This article discusses the challenges facing meta-analyses considering multiple sources of heterogeneity in peer-reviewed studies. While quantitative meta-analysis can control or correct sources of heterogeneity, it focuses on the search for statistical consistency which limits the extent of causal inference and generalisation. In contrast, qualitative approaches for meta-analysis tend to integrate various sources of heterogeneity through interpretive approaches. However, this integrative strategy leads to a constitutive or functional relationship between analysed constructs with limited capacity for theory testing. Comparative configurational meta-analysis (CFG-MA), a diversity-oriented approach, has an overall advantage in managing heterogeneity as it recognises the uniqueness of the cases and their within-case diversity by keeping the assumption of cross-case causal complexity. CFG-MA allows the integration of various sources of heterogeneity and the production of cross-cases and within-cases inference and generalisation.