This a follow-up question to a previous question about calculating CI's for pooled proportions, thank-you for your previous help The question is as follows: If a study uses some sort of stratified sampling and one is pooling it with a bunch of other studied what proportion does one use when using the weighted method (ie when there is no significant heterogenity) e.g. raw data uncorrected for sampling design 100/20000 data, corrected for sampling design 8000/2000000 Thanks again Paul Waraich