Shifting from cubical to equitable spatial task designs reduces gender differences in children’s mental rotation performance

Abstract Spatial ability, particularly mental rotation, has been widely studied in relation to gender differences. Ensuring equitable assessment of spatial skills is especially relevant for sustainability-related STEM (Science, Technology, Engineering and Mathematics) domains, where early educational experiences may shape later participation patterns. This study examined gender differences using a newly developed computerised Mental Rotation Test (nMRT) measuring accuracy and response time across 28 items varying in stimulus type and difficulty. To enhance inclusivity in task design, the test included concrete objects (e.g., cars, dolls, animals) and abstract figures (cubes, pellets, polyhedra), comprising male-stereotyped, female-stereotyped, and neutral objects. Stimuli were presented in two test parts, with rotations in picture plane followed by rotations in depth. Participants were 279 German primary school children ( Mage = 8.66 years; 134 girls, 145 boys), tested in classroom settings. Reliability analyses indicated high overall internal consistency of the nMRT ( α = 0.86). Multivariate analyses revealed no overall gender differences in accuracy or response time ( ps > 0.05), indicating broadly comparable performance across the full test. A small but reliable accuracy advantage for girls emerged only for picture-plane rotations of animals and letters ( ηp 2 = 0.03). Although boys demonstrated a small advantage for female-stereotyped stimuli rotated in depth, this effect did not remain statistically significant following correction for multiple comparisons. Controlling for age did not substantially alter the overall pattern of findings. Improved performance on some abstract stimuli across test parts was consistent with the effects of task familiarity. Overall, the findings suggest that gender differences in childhood mental rotation performance are sensitive to contextual and task-related factors, including stimulus content and rotation format, rather than emerging uniformly across spatial tasks.