Search for: All records

Abstract In this review, I summarize what we know about the development of sex‐related differences in spatial skills, their potential malleability, and their possible causes. Current evidence suggests that sex differences increase in size with age, at least for skills with assessments suitable for use across development. However, male advantages vary from nonexistent to substantial, depending on the skills assessed, task parameters, and culture. Training and practice can improve spatial skills, although interventions tested so far do not eliminate the male advantage (when there is one). The complex pattern of increases with age, task variation, and cultural variation challenges efforts to theorize about causation. Progress requires identifying the underlying cognitive and neural structure of the spatial domain, developing reliable and valid assessment tools suitable for use across wide age ranges, gathering large datasets from a variety of cultural settings, and identifying and investigating specific mechanisms for growth and change. 

Research on spatial thinking requires reliable and valid measures of individual differences in various component skills. Spatial perspective taking (PT)-the ability to represent viewpoints different from one's own-is one kind of spatial skill that is especially relevant to navigation. This study had two goals. First, the psychometric properties of four PT tests were examined: Four Mountains Task (FMT), Spatial Orientation Task (SOT), Perspective-Taking Task for Adults (PTT-A), and Photographic Perspective-Taking Task (PPTT). Using item response theory (IRT), item difficulty, discriminability, and efficiency of item information functions were evaluated. Second, the relation of PT scores to general intelligence, working memory, and mental rotation (MR) was assessed. All tasks showed good construct validity except for FMT. PPTT tapped a wide range of PT ability, with maximum measurement precision at average ability. PTT-A captured a lower range of ability. Although SOT contributed less measurement information than other tasks, it did well across a wide range of PT ability. After controlling for general intelligence and working memory, original and IRT-refined versions of PT tasks were each related to MR. PTT-A and PPTT showed relatively more divergent validity from MR than SOT. Tests of dimensionality indicated that PT tasks share one common PT dimension, with secondary task-specific factors also impacting the measurement of individual differences in performance. Advantages and disadvantages of a hybrid PT test that includes a combination of items across tasks are discussed.