Hello All,
I am having difficulty using the metan command, both from the command
line and from the (installed) dialog screen.
I have a standardized mean difference effect size for theta, and the
standard error for se(theta). I also have an inverse variance weight,
if that is needed.
Following the directions on the metan dialog screen, I have entered the
basic command [theta] and [se(theta)] as:
. metan ES se_ES
And get an error that "too few variables" are specified. To resolve
this, I have tried various ways of entering the two variables (e.g.,
with and without the comma), and entering variables in most of the other
fields on the dialog screen (e.g., specifying my inverse variance weight
in the [weight] field; selecting Cohen's d, specifying random effects,
so on), but continue to get a range of errors. I have looked up quite a
few articles and help files on metan, but so far all examples use binary
data, or other approaches besides the Effect and se_Effect variables to
compute the meta-analysis stats.
I would appreciate someone sharing an example of the [metan] command
using "theta, se(theta)", or pointing me to an article or help file with
an example of metan used for this type of continuous effect size (i.e.,
Cohen's d), especially one using random effects model.
Thanks very much,
Wendy
*
* 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/