@JeffGreeneLearn Harring and I have a couple papers on this marginal GMM approach and have found that it vastly improves convergence and is more accurate compared to a random effects GEE, even if the model is wrong. 2/ https://t.co/TJQBywkFEs https://t.c
RT @dmcneish18: New paper - Continuous latent variables in Growth Mixtures make estimation hard & lead to nonconvergence. The discrete late…
RT @dmcneish18: New paper - Continuous latent variables in Growth Mixtures make estimation hard & lead to nonconvergence. The discrete late…
RT @dmcneish18: New paper - Continuous latent variables in Growth Mixtures make estimation hard & lead to nonconvergence. The discrete late…
Excellent paper. It's clear, pedagogical, has a pragmatic focus, and is surfeit with appendices and OSF links. It just so happens I’m in the early stages of an applied GMM paper. No longer. I’m switching my approach to Dan's CPMM.
Great work by alum @dmcneish18 and EDMS prof @Dr_J_Rob!
RT @dmcneish18: New paper - Continuous latent variables in Growth Mixtures make estimation hard & lead to nonconvergence. The discrete late…
RT @dmcneish18: New paper - Continuous latent variables in Growth Mixtures make estimation hard & lead to nonconvergence. The discrete late…
RT @dmcneish18: New paper - Continuous latent variables in Growth Mixtures make estimation hard & lead to nonconvergence. The discrete late…
RT @dmcneish18: New paper - Continuous latent variables in Growth Mixtures make estimation hard & lead to nonconvergence. The discrete late…
RT @dmcneish18: New paper - Continuous latent variables in Growth Mixtures make estimation hard & lead to nonconvergence. The discrete late…
This is a very solid option for situations where you intend to choose growth mixture modeling. Take a look.
New paper - Continuous latent variables in Growth Mixtures make estimation hard & lead to nonconvergence. The discrete latent is the main interest, so the cont. latent growth parms just pattern the covariance and are unnecessary to this main goal. 1
Covariance pattern mixture models: Eliminating random effects to improve convergence and performance https://t.co/oaBQaIfDUv BehResM