Bayesian meta-analysis Theory, application, and implementation

This dissertation develops, applies, and implements a comprehensive Bayesian meta-analytic framework that combines Bayesian estimation, hypothesis testing, and model averaging. The framework is extended to accommodate different outcome types, incorporate several types of publication bias adjustment methods, and support meta-regression and multilevel meta-analysis. The dissertation also highlights the need for neutral and cumulative method assessment for future methodological improvements. Applications to large collections of meta-analyses quantify the footprint of publication bias in the published literature. Finally, implementation in user-friendly software (JASP) lowers the barrier for applied researchers to adopt the newly developed Bayesian meta-analytic framework.