.-
help for ^gllapred^
.-

Predict command for gllamm6 (a first attempt)
---------------------------------------------

^gllapred^ varname [, ^xb^ ^u^ ^p^]

where only one option may be used at a time.


Description
-----------

^gllapred^ provides empirical Bayes estimates of the latent variables
or random effects from a model estimated using gllamm6.  The command 
currently only works for two-level models. Posterior probabilities can 
also be computed. If gllamm6 has estimated a model with discrete 
random effects, i.e., it was run using the ip(f) option, posterior
probabilities are computed for all the "classes". If gllamm6 has
estimated a model with (multivariate) normal random effects,
posterior probabilities are computed for the quadrature locations
used to approximate the likelihood.

The likelihood is returned and should be the same as the lieklihood
printed by gllamm6 - if not, there is a bug!


Options
--------

^xb^ the linear predictor for the fixed effects is returned

^u^ the posterior means of the random effects are computed
      and returned in variable names "varname"m1, "varname"m2,
      etc. where the m1 suffix refers to the first random effect,
      i.e. the first equation given in eqs(), etc. One posterior
      mean is returned per level-2 unit and the other observations
      contain missing values.

^p^ as well as computing the posterior means, gllapred returns the 
      posterior probabilities in "varname"1, "varname"2,
      etc. giving the probabilities of classes 1,2, etc., or in the
      case of quadrature, of quadrature points 1,2, etc.
      In the case of multiple quadrature, e.g. for two random effects,
      "varname"ij is returned giving the probability of the combination
      of location i for the first random effect and j for the second
      random effect. The returned variables have non-missing values
      for only one observation per level 2 units. gllapred also prints 
      out the (prior) probability and location matrices to help interpret
      the posterior probabilities.


Author
------
Sophia Rabe-Hesketh (spaksrh@@iop.kcl.ac.uk)