Re: st: Metandi- diagnostic meta-analysis- error calculating cooks distance

[Roger Harbord <rmharbord@gmail.com>]

Dear Jannik,
I haven't seen this error message before. It originates from
-gllapred- rather than from -metandi- itself, as -predict- after
-metandi ..., gllamm- is essentially a wrapper for -gllapred-.

I tried googling the error message and got this recent thread
<http://www.stata.com/statalist/archive/2010-11/msg00467.html> which
suggests you should first make sure your version of -gllamm- is up to
date (-adoupdate gllamm, update- or -ssc update gllamm, replace-).

If that doesn't help, feel free to email me off-list following the
procedure outlined at
<http://www.stata.com/support/faqs/techsup/sendout.html>. Please also
include the commands -which metandi- and -which gllamm-.

Regards,
Roger.
-- 
Roger Harbord
http://www.epi.bris.ac.uk/staff/rharbord.htm


On 4 January 2011 12:37, Jannik Helweg-Larsen <jhelweg@dadlnet.dk> wrote:
> I get this error message when trying to check for particularly
> influential observations using Cook's distance in the user-written "
> metandi"-
>
> (*! version 2.0 May 27, 2008 @ 15:44:22-*! Author: Roger Harbord,
> University of Bristol- from ssc)
>
> predict cook, cooksd
>
> (Cook's distance may take a few seconds...)
> can't get correct log-likelihood: -292.53257 should be -292.52883
> something went wrong in comprob3
> r(198);
>
> this option is described in  the help  file " help metandi postestimation"
>
> here are the details of the preceeding command:
>
> metandi tp fp fn tn , gllamm force
>
<snip>

*
*   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/