Measuring Perceived Ease of Artificial Intelligence Utilisation: An Adaptation and Psychometric Validation of a Scale among Preservice Mathematics Teachers in Nigerian Context

Abstract

Perceived ease of use is a foundational construct in technology-acceptance research, yet instruments developed to measure it in relation to artificial intelligence (AI) remain scarce within African teacher-education contexts. Existing scales are rarely re-examined for the specific referent of AI-assisted mathematics learning. This study adapted a six-item Perceived Ease of Use (PEOU) from the literature to an AI-in-mathematics-learning context and examined its psychometric properties among drawn 221 preservice mathematics teachers using a multistage non-probability sampling procedure from three higher institutions in the Federal Capital Territory (FCT), Abuja, Nigeria. Confirmatory factor analysis of the initial six-item one-factor model showed inadequate fit (CFI = .886, TLI = .811). However, the removal of two items with low standardised loadings produced a four-item model with an excellent fit (χ²(2) = 2.770, p = .250, CFI = .991, TLI = .972, RMSEA = .042, SRMR = .026). The retained scale demonstrated acceptable internal consistency (ω = .723). Non-parametric comparisons revealed no statistically significant difference in PEOU AI across institutional type (Kruskal–Wallis χ²(2) = 3.637, p = .162) or gender (Mann–Whitney U, z = -.567, p = .571). The findings support a psychometrically sound, unidimensional four-item PEOU-AI scale for use in Nigerian mathematics teacher education and suggest that ease-of-use perceptions of AI tools are broadly comparable across the sampled institutions and gender groups.