A heteroscedastic Bayesian model for method comparison data

Abstract When implementing newly proposed methods on measurements taken from a human body in clinical trials, the researchers carefully consider whether have maximum accuracy. Further, they verified validity of new method before being implemented society. Method comparison evaluates agreement between two continuous variables to determine those agree enough interchange methods. Special consideration our work is variation with magnitude measurement. We propose evaluate when are heteroscedastic using Bayesian inference since this offers more accurate, flexible, clear, and direct model all available information. A simulation study was carried out verify characteristics accuracy different settings sample sizes. gold particle dataset analyzed examine practical viewpoint model. This shows that coverage probabilities parameters greater than 0.95. Moreover, relatively low error values, implies deals higher heteroscedasticity data others. In each setting, performs best size 500.