Question about the newest version for GAMLj3::gamlj_mixed argument about gamlj HOT 6 CLOSED

Please reinstall it from github: it should work now

from gamlj.

If I may, [I'm not sure this is an appropriate place for this question],

when the autocorrelation is adjusted for the longitudinal analysis

Random Effect LFT does not show any result. I am sort of guessing this is a normal finding as the residuals structure for repeated meausures is set to autoregressive... Am I thinking right ?

Thank you for the response, again..
Sincerely yours...

from gamlj.

from gamlj.

I just uploaded the .omv file that I mentioned.
example.omv.zip

from gamlj.

Yes, that is expected behavior. LR tests for random effects are not implemented when residual variance-covariance is not "Identity".

from gamlj.

Since version 3.1.6 (coming out soon), you can make a model comparison to test random variance of the intercept when using autoregressive residual variance. This can be done activating "model comparison" and specifying a nested model with the same fixed effects as the original model but no random coefficients.
However, although this is now technically possible, it does not mean that the test is sensible. There's very scarce literature supporting this particular case, so it is up to the user to justify this test.

from gamlj.