NEWS.md

kaefa 0.1.428.1

New Features

• Added fitdistrplus integration for empirical theta prior estimation

  • fitThetaPrior(): Fit distributions to raw scores using fitdistrplus
  • testThetaPriorCalibration(): Test if calibration works for non-nominal models by comparing raw score distributions with theta estimates
  • applyThetaPrior(): Apply fitted distribution during model calibration with automatic validation
  • New vignette "Setting Theta Priors with fitdistrplus" documenting the functionality
  • Comprehensive examples in examples/ directory

• Added fitdistrplus as a package dependency

Purpose

This update addresses the need to set theta priors based on empirical raw score distributions, particularly useful for:

• Non-representative samples
• Validating model calibration for non-nominal models
• Documenting empirical ability distributions
• Improving reproducibility of IRT analyses

kaefa 0.1.0.9000

• Added a NEWS.md file to track changes to the package.