Skip to main content


Fig. 3 | BMC Medical Research Methodology

Fig. 3

From: Prediction models for clustered data with informative priors for the random effects: a simulation study

Fig. 3

Calibration plots for Bayesian models using discrepant expert opinion as prior information for the random effects. Predicted risks are plotted against the true latent underlying risks for 5000 subjects from 50 equal sized clusters. Clusters using optimal expert opinion are displayed in grey color, whereas clusters using discrepant expert opinion are addressed in black color. The diagonal line is the line of identity (i.e., predicted risks are equal to the true risks). Each dot represents a subject, and each line formed by the dots represents a cluster

Back to article page