Code for Hierarchical Bayesian Non-response Models for Error Rates in Forensic Black-Box Studies
This dataset contains all R code used to produce the results in the paper Khan, K., & Carriquiry, A. (2023). Hierarchical Bayesian non-response models for error rates in forensic black-box studies. Philosophical transactions. Series A, Mathematical, physical, and engineering sciences, 381(2247), 20220157. https://doi.org/10.1098/rsta.2022.0157.
In particular, it includes the JAGS models and run code for the naive, ignorable, and non-ignorable models discussed in the above-referenced paper. The naive and ignorable modes are fully Bayesian, but the non-ignorable is fit using an empirical Bayes approach. The code used to obtain the MLEs for the associated hyper-parameters in this model is also included. Finally, code to estimate error rates using the relevant posterior predective distribution is here, as well.