Redesigning STEAM Education for Engineering Creativity in AI-Driven Learning Environments

This paper contributes to the literature on redesigning STEAM education for the engineering domain by introducing structural equation modeling (SEM) and the Heckman selection model in the empirical analysis in addition to the conventional regression-based creativity assessment at the individual learner level (engineering students), learning modality variation (AI-supported instruction). Using secondary data from higher education engineering programs for the period of recent academic cohorts, we aim to examine whether AI-driven learning environments affect engineering creativity outcomes. To achieve this aim, the Heckman selection model for creativity participation was used for the first time—instead of using the standard regression approach—to incorporate the selection process into the outcome equation as an endogenous factor in creativity performance models. We used engineering creativity scores per student as our main dependent variable for creativity outcomes, as well as other instructional and learner characteristics, in our empirical models, which were estimated using structural equation modeling (SEM). Considering the presence of sample selection bias, estimation of the creativity models was corrected using the two-step method of Heckman. Indeed, there is an empirically significant relationship between AI-driven learning environments and engineering creativity outcomes, indicating that the result supports the proposed conceptual framework. Estimation results indicated that AI-supported STEAM instruction increased creativity levels, and that there was a statistically significant effect from AI-driven learning to creativity for engineering students in the selected sample. In conclusion, it appears that traditional STEAM instruction has less influence on the observed creativity outcomes than in other instructional configurations.

Hali tarjima qilinmagan