Development of the nursing artificial intelligence readiness scale for nursing students: a validity and reliability study.

BACKGROUND: The rapid integration of artificial intelligence (AI) into healthcare has amplified the need for nurses who can engage with AI-supported systems safely and effectively. Assessing nursing students' AI readiness through a psychometrically sound instrument is essential for guiding curriculum design and targeted educational interventions. This study aimed to develop and psychometrically evaluate the Nursing Artificial Intelligence Readiness Scale (NAIRS) for nursing students. METHODS: This methodological scale development study was conducted with 416 nursing students enrolled during the 2025-2026 academic year at a public university in Türkiye. To avoid conducting exploratory and confirmatory factor analyses on the same dataset, the full sample was randomly split into two independent subsamples (EFA: n = 200; CFA: n = 216). The two subsamples were comparable in key demographic and AI-related characteristics. An initial 40-item draft was generated from the literature and reviewed by 10 faculty experts using a content validity evaluation; the mean CVR was 0.88. Following expert review, five items were removed, reducing the pool to 35 items. The revised draft was then pilot tested with 50 nursing students to assess item clarity, readability, comprehensibility, and administration flow, and minor revisions were made prior to the main application. The final 20-item scale was structured across four theoretical domains: Knowledge/Awareness, Willingness to Use AI, Self-efficacy, and Ethical Awareness. Construct validity was examined using EFA (minres extraction; Promax rotation) and first-order CFA (ML estimation) with multiple model fit indices. Convergent and discriminant validity were assessed using CR/AVE and the Fornell-Larcker criterion, respectively. Internal consistency was evaluated with Cronbach's alpha in both subsamples, and temporal stability was tested via a 2-week test-retest application in a subgroup (n = 45) using ICC (two-way random, single measures). RESULTS: The final scale demonstrated a four-factor structure explaining 52.08% of the total variance, with factor loadings ranging from 0.351 to 0.909. CFA supported the proposed model with good fit (χ²/df = 1.426, CFI = 0.971, TLI = 0.967, GFI = 0.910, RMSEA = 0.044, SRMR = 0.049). Internal consistency was high: subscale alphas ranged 0.801-0.828 (EFA sample) and 0.858-0.902 (CFA sample), and the total scale alpha was 0.911. Test-retest reliability indicated strong stability, with ICC values ranging 0.896-0.963 across subscales and 0.952 for the total scale. CONCLUSIONS: The findings provide initial evidence that the NAIRS is a valid and reliable instrument for assessing nursing students' readiness for artificial intelligence across knowledge/awareness, willingness to use AI, self-efficacy, and ethical awareness domains. The scale may be useful for educational needs assessment and curriculum planning in nursing education. CLINICAL TRIAL NUMBER: Not applicable.