I'm in the process of doing a meta-analysis and have encountered some problems with the RCT data. One of my outcom is muscle strength. In one study, I have three different measurements of muscle strength for the knee joint (isometric, concentric, eccentric). I wonder how to enter data into the meta analysis. If I give them separately, I increase the number artificially (n). The best form would probably be to combine them within this one study, because in other studies included in the meta analysis, the authors give only one strength measurement.
Thank you all for any help.
Yes, I recommend that you either combine the three outcomes in one measure, or choose the one outcome which most resembles the muscle strength outcome in the other studies. If you included all three measures, the present study would have (artificial) greater weight in the meta-analysis, and study samples would not be independent, since the study population is included three times.
I agree with Jesper Fabricius also I recommend you check Cochrane Manual. They have quite good information about combine data in meta-analysis and if meta analysis is not possible how you can do synthesis using other methods.
https://training.cochrane.org/handbook/current
So idea would be to:
1. perform meta-analysis and combine data from this one manuscript (3 outcomes)
2. and then put results (d, CI, N etc.) to the main meta-analysis as an one outcome.
In this way, the effect size and SE also correspond to the multiplied number of subjects - N. I found that if I divide N / the number of variables it is converted correctly.
- Badges
- Science topic