Key points are not available for this paper at this time.
Digital-intelligence transformation in education has made pre-service teachers’ Technological Pedagogical Content Knowledge (TPACK) a strategic concern in teacher preparation. Survey data from 11,818 pre-service teachers across 17 local normal universities in China were analyzed through hierarchical regression, quantile regression, and structural equation modeling to examine how perceived university ICT support and perceived AI support in education are associated with self-reported TPACK. Both forms of support showed significant direct and model-conform indirect associations with self-reported TPACK, but the quantile coefficients varied across the TPACK distribution: university ICT support showed a modestly fluctuating descriptive pattern, whereas AI support in education peaked at the median and attenuated at upper quantiles. ICT self-efficacy and AI competency expectancy each formed significant indirect pathways in the hypothesized model, although the ICT pathway was more strongly indirect and the AI pathway remained more strongly direct. Additional checks of university-level ICCs, cluster-robust standard errors, and measurement invariance across key subgroups supported the robustness and comparability of the findings. These patterns clarify how perceived ICT and AI support are differentially associated with self-reported TPACK and provide empirical grounds for more precise, human-in-the-loop support designs in teacher education.
Liu et al. (Thu,) studied this question.