From Simulation to Hardware: Gender Dynamics and Transfer of Computational Thinking Skills in Educational Robotics

Despite the emphasis on computational thinking (CT) as a key 21st‑century skill, longitudinal evidence remains limited on how CT skills develop and transfer across contexts, particularly from sim‑to‑real learning environments. This longitudinal study involved 69 pre‑university science students who participated in three iterative phases of Arduino‑based robotics labs focused on autonomous navigation challenges. Data sources included CT self‑efficacy surveys administered at three time points, near‑transfer tasks assessing algorithmic reasoning, and performance scores on both virtual simulation and physical hardware challenges. A 2×3 repeated‑measures ANOVA revealed significant growth in CT skills over time, with males initially outperforming females; however, these gender differences diminished by the end of the semester, resulting in no gender effects on final virtual or physical performance. Regression analyses indicated that early CT skills strongly predicted near transfer, while success in virtual tasks predicted far transfer to physical environments, supporting the effectiveness of sim‑to‑real approaches for adaptable problem‑solving. Overall, the findings underscore the potential of inclusive robotics curricula to reduce gender inequities in CT development and transfer, while highlighting the bridging role of hybrid virtual‑physical designs.