Thank you for you helpful video. Can we use a 3 step approach to investigate class membership if the variables are time-varying (I don't want to use baseline variables)?
Thank you so much for this. i wonder if one can use a categorical variable as a distal outcome (dependent variable) and include the covariates and the posterior class membership in a multinomial regression? What assumption is it likely to break?
Thanks for sharing your lesson Dr Vermunt. Are you aware of the implementation of BCH in R? I did my step 1 with mClust but now I am not able to do anything else than the standard (non-adjusted) 3step process.
In Latent GOLD, one can get classifications when running a step3 model. In the Syntax there are several options which are relevant for this (mainly "simultaneous" and "noignore"). It may be an idea to use the Bakk-Kuha method if your purpose is to improve classifications via the step3 analysis.
Great explanation! How can I change the reference class when doing multinomial regression? I am modifying the order in the syntax, but I don't know if it is correct.
Yes, the 1-step approach is also a latent class cluster analysis. The step-3 model of the 3-step approach I would not call latent class cluster analysis.
This is really helpful! Thanks for taking the time to make this video, Dr. Vermunt!
Thank you for you helpful video. Can we use a 3 step approach to investigate class membership if the variables are time-varying (I don't want to use baseline variables)?
Thank you so much for this. i wonder if one can use a categorical variable as a distal outcome (dependent variable) and include the covariates and the posterior class membership in a multinomial regression? What assumption is it likely to break?
Thanks for sharing your lesson Dr Vermunt. Are you aware of the implementation of BCH in R? I did my step 1 with mClust but now I am not able to do anything else than the standard (non-adjusted) 3step process.
As far I know this is not available in R.
Very useful guide! How should we obtain the new cluster membership after bias adjustment via step 3?
In Latent GOLD, one can get classifications when running a step3 model. In the Syntax there are several options which are relevant for this (mainly "simultaneous" and "noignore"). It may be an idea to use the Bakk-Kuha method if your purpose is to improve classifications via the step3 analysis.
Thank you for the video! Is there a way to include both covariates and distal outcome in Step3.
Yes, this is possible. In Latent GOLD this requires using the Syntax instead of the Step3 point-and-click module.
Great explanation! How can I change the reference class when doing multinomial regression? I am modifying the order in the syntax, but I don't know if it is correct.
This trick yields the correct results. In the Latent GOLD Syntax you have full flexibility to select the reference category.
Very interesting video Dr. Vermunt! Just wondering that this model with covariate also as Latent Class Cluster Analysis? Thanks!
Yes, the 1-step approach is also a latent class cluster analysis. The step-3 model of the 3-step approach I would not call latent class cluster analysis.