Many thanks to both Maarten and Bobby. It took my computer a couple of days to run Bobby's suggestion. But I am not sure I understand what it does? Is obs just a third level? I xtset the data on panelvar (id) and timevar; should I also somehow include obs?
I would like to also include random effects for id and age. How do I then include obs? As separate random intercept, like
> . xtmepoisson ... || id: age || obs:, ...
or a random slope?
Many thanks in advance,
Margit Averdijk
> From: [email protected]
> To: [email protected]
> Subject: Re: st: xtmepoisson error variance
> Date: Fri, 3 Jul 2009 10:49:03 -0500
>
> Margit Averdijk asks:
>
>> I am estimating xtmepoisson models. As I understand, the variance of the
>> error term is fixed at a minimum. Is there any way to set the variance of
>> the error term 'free' so that I can control for overdispersion?
>
> Define a new level of the model comprised of the observations themselves.
>
> . gen obs = _n
>
> When used as the final (lowest) level in -xtmepoisson-, the variance component
> corresponding to -obs- will then estimate the overdispersion.
>
> . xtmepoisson ... || obs:, ...
>
> --Bobby
> [email protected]
> *
> * 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/
_________________________________________________________________
Lauren found her dream laptop. Find the PC that’s right for you.
http://www.microsoft.com/windows/choosepc/?ocid=ftp_val_wl_290
*
* 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/
