Dear All,
I am writing a meta analysis. My data from the studies only shows Pre: Mean (SD) Post: Mean (SD) for both treatment and condition. I want to summarize their findings.
I can calculate the difference of means and the SE. However, because they use different scales for the outcome 1-10, 1-21 or 1-40 (for pain), I think I need to transfer it to SMD. But How do I calculate that? And consequently how do I calculate the SE for each SMD per study because I need to provide that in RevMan for the generic inverse variance...
Lots of thanks!
https://stats.stackexchange.com/questions/447845/compute-standardized-mean-difference-and-its-standard-error-from-regression-adju
formulae for smd and sesmd for unadjusted means are given in the attached link...
QUESTION 1. You can refer to the "Cochrane Handbook of Systematic Reviews", where it is explained how to impute SDs for change-from-baseline values when you only have pre- and post-test data.
See here for further details: https://handbook-5-1.cochrane.org/chapter_16/16_1_3_2_imputing_standard_deviations_for_changes_from_baseline.htm
QUESTION 2. Once you have calculated changes from baseline (and their SDs) of intervention and control groups, you can use the SMD as a summary statistic. This can apply to your case because you report to have the same outcome measured in a variety of ways.
See here for further details: https://handbook-5-1.cochrane.org/chapter_9/9_2_3_2_the_standardized_mean_difference.htm
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis