# Re: st: thanks / a new question - heterogeneity test

 From Nicola Orsini To statalist@hsphsun2.harvard.edu Subject Re: st: thanks / a new question - heterogeneity test Date Fri, 25 Nov 2005 09:31:12 +0100

Hi Andrei,

you can get the Q statistic, its p-value and the I2 statistic in different ways:

1) Use -metan- instead of -meta- command

The Q-test is stored in r(het)
The p-value is stored in r(p_het)

2) Use -metareg- instead of -meta- command

The Q-test is stored in e(Q)
The degrees of freedom of Q is stored in e(df_Q)

Then, you can easily calculate the p-value by yourself
scalar p_het = chiprob(e(df_Q), e(Q))

Nicola

Andrei Malinovschi wrote:

Dear statalisters,

I would like to start by thanking you for the quick
answer - special thanks to Maarten Buis, Neil
Shephard, Roger Newson and Hervé Stolowy. I got both
the answer + further ideas how to improve the process
/ my analysis.

I have now another problem when doing a metaanalysis
and running the meta command. I am interested in the
random pooled estimate (95% CI) and heterogeneity test
(Q and p-value). I have found that the random effect
(95%CI) is stored in the global macros S_1, S_3, S_4,
but I cannot find where Q and the p-value for
heterogeneity test are stored. Any suggestions or
another way to retrieve them?

A possible solution that I was thinking about and it
is probably too complicated would be to calculate Q
(from the mathematical formula) and then use heterogi
to get both the p-value for heterogeneity test and
even I^2 that is not given by meta command.

Best regards,
Andrei

__________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
```*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```