Development and Initial Psychometric Evaluation of the Digital-Intelligence Literacy Scale for Medical Students: Measurement Dimensions and Structural Analysis

Abstract

Background/Objectives: Medical education requires assessment of students’ capabilities across digital technologies, health data, and artificial intelligence (AI). Existing instruments address related constructs but differ in their measurement scope. This study aimed to develop and initially evaluate the Digital-Intelligence Literacy Scale for Medical Students (DIL-MS). Methods: Literature-informed content analysis guided conceptual framework development and generation of 47 candidate items. A cross-sectional online survey using purposive sampling yielded 373 usable responses from fifth-year clinical medicine and anesthesiology students at Southern Medical University, China. Principal component analysis with Promax rotation and item analysis guided item refinement. Confirmatory factor analysis assessed the proposed five-factor model. Internal consistency, composite reliability, and convergent validity were examined. Results: The final scale comprised 34 items across five dimensions: Information Management, Digital-Intelligent Technology Application, Human–AI Collaboration, Ethics and Safety, and Critical Thinking and Innovation. The five-factor model showed moderate fit (CFI = 0.874, TLI = 0.864, RMSEA = 0.082). Overall Cronbach’s α was 0.962 (95% CI: 0.956–0.967), with subscale coefficients ranging from 0.902 to 0.939. Composite reliability ranged from 0.904 to 0.939, and average variance extracted ranged from 0.580 to 0.698. Conclusions: The DIL-MS demonstrated high internal consistency and preliminary evidence of convergent validity within a model with moderate fit. It may help characterize medical students’ self-reported learning needs and inform curriculum planning. Independent validation and further evaluation of content coverage, discriminant validity, and relationships with demonstrated performance are needed.