Research Summary

Mental rotation has been a central topic in cognitive psychology and spatial cognition since the pioneering experiments of Roger Shepard and Jacqueline Metzler (1971). Their work demonstrated that humans appear to internally transform spatial representations through a continuous mental process, providing important insights into mental imagery and the way the mind manipulates objects in space.

While decades of research have investigated the cognitive mechanisms underlying mental rotation, an equally important question concerns how spatial ability itself is measured. Traditional mental rotation assessments generally rely on fixed sets of items, where every individual receives the same questions and performance is evaluated primarily through a total score. Although these methods have contributed greatly to cognitive research, they provide limited information about the characteristics of individual items, how difficulty emerges, and how precisely different levels of spatial ability are measured.

This project represents an early step toward the development of a more adaptive and personalized approach to mental rotation assessment. The current assessment is designed as an item calibration study using Item Response Theory (IRT), a modern psychometric framework that examines the relationship between individual responses, item characteristics, and underlying cognitive ability.

Unlike traditional scoring methods that mainly focus on the total number of correct answers, IRT models mental rotation ability as a latent trait and estimates how each item functions along this ability dimension. Through participant responses, the model can investigate important item properties such as difficulty, discrimination, and the amount of information each item provides at different ability levels.

The three-dimensional stimuli used in this assessment were computationally generated to explore how specific spatial features influence cognitive demand. Characteristics such as object complexity, rotational transformations, structural similarity between alternatives, and distractor design are examined to better understand why some items are easier or more difficult than others. The objective is not only to measure performance, but to construct a calibrated map of mental rotation items across different levels of spatial ability.

This calibration process represents the foundation required for future Computerized Adaptive Testing (CAT). Before an adaptive system can select questions automatically, it first requires a well-understood item bank where each item has known psychometric properties. Using information obtained from IRT analysis, future adaptive algorithms could dynamically select the most appropriate items for each individual, allowing accurate estimation of spatial ability with fewer questions and reduced testing burden.

Beyond assessment optimization, this approach may contribute to broader applications in cognitive science, education, and rehabilitation. A more precise understanding of spatial ability measurement could support personalized learning environments, track changes in visuospatial cognition over time, or assist future cognitive training and rehabilitation programs where individual progress needs to be monitored accurately.

Ultimately, this research combines spatial cognition, computational stimulus generation, and psychometric modeling to explore how mental rotation ability can be measured more precisely. This first stage focuses on understanding the structure of the assessment itself, transforming a collection of spatial problems into a scientifically calibrated measurement system capable of supporting future adaptive technologies.