A novel numerical algorithm that provides functionality for estimating the exact 95% confidence interval of the location parameter in the random effects model, and is much faster than the naive method. Works best when the number of studies is between 6-20.
