Meta-Analysis Resources

Tools for Those Who Summarize the Evidence Base

Resources and networking for those who conduct or interpret meta-analyses related to any phenomenon that is gauged in multiple studies.

I have seen this question before, but I have a similar one and the answers here in the forum didn't help me much. I am trying to include a 3 arm study for meta-analysis .

My intervention of interest lets say A was compared with placebo B, and another active intervention say C. How do I extract the sample size data for these?

should it be A vs. (B/2 +C/2)? or should I concentrate on A vs. B and ignore intervention C?

will appreciate your thoughts

Views: 286

Reply to This

Replies to This Discussion

If the report does not indicate what the sample sizes were for each cell, then it seems you have two alternatives: (a) make an assumption about the cells having equal size; or (b) contact the author(s) to ask.

As to deciding what comparison is best, I usually counsel people to pick the conditions that best match the practices in the other studies. You seem to be interested only in A; if so, you would omit C. But I also counsel to make these types of decisions crystal clear when you report your meta-analysis--so that critics can see what you have done and why.

Hope that helps!

Thanks Blair, I appreciate it. It was is helpful.

I have one other questions which I urgently need suggestion with:

How do I compute Standardized Mean difference  (SMD)  (or the effect size or Chens d with Confidence Intervals) from studies that only report between group difference and confidence intervals, without any data
for change, (mean +/-SD) from baseline. Given the change in (mean and
SD) from baseline, I know meta-analyst softwares can generate either SMD of MD
depending on which metric I chose on the software.

Summary of my questions:

Given between group differences, how can I generate SMD for
meta-analyst software in other to run my analysis or how do I feed
the between group difference into the software when what I am actually
interested in is the SMD? is there any way of back calculating the change (mean+/-SD) from baseline us the MD and confidence intervals between the two groups?

I appreciate your help.


Sorry, I am no expert on meta-analyst software! Maybe someone else can help? You might think about posting your question in the effect size calculation group, which is dedicated to this purpose. -Blair

Thanks  a million, Blair, will post it on the ES blog.


If you are interested in statistical material, you can read how the meta-analysis method was used to study the effects of cbd. There are several reports of such studies in the public domain on medical portals. I think it will be interesting to you. I used these materials when I studied the natural treatment of cancerous tumors.

Reply to Discussion


© 2022   Created by Blair T. Johnson.   Powered by

Badges  |  Report an Issue  |  Terms of Service