Writing equations for latent growth curve models

[omnibus]

Hi all,

I'm a PhD student currently working on an independent research project using latent growth curve modeling. This is my first time using this statistic, and I was wondering if anyone had a good reference that discusses how to form the statistical (i.e., mathematical) equations. I like to grasp what I'm doing before "plugging and chugging' so to speak.

Specifically, I'm investigating factors that influence change in polydrug use over three waves of data in a sample of urban Hispanic/Latino youth.

Also, if anyone knows of a reference for equation writing with parallel process models, my overly ambitious colleague would be appreciative.

Thank you for your assistance!

 

[spunky]

how to form the statistical (i.e., mathematical) equations. I like to grasp what I'm doing before "plugging and chugging' so to speak.

Muthen's website's always a good source to find the why-things-work for everything:

http://pages.gseis.ucla.edu/faculty/muthen/articles/Article_071.pdf

although i'm confused... using the Bentler-Weeks notation (and assuming you're already familiar with the basics of SEM), how would the structural equations be any different from the regular ones, just by choosing a particular set of loadings for the slope and intercept?

 

[omnibus]

Thank you for the paper. I'm familiar with the basics of SEM, but I've only heard of latent mixture models in formal settings where it is assumed the audience is familiar with the technique. While presenters often show an "SEM" looking diagram, they seldom discuss its relation to SEM. At any rate, thanks again for the reference.

 

[spunky]

wait a minute...


first you said:

working on an independent research project using latent growth curve modeling

now you're saying:

but I've only heard of latent mixture models in

latent growth curve models and latent mixture models are not the same thing. the reference i provided you with is only for latent growth curves. although, just to make things interesting, there is such a thing as LATENT GROWTH MIXTURE models... so... which ones are you intersted in? latent growth models? latent mixture models? or latent growth mixture models? they are related but are not the same thing.

 

[omnibus]

You're right. It was a typo on my end. I am interested in the latter--latent growth mixture modeling. The Muthen site you recommended is very useful and has several papers on the topic I am reviewing. Thanks again and sorry for the miscommunication.

 

[spunky]

yah, the Mplus website has everything on everything to be known, lol... but i just wanted to clarify that because maybe you were gonna go to the link i sent to you and realize very quickly there isn't much "mixture modeling" (but a hell lot of "latent growth curve") going on there...