|
|
Estimation of non-linear effects |
|
Message/Author |
|
|
With an old (ca 2011) computer, I used ANES data to model the responses to the two presidential candidates in each of the 2004 - 2016 elections: Each model contains 4 latent variables (2 per candidate). I currently estimate the quadratic effects of 2 and the product of the other 2 on the vote by saving the results of an initial model and treating the latents as observed variables in subsequent models. Could current personal computers provide enough power and memory to estimate these effects in the initial model using the xwith command? If so, what would be the specs.? |
|
|
Yes, current computers can handle that, particularly when using our new Bayes features; see the paper on our website: Asparouhov, T. & Muthén, B. (2019). Bayesian estimation of single and multilevel models with latent variable interactions. Submitted for publication. You can do it with a Windows i7 or i9 cpu. |
|
Back to top |
|
|