# Bridging Item Response Theory and Factor Analysis: A Four-Parameter Mixture-Dichotomized Model with Bayesian Estimation

arXiv: https://arxiv.org/abs/2407.04071 · HTML: https://mlsanity.com/paper/2407.04071/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Jul 2024.

## Built on by (most cited first, 1 of 1)

- [Response-free item difficulty modelling for multiple-choice items with fine-tuned transformers: Component-wise representation and multi-task learning](https://mlsanity.com/paper/2605.16991/) (2026)

## Abstract (arXiv)

Item Response Theory (IRT) and Factor Analysis (FA) are two major frameworks for modeling multi-item measurements of latent traits. While the relationship between two-parameter IRT models and dichotomized FA models is well established, FA formulations for IRT models with additional parameters are less common. We focus on the four-parameter factor-analytic (4P FA) model that extends the traditional dichotomized single-factor FA model through a hierarchical mixture formulation accounting for guessing and inattention effects. We analytically establish the equivalence of the 4P FA and 4P IRT models, extending the FA--IRT correspondence beyond the two-parameter case. A Bayesian estimation procedure is developed for model estimation, to estimate the four item parameters, the respondents' latent scores, and the scores adjusted for guessing and inattention effects. The proposed algorithm is implemented in textttR and textttPython. A simulation study compares estimation under the FA and IRT formulations of the 4P model and evaluates the practical implications of the FA parametrization. Empirical examples based on an admission test and an anxiety inventory demonstrate the correspondence between the 4P FA and 4P IRT models and illustrate the application of the proposed methodology.