11. Latent class regression and latent class growth models
Вставка
- Опубліковано 23 лют 2021
- This video discusses latent class models which can be used when a single dependent variable is observed multiple times and a regression model is specified for this dependent variable. The latent classes are subgroups which differ in the intercept and the predictor effects of this regression model. Applications concern the analysis of repeated measures experiments with within-subject factors, regression models for two-level data sets, and growth models for longitudinal data sets. The latter application type is referred to as LC growth, latent trajectory, or group-based trajectory modeling. I use Latent GOLD 6.0 for illustrating how a LC regression and a LC growth analysis is performed.
I had been searching for something to understand Latent Class Regression in a simpler way. My search ended here. Thank YOU!!
Thank you so much Prof. Vermunt for this informative abd very helpful video.
Dear Jeroen, if you are planning to record other videos could you do something about Latent Class Tree, please? Would be appreciated! Thank you anyway for your tutorials!
Yes, I have plans to create some additional videos. Topics I have in mind are LC tree models, latent Markov/transition models, and multilevel LC models.
Hello Prof. Vermunt.
Is it possible to conduct LPA of repeated measures data in LatentGold, instead of either of these two approaches for repeated data? What I've seen described occasionally as RMLPA.
Thanks :)
De term RMLPA is new to me, but it seems it simply involves treating the repeated measures as separate indicators in a standard LPA. Such a model could also be seen as a LC growth model with an unrestricted time trend (that is, using time as a nominal predictor).
Thanks so much for your quick reply.
This term was new to me also, and I too interpreted it as using repeated measures as separate indicators in a standard LPA (similar to RMLCA; Collins & Lanza, 2010).
Would that be incorrect given the indicators are not independent? I am asking because I have a measure of anxiety collected at five time points and would like to know if I could use these in an LPA, specifically in LatentGOLD?
Thanks :)
Yes, you can use these repeated measurements as separate indicators. Note that indicators are typically not independent. The dependence between indicators is used (and needed) to identify the classes.
Hi Jess! I am writing my master thesis and I also have longitudinal data on psychopathologies that I want to analyze with latent gold.. i saw your comment and found the idea very interesting. Could I possibly ask you about your project?
@@linnearitter5724 Hello :)
I did analyse this data in Latent gold - profile analysis - highly recommend it!
thank you so much