The Multidimensional Assessment of Interoceptive Awareness (MAIA) has gained ground as a psychometric approach to tracking positive changes in aspects of interoceptive sensibility in the form of self-reported tendencies in beliefs and attitudes toward bodily signals. The instrument has been translated and validated across of many contexts throughout the world. However, issues in validation studies have suggested the presence of some poorly fitting items leading to inconsistent reliability on a few of the subscales. Modeling of item responses would extend the understanding of these issues through estimation and assessment of measurement precision across groups through parameters such as item difficulty and discrimination. This study uses item response theory to analyze data from a survey of the English MAIA and Japanese MAIA as well as a pre-post investigation of the Japanese MAIA in an applied psychoeducational context. Graded response models for the polytomous items of the MAIA were generated in R and fit was compared between constrained and unconstrained item discrimination estimates. The results are forthcoming, but the research objective is to determine if items perform similarly or differently across samples, and whether stability among item performance can be observed in the Japanese version of the MAIA.