A paper by Ryuya Komuro (JSPS Research Fellow PD) of the Department of Applied Linguistics has been published in Sage Journals.

This paper introduces mfrmr, a free, open-source package for the statistical software R. It centers on the many-facet Rasch model, a statistical approach for situations where human judges assign scores — for instance, grading English essays or speaking tests. In such settings, factors beyond the test-taker's own ability, such as how strict a particular rater is or how difficult a given task turns out to be, can influence the final score. The many-facet Rasch model takes these factors into account to produce scores that are fairer and more reproducible.
mfrmr supports several model variants, including the rating-scale model and the partial-credit model. Its key strength lies in offering one continuous workflow: fitting a model to the data, checking model fit, screening for bias between raters, running generalizability analyses, and generating reports or exportable files, all within a single package. By allowing every step from raw ratings to final estimates, diagnostics, and reports to be traced, mfrmr supports a more transparent approach to assessment.
Designed specifically for rating data involving multiple judges, it is expected to be useful in both educational and research settings. The official package is freely available on CRAN (https://cran.r-project.org/web/packages/mfrmr/index.html).

Title: mfrmr: Estimation and Diagnostics for Many-Facet Measurement Models in R

For more information, please see the published article below:
https://doi.org/10.1177/01466216261470436