Last 3 Days
Quantifying the effect of a continuou...
Structural Equation Modeling
posted on Sunday, February 10, 2008 - 2:21 am
I have no success finding an option I believed Mplus had: the ability to estimate and quantify the effect of a continuous moderator on a regression (or a correlation)
In LCA, Mplus can estimate the effect of latent classes on associations between two variables. I would like to have the possibility to estimate moderation for a non-categorical variable, having
(1) a quantification of its effect on the association between two variables
(2) being able to develop a model with longitudinal data and three-wave data where the influence of the moderator is estimated at all three measurements within one model.
Is that possible. If so, how do I proceed?
Thanks for any help.
posted on Sunday, February 10, 2008 - 4:24 am
I might add... I know of the option to estimate interaction effects even between two latents (XWITH).
That is an alternative. But it is not quite what I was hoping for when testing the moderating effect of a continuous variable.
Bengt O. Muthen
posted on Monday, February 11, 2008 - 7:55 am
Ex5.23 uses the continuous variable "pihat" to moderate a correlation in a "QTL" analysis.
Moderating a regression slope gets back to the usual interactions between variables.
Back to top
Add Your Message Here
This is a private posting area. Only registered users and moderators may post messages here.
Enable HTML code in message
Automatically activate URLs in message