Skip to main content
. Author manuscript; available in PMC: 2016 Jan 1.
Published in final edited form as: J R Stat Soc Ser C Appl Stat. 2014 Dec 5;64(3):469–489. doi: 10.1111/rssc.12087

Figure 1.

Figure 1

Bias and coverage for estimated disease prevalence, sensitivity, specificity, PPV and NPV from the full likelihood (FL) and the composite likelihood (CL) methods. The true overall disease prevalence is 0.2, sensitivity is 0.9, and specificity is 0.9. The data are generated from bivariate GLMM (for case-control studies) and trivariate GLMM (for cohort studies). Results are summarized from 1000 simulations. The x-axis represents for the different settings of pairwise correlations among study-specific prevalence, sensitivity and specificity (in logit scale).