MUSIC for Medical Students: Confirming the Reliability and Validity of A Multi-Factorial Measure of Academic Motivation for Medical Education

Abstract

Construct: The MUSIC® Inventory measures the construct of academic motivation across five factors: empowerment, usefulness, success, interest, and caring. The factors are defined in terms of the degree the student perceives that they have control over their environment, that the coursework is useful to their future, that they can succeed in the course, that the course and instructional methods are interesting, and that the teacher cares about their wellbeing and their success respectively. Background: A valid measure of medical students’ academic motivation would provide medical teachers with a method for evaluating the motivational aspect of their course, and provide focus for changes in teaching and learning to improve medical student engagement. While the MUSIC® Inventory structure has been validated in the tertiary setting and with several professional programs, it has not been validated with medical students. The aim of this study was to use both classical test theory and Rasch modelling to assess the reliability and confirm the structure of the five-factor model of the MUSIC® Inventory with medical students. Approach: 152 medical students completed the 26-item inventory. Descriptive statistics, internal consistency, correlations between factors, confirmatory factor analysis, and Rasch analysis using the rating scale model were performed to determine reliability and validity. Findings: The five factors showed good internal consistency (Cronbach’s alpha 0.87 – 0.92). Correlations between factors were moderate to high (r=.38 – .89). Confirmatory factor analysis highlighted inconsistencies in factor loadings of three of the items hypothesized to measure interest. Rasch analysis using the rating scale model showed that all items for each factor had good item fit (0.65 – 1.37). Person separation (2.28 – 2.85) and reliability (.84 –.91) scores indicated that the scales were able to differentiate different levels of respondents. Item separation (2.25 – 6.97) and reliability scores (.83 –.98) indicated that the items of the scales were being differentiated by the respondents. Conclusions: Rasch analysis indicates that the five factors of academic motivation measured by the MUSIC® Inventory account for the response patterns in data from medical students. However, while the factors of empowerment, usefulness, success, and caring showed expected reliability and validity using classical analysis, three of the interest items cross-loaded on to the usefulness factor. Possible reasons may include ambiguity of language for the items or medical students' conception of usefulness and interest. Future research will explore medical students' understanding of the language used to measure these factors in further detail.