Anonymous posted on Tuesday, March 22, 2005 - 7:24 pm
Hi, I have three questions on latent growth curve analysis using mplus:
1a: Is it correct to interpret the fixed estimates of the latent intercept variable (the '1')for a continous indicator as 'factor loadings'?
1b: Is there a different interpretation for 1a when conducting latent growth curve analysis for categorical data?
2: In previous analysis I found out that the latent slope of my latent grwoth curve model is insignificant when I set the parameters of time 2 and time 3 free. However, the latent slope is significant if I impose linear measurement restrictions (time 2= 1, time 3= 2). Is there any explanation for that?
3: Does Mplus produce standardized or unstandardized latent group means for the latent intercept resp. latent slope variable?
Thanks for your efforts, Tom
bmuthen posted on Tuesday, March 22, 2005 - 7:40 pm
2. I assume you refer to freeing the loadings for the slope when you talk about setting parameters free at t2 and t3 and that you say that the mean of the slope is (in)significant. I am confused, however, because with 2 growth factors you need 2 time points with fixed loadings. Regarding the slope mean, note that it has different meaning with fixed vs free loadings since change is captured by loadings*slope mean.
Anonymous posted on Tuesday, March 22, 2005 - 7:52 pm
thanks a lot, just for clarification of question 2: I have one slope factor for three measurement points. For measurement time one, I set the first path to zero. For measurement time two, I set the second to one. And for measurement time three, I set the third path to two.
bmuthen posted on Tuesday, March 22, 2005 - 7:54 pm
Ok, but what do you mean then by saying "I set the parameters of time 2 and time 3 free"?
Anonymous posted on Tuesday, March 22, 2005 - 8:32 pm
Sorry, I made a mistake. In these previous analyses I used another software and labelled the paths of time 2 and time 3 'N' respecively 'M'. In that sense, I did not "set them free". Greetings, Tom
bmuthen posted on Tuesday, March 22, 2005 - 8:43 pm
Ok; good - do you have all your questions answered then or do you want to (re)state a question?