Dear all,

I was searching on how to handle the missing SDs in a RCT. Sometimes only mean is given and SD is missing.

The cochrane handbook suggests that either we can impute the missing SD from other studies included in the same meta analysis or secondly we can estimate by calculating correlation coefficient.

I want to know which method is more preferable. I know it is best to avoid these methods as much as possible but if we really need to include that study, which method can be used?

I would be really grateful if someone can answer.

Best wishes