Bayesian Estimation of Generalized Item Response Models with Flexible Tail Behavior

Description

This dataset provides a comprehensive computational framework for the Bayesian Estimation of Generalized Item Response Models with Flexible Tail Behavior. It features highly optimized algorithms for both the Generalized 2-Parameter Logistic Model (GL2PLM) and the Generalized Graded Response Model (GLGRM). The codebase is structured to address two primary estimation paradigms: (1) Known Shape Parameter: providing benchmark recovery analysis when $p$ is a pre-specified constant; and (2) Unknown Shape Parameter.

Steps to reproduce

Requirements: R (>= 4.0.0) R Packages : `Rcpp`, `RcppArmadillo`, `coda`, `ggplot2`, `BayesLogit` Compiler: A working C++ compiler (Rtools on Windows, Xcode on macOS, or GCC on Linux) File Structure scripts :Contains R scripts for simulation and analysis. src: Contains C++ source files for the MCMC engines. How to Reproduce 1. GL2PLM with known and unknown: Run `GL2PLM_simulation_revised.R` to verify parameter recovery for GL2PLM. 2. GLGRM with known p: Run `GL2PLM_p_known_code.R` to to verify parameter recovery for GLGRM. 3. GLGRM with unknown p: Run `GLGRM_p_unknown_code.R` to execute the MCMC chain with slice sampling for $p$.

Institutions

• Minnan Normal University

  Fujian, Zhangzhou

Categories

Funders

• National Social Science Fund of China

  Grant ID: 21BTJ036

Licence