I sought to perform meta analysis of continuous variables
I have extracted mean and SD from multiple studies for a variable
What is the statistical method to compute a "global" mean + SD from those variables ?
Asked
Active
Viewed 19 times
1
-
Well your question already contains the answer: do a meta-analysis. So what is your question exactly? – mdewey Nov 14 '21 at 11:42
-
1I think you're looking for the term 'inverse variance', but there's more to it than just that. For example, are these single arm or comparative studies? Are they all using the same scale? Cochrane Handbook is a good place to start reading up on these and more. – abousetta Nov 14 '21 at 12:28
-
Well Infact it seems that what I was looking for was how to calculate weighted mean. However, I still dont know how to compute multiple stadard deviation ? Is there a method to calculate "weighted" SD ? – rupescissa Nov 17 '21 at 12:27
-
Yes. There's a way to combine any moments from multiple datasets (for what it's worth). See https://stats.stackexchange.com/a/51927/919. It is related to the formulas for combining moments with weights, https://stats.stackexchange.com/a/16609/919. – whuber Nov 18 '21 at 19:18