Figure 3
From: Longitudinal hierarchical Bayesian models of covariate effects on airway and alveolar nitric oxide
Power and Type I errors for the 6 methods (distinguished by color) for true β of 0.1, with power for a given NO parameter’s association calculated from the simulation scenario where only that association is non-zero (e.g., power for \({\beta }_{Ca}\) from Scenario 2 where \({\beta }_{Ca}=0.1, {\beta }_{logCaw}=0, {\beta }_{logDaw}=0\)) and Type I error calculated under two scenarios (e.g., Type I error for \({\beta }_{Ca}\) under S3: \({\beta }_{Ca}=0, {\beta }_{logCaw}=0.1, {\beta }_{logDaw}=0\) and under S4: \({\beta }_{Ca}=0, {\beta }_{logCaw}=0, {\beta }_{logDaw}=0.1\)), with the non-zero NO parameter association denoted by shape.
