It is fundamental to evaluate whether research findings are replicable for valid scientific discoveries. Systematic reviews and meta-analyses offer valuable opportunities to assess the replicability of studies on the same research topic. Due to many factors (e.g., differences in patients’ baseline characteristics), the studies in a systematic review could report fairly different results. In this talk, I will present a hybrid test for assessing the between-study inconsistency. The proposed approach could also be applied to investigate time-lag bias, where early studies may report more optimistic results than later studies. In addition, no rigorous statistical methods exist to characterize non-replicability in the current literature of systematic reviews and meta-analyses. I also introduce a new measure to quantify replicability based on standardized residuals from a leave-m-studies-out procedure. The measure examines the impact of non-replicability on the conclusion of a meta-analysis; it also differentiates non-replicability from heterogeneity. Simulations and real-data analyses are presented to illustrate the new methods’ performance.