Figure 3
From: Goodness-of-fit testing for meta-analysis of rare binary events
Comparison of empirical Type I error rates by proposed IPQ, parametric bootstrap method (PB), standardization method (STD), Naïve method (Naive), Posterior Predictive Check (PPC), Sampled Posterior Check (SPC), and Pivotal Quantities method with cutoffs of 0.1 and 0.25 (PQ-0.1, PQ-0.25). Data were generated from the null cases with different \((I,\mu ,\rho )\) combinations, each 200 replicates. Tests were conducted at the significant level \(\alpha =0.05\).
