RE: st: Meta analysis of effect using effect/CI method

["Tiago V. Pereira" <tiago.pereira@mbe.bio.br>]

Will,

If your LS mean differences are approximately normally distributed, I
believe that the command:

metan lsmean lowerci upperci, [options]

should work fine as long as you have an appropriate version of -metan-.
Make it sure you don't have an older version of it. I remember that the
earlier version accepts only either four or six variables (a b c d for 2x2
tables, and n1 m1 sd1 n2 m2 sd2 for mean differences, respectively). Since
you are on Stata 8.2, I guess that some of the newest versions of -metan-
may not work. So, try to find out the version 1.86 1Apr2004.

If you need further assistance, please, do not hesitate to contact me.

Cheers!

Tiago



------

I am trying to use the dialog box option to execute a meta-analysis of
continuous data in which I was provide the LS Mean difference as the
measure of effect, and the 95% confidence interval of the estimate. I am
using Stata version 8.2. I selected the "Effect/CI" option for the
analysis in the dialog. The dialog tells me to enter theta (effect), the
LCL and UCL in that order into the text box. When I execute the program I
get an error stating that I have provided too few variables. The software
executes metan when trying to conduct the analysis.

Thanks,
Will


*
*   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/