Gaze patterns can reveal different task strategies, which have been shown to predict performance on analogical reasoning tasks. In this experiment, we apply gaze analysis methods to a fraction comparison task designed to elicit alternate strategies (and thus differing gaze patterns) in different conditions. Preliminary analyses show that gaze patterns indicating conversion to equivalent fractions is both the most prevalent and the most accurate strategy, over benchmarking by magnitude. Cross-multiplication is most effective in the most difficult condition. We also explore the costs and benefits of consistently using a particular strategy versus flexibly switching strategies according to the specific fraction pairs presented.

Session Type: Poster Session