you need the standard error (SEi) of the effect coefficient bi for each study . The weight for each study i is wi =1/ SEi2 . The weghted mean effect coefficient is then b = Sigma (wi bi)/ Sigma wi

If you do not know the variance apriori, a simple thing to do will be to assume Poisson distribution. It means that you assume that the variance is proportional to the sample size and variance of the mean is inversely proportional to the sample size ....

you need the standard error (SEi) of the effect coefficient bi for each study . The weight for each study i is wi =1/ SEi2 . The weghted mean effect coefficient is then b = Sigma (wi bi)/ Sigma wi

Dear James,

Thank you very much for your answer.