Individual differences in children's strategy selection in a computational estimation task with two versus three available strategies.

We investigated how children's strategy selection on different problem types was influenced by whether two or three strategies were available in a computational estimation task. Importantly, we examined the influence of individual differences in working memory updating on these effects. Third and fourth graders (N = 725) were asked to indicate the best strategy for two-digit addition problems (e.g., 47 + 24) without calculating estimates. Homogeneous problems (i.e., both unit digits smaller than 5 or larger than 5) and heterogeneous problems (i.e., one operand's unit digit smaller than 5 and the other's unit digit larger than 5) were included. Children completed selection tasks under two conditions: (a) a three-strategy condition, in which they could choose among the rounding-down strategy (i.e., rounding both operands down), the mixed-rounding strategy (i.e., rounding one operand down and the other operand up), and the rounding-up strategy (i.e., rounding both operands up), and (b) a two-strategy condition, in which they could select between the rounding-down strategy and the rounding-up strategy only. As expected, children chose the best available strategy more often under the three-strategy condition than under the two-strategy condition and more often on homogeneous problems than on heterogeneous problems. Importantly, these effects were moderated by children's updating capacities. That is, children with less efficient updating showed worse selection performance on heterogeneous problems than on homogeneous problems under both conditions. In turn, children with more efficient updating showed comparable performance for both problem types under both conditions. These findings have important implications to further our understanding of underlying processes in children's strategy selection in computational estimation.