FUN For Fit
FUN For Fit is an R Shiny tool for evaluating the practical significance of item misfit in psychological and educational tests — methods central to Clinical Outcome Assessment (COA) validation.
Features
- Rasch model (1PL) and 2PL IRT analysis of weighted MNSQ fit statistics (TAM package)
- 2D IRT modelling to estimate correlation bounds after Köhler & Hartig (2017)
- Step-guided interface enabling non-technical researchers to conduct psychometric instrument evaluation
- Interactive result visualisation (Plotly)
Relevance
The methods implemented mirror those used in COA instrument validation in clinical trials — assessing whether individual items of a patient-reported outcome (PRO) measure behave as expected under an IRT model, and quantifying the practical consequences of misfit.
Axel Budde
Patient-Centered Outcomes Researcher & Clinical Data Scientist
Psychologist and PhD candidate specialising in patient-centred outcomes, psychometric validation, and clinical data science.