Re: st: xtmixed with autoregression--mostly equal time intervals, but a few unequal

[Lucas <lucaselastic@gmail.com>]

Dear Jordan,

I've never done this, but perhaps the cov(pattern) subcommand of mixed
(stata 13) would get you what you want.

Sam

On Tue, Jul 16, 2013 at 4:42 PM, Jordan Silberman
<silberman.stata@gmail.com> wrote:
> I'm using xtmixed to estimate a longitudinal model. Each subject was
> measured on 16 occasions, so we have 15 intervals between time points.
> Twelve of the 15 intervals between time points are all equal in
> duration (each time interval is 1 day). The other 3 are not equal
> (each of these time intervals is about 4 days). I want to estimate
> this model with autoregressive residuals (including autoregressive
> residuals significantly improves the model fit). But my understanding
> is that this is not "kosher" with unequally spaced time points. Is
> there a way to model autoregressive residuals in xtmixed, but not for
> all time points? Or are there other ways to address this problem? Any
> advice would be greatly appreciated. Thanks, Jordan
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/