths).

The first question regards the R-square output provided. The output provides the r-square for all of the variables in the model which have other variable(s) predicitng them (i.e., all except the DV). (1) is it correct to interpret the r-square coefficient as .176 = 17.6% of the variance explained in __ variable using this model? and (2) we are interested in the r-square in the IV accounted for by each of the explaining variables (IV, and each Mediator) - is this possible to obtain? or can we only see the total amount explained by the model itself?

My second question regards the use of control variables (i.e., does our model account for the DV, above and beyond the variance explained by a certain set of variables -- e.g., controlling for the effect of gender on blood pressure). In this model, would I include a single path "dv ON control"? or, would I include the control variable in each of the paths predicting the outcome: "dv ON control iv"?

Your help is greatly appreciated!
Thanks
 Bengt O. Muthen posted on Saturday, November 13, 2010 - 9:59 pm
(1) Yes.

(2) I think you meant to say DV and not IV here - no, you can't divide up the variance explained because the explaining variables are correlated.

Each of the paths.
 Gavin Elder posted on Sunday, November 14, 2010 - 5:31 pm
Dr. Muthen, thank you much!
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: