On the estimation of the median odds ratio for measuring contextual effects in multilevel binary data from complex survey designs (opens in new tab)

In studies with clustered or hierarchical data structures, quantifying between-cluster heterogeneity, referred to as contextual effects, is crucial for valid cluster-level inference. The median odds ratio (MOR), derived from random effects (RE) logistic regression models for clustered binary data, provides an intuitive assessment of contextual effects. Most existing research focuses on point estimation of the MOR for two-level models, with limit...