Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: metan with repeated-measures designs?


From   "Ploutz-Snyder, Robert (JSC-SK)[USRA]" <robert.ploutz-snyder-1@nasa.gov>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: metan with repeated-measures designs?
Date   Thu, 22 Jul 2010 15:56:22 -0500

Good afternoon Statalisters;

I am considering a meta analysis of studies that were all similarly designed as 2-factor, completely repeated measures factorials. The outcome of interest is continuous/normal.
 
The two factors are Time (pre, post) and Leg (left, right), where an experimental intervention was applied to one leg per human subject, with their other leg serving as (obviously within-subject) control.  Thus each study is a straight forward 2(time) x 2(leg) completely within-subjects factorial design, and I'm hoping to get a meta analytic estimate of across-study effects.  I'd also like to factor in a couple of covariates (study sample size, duration of the intervention) eventually.

As typical of meta analytic projects, the various papers don't supply me with all that I'd like. Most, however, report means(SD) of the continuously scaled outcomes,  pre and post, for each leg.  So I can construct a dataset that is kinda/sorta like the data that user-written -metan- command would typically use.


My admittedly meager understanding of the -metan- command is that it is designed for completely independent measures design, and furthermore with only one mean(sd) and n per group.  Is there a method for somehow adjusting for the repeated-measures design if I feed -metan- the pre and post means and SD of only my TREATED leg (ignoring the already established null change in control legs).  Is it possible to expand to a two factor design?  

I'm open to any thoughts on a better approach for meta analysis with this sort of experimental design??

Much obliged.




*
*   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   |   Site index