On the relative speed account of number-size interference in comparative judgments of numerals

Humans show systematic congruency effects due to irrelevant variations of the numerical value or the physical size of digits in judgments about either of these 2 attributes alone. According to influential models (e.g., J. Tzelgov, J. Meyer, & A. Henik, 1992), these effects are characterized by genuine asymmetries of size and number processing not accounted for by simple relative speed considerations, whereas some recent work (e.g., A. Pansky & D. Algom, 1999) partly challenges this view. This article presents 2 qualitative gradient-based predictions made by relative speed models and a diffusion-based implementation of the relative speed view to quantitatively account for response times and error rates in comparative judgments of digits. The results of 2 experiments using a completely task-symmetric design are in accord with these detailed predictions; they are also consistent with the view that both number and size are converted into magnitude representations of similar structure.