Evaluation of Cognitive Architecture Through Embodied Learning Applied to Robotic Programming From a Gender Perspective

Abstract

Traditional evaluations of educational interventions often prioritize average treatment effects, obscuring two critical dimensions: the distributional equity of outcomes and the structural integration of cognitive skills. We address these gaps by comparing project-based Arduino robot application (PARA) instruction to screen-based active learning (SBAL) across three domains: Robotics achievement, computational thinking, and task persistence. In a quasi-experimental study of N = 55 undergraduates (PARA: n = 16; SBAL: n = 39), we employ a multimethod framework combining psychometric network topology, fuzzy-set qualitative comparative analysis (fsQCA), and distributional metrics. Results suggest that PARA was associated with a more integrated pattern of relationships among achievement, computational thinking, and persistence within this sample. Network analysis indicated a comparatively denser exploratory triad in the PARA condition, where persistence, computational thinking, and achievement were more strongly interrelated (density = 66.7%). The SBAL condition showed a comparatively sparse pattern of interdomain associations (density = 0.0%). Furthermore, fsQCA identified exploratory configurations associated with high achievement under PARA, including patterns involving computational thinking and task persistence; these subgroup patterns are interpreted cautiously given the small and unequal cells. Distributional analyses suggest that PARA may reduce outcome inequality and support educational mobility within the present sample. Taken together, the findings are consistent with the interpretation that instructional modality may shape how learners coordinate cognitive and affective resources, although stronger causal claims require replication. The PARA method’s emphasis on embodied debugging may help legitimize persistence as one route to mastery. Evidence suggests that equity-conscious science, technology, engineering, and mathematics pedagogy should be designed and evaluated not only for average achievement gains, but also for its capacity to support integrated cognitive-affective learning processes.