I am trying to analyse some complex data looking at the effect of an supplement on exercise recovery. The problem is I have data from several studies that look at repeated measures data over time and some that also have different doses of the supplement introducing several kinds of dependency. I have calculated hedges g' for effect sizes but I want to preform a multivariate meta-analysis with study Id as a level but the correlation is unknown. Is it best to do a metaregression to estimate an overall correlation or to use several correlations to see the impact. Any help with the syntax would also be helpful.
Data is set as so:
ID | Year | ES | SE |
1 | 2010 | -1.23859 | 0.48 |
2 | 2011 | 2.392134 | 0.54 |
3 | 2011 | 2.113046 | 0.51 |
4 | 2014 | -1.08273 | 0.54 |
5 | 2015 | -0.73929 | 0.52 |
5 | 2015 | -0.90916 | 0.53 |
6 | 2014 | -0.20178 | 0.46 |
7 | 2015 | -0.045 | 0.42 |
8 | 2016 | -1.13688 | 0.55 |
9 | 2016 | -1.22741 | 0.56 |
10 | 2016 | -0.00641 | 0.39 |
11 | 2016 | 0.914 | 0.44 |
12 | 2020 | -0.04651 | 0.34 |
12 | 2020 | -0.39975 | 0.35 |
12 | 2020 | -0.0757 | 0.34 |
12 | 2020 | -1.23925 | 0.38 |
13 | 2020 | -0.393 | 0.32 |
0 Response to Using Mvmete for multiple outcomes and multiple treatments?
Post a Comment