Re: st: bootstrap with mixed logit

[Stas Kolenikov <skolenik@gmail.com>]

I would not expect this work, as the class labels will always be
different in different bootstrap samples: what is class 1 in the
original sample may be class 3 in the first bootstrap subsample, and
class 4 in the second one, etc. With a single run, a permutation of
class labels is irrelevant, but matching the classes over repeated
runs, as is needed for the bootstrap, is extremely cumbersome.
Moreover, some smaller classes may disappear if they don't have enough
observations in a given bootstrap sub-sample. So I don't think the
bootstrap is a sensible method for this problem.

On Fri, Jan 27, 2012 at 8:42 AM, Marco N. Anterpi <anterpin@hotmail.it> wrote:
> Hello everyone,
>
> I have an unbalanced panel with a binary dependent
> variable and I estimated a mixed logit model. However, I do not manage
> to compute the standard errors for the summary statistics with the
> command bootstrap.
>
> More in detail, I estimated a mixed logit model with 5 latent classes via EM algorithm,
> using the command:
>  lclogit choice price income gender, group(ind) id(dec) nclasses(5)
>
> I can display the list of parameters
> (matrix B) and the probability of class membership (matrix P) and 5
> variance-covariance matrices of the estimators ( v1 v2 v3 v4 v5 one for each class).
>
> My problem is that I do not manage to produce the standard errors
> for the summary statistics of the different variables and I am probably
> making a silly mistake in handling the data.
>
> Can anyone tell me the exact steps I should follow?

-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/