Calculate the treatment mean of the meta-analysis. Locate the treatment means of all of the individual studies. Add these treatment means together, dividing by the number of studies looked at. The resulting number is the meta-analysis treatment mean.
Calculate the control mean of the meta-analysis. Locate the control means of all of the individual studies. Add these treatment means together, dividing by the number of studies looked at. The resulting number is the meta-analysis control mean.
Compute the meta-analysis variance. This variance is the pooled variance of all the studies included in the meta-analysis. Find the sample sizes for each study and label them as "n1" for the first study, "n2 for the second study" and so on. Find the variances for each study and label them "v1" for the first study, "v2" for the second study and so on. Calculate the pooled variance with the equation "Sigma[(ni-1)*(vi)^2]/[Sigma(ni)-g], where "Sigma" represents the function that sums the variables over the entire set of studies and "g" represents the number of studies included in the meta analysis. This pooled variance is the variance for the meta-analysis.
Determine the effect size. Subtract the meta-analysis control mean from the meta-analysis treatment mean. Divide the result by the square of the meta-analysis variance. The resulting number is the effect size for the meta-analysis.