I have a dichotomous outcome variable and measuring OR as effect size. I have got data from two studies and both represent data in different form. One study gives experimental & control group frequencies and another gives odds ratio with total number of events. I don't know how to enter both type of entries in one outcome...
1: Try contacting the authors
2: Try calculator within RevMan
3: Try recalculating the number of events if there are enough data for separate groups.
4: If the data is not usable and you cannot access the data contacting the authors then you cannot use these data.
To calculate the pooled OR, you need a lo[g[OR] and standard error. You can calculate the OR for the first study since they provide the frequency for both groups, both calculating SE is necessary for both studies.
Do the authors provide 95% C.I. ? That helps to compute the SE.
You may refer to my recent meta-analysis which also pools the odds ratio in various studies:
Article Subjective sleep quality, blood pressure, and hypertension: ...
I had the same question. Very useful comments, thank you.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis