Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: Meta analysis


From   "Tiago V. Pereira" <tiago.pereira@incor.usp.br>
To   statalist@hsphsun2.harvard.edu
Subject   RE: st: Meta analysis
Date   Wed, 13 Aug 2008 23:50:50 -0300 (BRT)

Cynthia,

It is not clear what kind of effect size you are dealing with;
unstandardized mean difference? or standardized mean difference? rates?

In any case, -metan- only works on a log scale with raw data and using
binary outcomes (i.e. effec sizes measuref as risk ratio/odds ratio). If
you donīt have binary outcomes, and want to see your effect sizes
approximately symmetrical around the point estimate, you need to transform
the values. However, if you have negative values this will not possible,
and you have to work on the original scale. there is no way out as far as
I know.

Regarding the binomial option and aprox. and absolute methods, maybe other
listers can help, because I dont have these options in my current version
of metan.

Cheers!

Tiago



*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index