Bayesian factorial regression of perceived barriers in robotics education: heterogeneous associations by career stage, gender, and training

Designing successful professional development requires an understanding of perceived obstacles to integrating robotics in classroom; however, there are currently few validated assessment instruments and analyses of subgroup differences across educator career stages. This study looked at how pre-service teachers (PSTs) and teacher educators perceive barriers differently and validated a Perceived Implementation Barriers Composite (PIBC). From robotics education modules in African teacher education institutions, 109 educators in all (50 teacher educators and 59 PSTs) were enlisted. While a 2×2×2 Bayesian factorial regression with regional fixed effects looked at heterogeneity across position, gender, and training subgroups, confirmatory factor analysis (CFA) with WLSMV estimation and polychoric reliability assessment was used to evaluate the PIBC. Although single-factor CFA results showed limited model fit, the PIBC showed strong reliability (polychoric α=0.882). Multigroup CFA suggested caution in mean comparisons by supporting configural and metric invariance but not full scalar invariance. Compared to PSTs, teacher educators reported fewer obstacles (effect=−0.223,95%CrI:−0.421,−0.025). Although credible intervals overlapped, interaction analyses suggested potential differences by gender and prior training. The results highlight the need for tailored, context-sensitive professional development in robotics education by indicating that career stage may moderate the impact of gender and training on barrier perceptions.