I'm trying to calculate the total R-squared in a structural equation model. Basically, I want to compare models based on the amount of variance accounted for by the endogenous variables in general (I'm not necessarily interested in the specific R-squared for each variable though).

I've searched for some options, but I'm still unsure of what to do. I'm open to doing this in Mplus, but I'm looking for a way to do this with lavaan in R.

  • $\begingroup$ Do you mean the total R-squared for the model, or the total R-squared for each variable. (I'm not sure what the first of those would mean). $\endgroup$ Mar 29 '16 at 16:50
  • $\begingroup$ See ?fitMeasures and the option fit.measures="rmsea". The MSE is intrinsically related to R squared. cran.r-project.org/web/packages/lavaan/lavaan.pdf $\endgroup$
    – AdamO
    Mar 29 '16 at 16:56
  • $\begingroup$ @Jeremy Miles I'm interested in the total R-squared. I want to compare models based on which one accounts for the most variance. The Models are all with the same participants, but refer to different contexts (e.g., emotions in math over time compared to emotions in physics over time). $\endgroup$
    – John R
    Mar 30 '16 at 14:49
  • $\begingroup$ @AdamO Do you know how to compare models based on MSE? $\endgroup$
    – John R
    Mar 30 '16 at 14:51
  • $\begingroup$ I think you need to elaborate on what you mean by total R-squared. Can you give an example? (If @AdamO's answer isn't what you want.) $\endgroup$ Mar 30 '16 at 14:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.