Author:

Keywords:

Cognitive development, Experimental studies, Mathematics, numeracy

Abstract:

Participants’ estimation pattern on a number line estimation task is usually interpreted as the reflection of their underlying mental number line. However, recent findings (Ashcraft & Moore, 2012; Barth & Paladino, 2011) challenge this widespread assumption by suggesting that strategies might play an important role when solving this task. The present study tested this assumption directly by gathering trial-by-trial verbal strategy reports when solving a number line estimation task. Sixty-three adults made number line estimations on a 0 to 1000 number line. Participants were assigned to one of three conditions in which the number of benchmarks on the number line was varied to elicit potential benchmark-based estimation strategies: (a) only the origin and endpoint were indicated (bounded condition); (b) an extra benchmark at the midpoint (500) was presented (midpoint condition); (c) three additional benchmarks (250, 500, and 750) were specified (quartile condition). Results indicated that participants in the midpoint and quartile condition estimated more accurately than in the bounded condition. Furthermore, the verbal strategy reports revealed that, as the number of provided benchmarks increased, participants relied more frequently on the halving strategy, in which they split the number line into halves, quarters and sometimes even eights. Importantly, we observed that in all three conditions, participants not only made use of the externally presented benchmarks, but also used them to create more refined internal benchmarks when making number line estimations. These findings have both theoretical and educational implications.