Development and psychometric validation of the barriers to AI integration in nursing education scale (BAI-NE): a multi-institutional study

This multi-institutional study aimed to develop and psychometrically validate the Barriers to AI Integration in Nursing Education Scale (BAI-NE). A methodological study with a mixed-methods exploratory design was conducted in two phases: scale development and content validation, followed by psychometric evaluation. The 15-item BAI-NE was developed based on the Unified Theory of Acceptance and Use of Technology (UTAUT) and an extensive literature review.

Content validity was assessed by a panel of ten experts. A convenience sample of 470 nurse educators from seven faculties of nursing participated in the study. The sample was randomly divided for Exploratory Factor Analysis (EFA; n = 235) and Confirmatory Factor Analysis (CFA; n = 235).

EFA identified a four-factor structure explaining 42.7% of the total variance, comprising Technical and Infrastructure Barriers, Knowledge and Training Barriers, Ethical and Regulatory Barriers, and Attitudinal and Organizational Barriers. EFA factor loadings ranged from 0.566 to 0.738. CFA provided preliminary support for the proposed model, with good fit indices (χ²/df = 0.90, RMSEA = 0.00, CFI = 1.000, TLI = 1.015, SRMR = 0.043, GFI = 0.961); because χ² was lower than df and CFI and RMSEA were at their bounds, these values should be interpreted cautiously and replicated in independent samples.

Standardized CFA loadings ranged from 0.54 to 0.82. Internal consistency was satisfactory, with Cronbach’s alpha values ranging from 0.713 to 0.754 and McDonald’s omega values ranging from 0.715 to 0.754. Test–retest reliability over a two-week interval was excellent for all subscales (ICC = 0.85–0.92).

Convergent validity was modest, with composite reliability values exceeding 0.70, while AVE values ranged from 0.426 to 0.451. Discriminant validity was supported using the Fornell-Larcker criterion and HTMT values