Skip to main content
Fig. 3 | BMC Public Health

Fig. 3

From: Comparison of Bayesian and frequentist methods for prevalence estimation under misclassification

Fig. 3

Confidence interval coverage. Coverage by parameter set for several methods of 95% confidence interval (CI) construction. The nominal coverage of 95% is marked by a solid gray line, and the dahed lines mark coverage values of 90% and 100%. The traditional CI’s (Clopper-Pearson, Blaker, Sterne, and Wald-Rogan-Gladen) all exhibit significant under coverage. The bottom left inset shows the Lang-Reiczigel CI’s and the Bayesian HDI’s coverage distributions in more detail. It reveals the Lang-Reiczigel CI tends to have some over coverage and the Bayesian HDI’s coverage appears to be more symmetrical around the nominal 95% value

Back to article page