Hi.
Meta-anlysis and stats dummie here.
Recently I started to experience a little with MA and RevMan.
I did some simulations using the raw values at post intervention (both mean and Sd).
However, some authors instead using the values of post-intervention, prefer to use the difference between post and pre values. While I understand the process for the mean, I am having troubles understanding the process to the determine the variance/Sd.
Can someone give me a pratical example, step by step (I know it will be simple). Here goes an example.
Intervention X - BP (baseline): 145 ± 8; BP (post intervention): 130 ± 6; N=35
Control: BP (baseline): 146 ± 9; BP (post intervention): 139 ± 4; n= 25
So, for Intervention X, I would insert 15 (mean) and 35 (total), in RevMan. For control would be 7 (mean) and 25 (total).
What would be the Sd/variance in those cases?
" Variances were calculated from the pooled SDs of change scores in the intervention and control groups." - An example of the methodological description using by those authors.
Thanks a lot!
Hi Jorge,
I suggest you to see links on topic.
Dealing with missing standard deviation and mean values in meta ...
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5842611/
16.1.3.2 Imputing standard deviations for changes from baseline
https://handbook-5-1.cochrane.org/.../16_1_3_2_imputing_standa...
Combining Effect Size Estimates in Meta-Analysis ... - Semantic Scholar
https://pdfs.semanticscholar.org/.../7804b7df89d229fd5db7a1c567...
Computation of different effect sizes like d, f, r and transformation of ...
https://www.psychometrica.de/effect_size.html
How to Select, Calculate, and Interpret Effect Sizes | Journal of ...
https://academic.oup.com/jpepsy/article/34/9/917/939415
Best regards
Thanks Mohamed-Mourad Lafifi .
I believe the answer is in that Cochrane link, option 1.
Do you know if there some software that makes that calculation more easily?
Lastly, do you know if I insert that value in the "SD" collumn of RevMan, or do I need to use a different setting or a different software to perform the calculation that this method requires?
Thanks
HI Jorge Teixeira
Finally, did you get any software to calculate more easily? I have the same problem than you. Thanks!
Hi,
The free open-source software JASP (Version 0.8.6. Amsterdam Netherlands) (JASP, Team 2018) for Inferential procedures. Also you have NCSS Statistical Software.
Best regards
Hello
I have collected the average yields of the controls and treatments. The standard deviations are missing in the articles, and I need them to calculate the size of the effect. Would you please help me to find a formula
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis