AMETA: Stata module to perform alternative and Bayesian meta-analysis

ameta offers a choice of two meta-analysis approaches, an alternative and a bayesian. On the one hand, the alternative method consists of eight different heterogeneity estimators for the calculation of the pooled intervention effect, its std. error and CI, the between-studies variance tau-squared, as well as other heterogeneity measures under a random-effects model and with the assumption that the pooled intervention effect is following a normal distribution. On the other hand, the bayesian method offers three Bayesian meta-analysis models for calculating parameters such as the posterior variance and expected value of mu and posterior variance and expected value of tau-squared of the unknown population parameter theta along with its credible interval. All three models are of the random-effects family so they take into account both the within and between study variation and estimate it. Moreover, since a Bayesian inference is used, we also assign prior probability distribution to the parameter theta and specifically an inverse gamma prior distribution with parameters a and b which are also known as shape and scale parameters of the distribution.