Strategic Limitations in Preschoolers’ Relational Spatial Reasoning: Insights from Eye‐Tracking and Instructional Intervention

ABSTRACT Background Relational spatial reasoning is foundational to later mathematical and navigational competencies, yet little is known about the strategic processes preschoolers employ when integrating relational information. Eye‐tracking offers a means to capture these processes, but evidence remains limited, particularly regarding how brief instructional interventions shape children's visual strategies. Aims This study examined whether a short, strategy‐focused instructional intervention enhances preschoolers’ relational spatial reasoning and whether improvements are reflected in behavioural accuracy, eye‐tracking indicators of visual strategy, and think‐aloud evidence of metacognitive engagement. Sample(s) Participants were 82 Taiwanese preschoolers aged 5–6 years, randomly assigned to an experimental group receiving relational spatial reasoning instruction or to a control group receiving time‐matched literacy activities. Methods Children completed three 20‐min instructional sessions followed by a post‐test relational map task while behavioural accuracy and eye‐tracking metrics were recorded. ANCOVA was conducted to compare post‐test performance while controlling for pretest scores. Eye‐tracking indicators included time to first fixation on key landmarks, distractor revisits, scanpath patterns, and fixation duration. Logistic regression examined predictors of responsiveness to instruction, and think‐aloud protocols were analysed to corroborate visual strategy patterns. Results The experimental group demonstrated higher post‐test accuracy and shorter time to first fixation on target landmarks compared with the control group. Instructed children made fewer revisits to distractors and showed more systematic scanpaths, although groups did not differ in total fixation duration. Logistic regression identified shorter time to first fixation and fewer distractor revisits as significant predictors of responsiveness, whereas total fixation duration was not predictive. Think‐aloud data indicated that responders engaged more frequently in planning, monitoring, and evaluation. Conclusions Findings clarify the strategic components underlying early relational spatial reasoning by distinguishing between search initiation, selective engagement, and sustained processing. The results underscore the value of promoting visual efficiency and attentional control in early spatial education and demonstrate the utility of eye‐tracking as a diagnostic tool for tailoring instructional support.