Selection of procedures in mental subtraction: Use of eye movements as a window on arithmetic processing

Adults who use mental procedures other than direct retrieval to solve simple arithmetic problems typically make more errors and respond more slowly than individuals who rely on retrieval. The present study examined how this extra time was distributed across problem components when adults (n = 40) solved small (e.g., 5 - 2) and large (e.g., 17 - 9) subtraction problems. Two performance groups (i.e., retrievers and procedure users) were created based on a 2-group cluster analysis using statistics derived from the ex-Gaussian model of reaction time (RT) distributions (i.e., μ and τ) for both small and large problems. Cluster results differentiated individuals based on the frequency with which they used retrieval versus procedural strategies, supporting the view that differences in mu and tau reflected differences in choice of strategies used. Patterns of eye movements over time were also dramatically different across clusters, and provide strong support for the view that individuals were using different mental procedures to solve these problems. We conclude that eye-movement patterns can be used to distinguish fluent individuals who readily use retrieval from those who rely more on procedural strategies, even if traditional self-report methods are unavailable. (PsycINFO Database Record