ABSTRACT Objective The 28-item Academic Motivation Scale (AMS) is a widely used measure of students’ motivation to learn but it has the common limitations of an ordinal scale such as low precision and unsuitability for parametric statistics. The current study aimed to evaluate the psychometric properties of the AMS using Rasch methodology and enhance the precision of the scale using ordinal-to-interval transformation. Method The Partial Credit Rash model was used to analyze responses of 429 New Zealand medical students who completed the scale in English. Results The initial poor fit of the AMS to the Rasch model was improved by creating four super-items combining dependent subscales/items that displayed higher residual correlations with each other. These modifications resulted in the best fit to the Rasch model with no significant deviations of scale parameters from the model expectations (χ2 (24) = 19.79, p = 0.71), invariance across sex, age and ethnic groups, unidimensionality and high reliability (PSI = 0.81). Conclusions This study supported the robust psychometric properties of the AMS and produced conversion tables to transform the ordinal AMS scores into interval-level data to enhance the precision of the scale and enable use of parametric statistics without altering the original scale structure. KEY POINTS What is already known about this topic: (1) The Academic Motivation Scale (AMS) was developed to measure academic motivation which is important in higher education settings to evaluate students’ motivation to learn. (2) The AMS is an ordinal scale which is not suitable for parametric statistical tests and has other limitations of an ordinal measure, such as low precision. (3) The full-scale AMS is the most reliable according to the recent evidence, while the individual subscales of the AMS have low generalisability across student population and response occasions. What this topic adds: (1) Using Rasch methodology provides a powerful tool to evaluate and enhance the psychometric properties of an ordinal scale such as the AMS. (2) Rasch model fit indicates that the AMS complies with fundamental principles of measurement, such as unidimensionality, invariance, and interval scale metrics. (3) This study produced ordinal-to-interval conversion tables to transform ordinal responses to the AMS into interval level data that increases precision of the instrument and its suitability for parametric statistics