Real-world object size as a magnitude dimension: Robust perceived size distance effects and context-sensitive spatial coding.

Research on numerical cognition suggests that magnitude processing supports both distance-sensitive comparison and directional spatial mapping. Whether real-world object size shows a similar profile, however, remains unclear. The present study examined this question across two preregistered experiments using task contexts adapted from paradigms in the numerical domain. In Experiment 1, participants judged which of two simultaneously presented objects was smaller or larger in real-world size. An arrangement-based compatibility pattern emerged in the small-object task but not in the large-object task; importantly, the task x compatibility interaction was not significant. In Experiment 2, the expanded model including mapping order revealed a SNARC-like tendency, but its expression was qualified by mapping order. A phase-based analysis further suggested that this modulation could be parsimoniously explained by the combination of a weak SNARC-like tendency and block-related baseline speeding. Across both experiments, responses became faster as perceived size distance increased, indicating robust sensitivity to relative magnitude structure. Together, these findings suggest that real-world object size is processed as a meaningful magnitude dimension, while its directional spatial expression is weak, context-sensitive, and not uniformly expressed across task contexts.