THE ACQUISITION AND USE OF AN ADAPTIVE STRATEGY FOR ESTIMATING NUMEROSITY

This study investigates the development of an adaptive strategy for th e estimation of numerosity from the theoretical perspective of ''strat egic change'' (Lemaire & Siegler, 1995; Siegler & Shipley, 1995). A si mple estimation task was used in which participants of three different age groups (20 university students, 20 sixth-graders and 10 second-gr aders) had to estimate 100 numerosities of (colored) blocks presented in a 10 x 10 rectangular grid. Generally speaking, this task allows fo r two distinct estimation procedures: either repeatedly adding estimat ions of groups of blocks (= addition procedure) or subtracting the est imated number of empty squares from the (estimated) total number of sq uares in the grid (= substraction procedure). A rational task analysis indicates that the most efficient overall estimation strategy consist s of the adaptive use of both procedures, depending on the ration of t he blocks to the empty square. The first hypothesis was that there wil l be a developmental difference in the adaptive use of the two procedu res, and according to the second hypothesis this adaptive use will res ult in better estimation accuracy. Converging evidence from different kinds of data (i.e., response times, error rates, and retrospective re ports) supported both hypotheses. From a methodological point of view, the study shows the potential of Beem's (1995a, 1995b) ''segmentation analysis'' for unravelling subjects' adaptive choices between differe nt procedures in cognitive tasks, and for examining the relationship b etween these adaptive choices and performance.