Abstract

The present study was designed to examine whether extensive training would alter the processes underlying young children s addition performance from reconstructive processes to automatic retrieval of facts. Children in the second grade were trained until their mean response times and errors on an addition verification task were equal to those of children in the sixth grade. Before and after training, automatic retrieval was assessed with number naming and number matching tasks. The effect of training on retrieval processes was evaluated by comparing changes in retrieval processes for children who were trained and who were not. The effect of training was very systematic on overall response times. On the verification task, mean response times decreased significantly over training sessions. Similarly,on the number matching and number naming tasks, mean response times decreased signifi­cantly after training. However, training on the addition verification task did not significantly affect the retrieval of addition knowledge. On the number matching task, addition knowledge was retrieved automatically in the experimental group as indicated by a significant probe effect. However, the effect of probe did not vary either with test or stimuli. In contrast, the control group did not show a significant effect of probe. These results imply that automatic retrieval observed in the experimental group is not due to training per se, but due to other factors associated with the experimental group. A similar result was obtained from the number naming task. A fairly constant magnitude of facilitation was observed at most SOAs. However, this effect of probe was not modified by group, test and stimuli as in the matching task, implying that addition training did not affect the effect of probe.