Choice autonomy and memory for spatial locations in six-year-old children.

Six-year-old children were presented with 12 identically labelled locations in a room and required to search non-redundantly for the six that contained rewards. On each day they could commit Across-Trial Memory (ATM) errors by visiting non-rewarded locations, or Within-Trial Memory (WTM) errors by revisiting locations previously visited that day. Groups were trained for nine days, either walking or pushed in a wheelchair, and with or without freedom of choice. They were then tested, walking with freedom of choice. Performance of the task improved significantly across training days in groups allowed free choice, whether walking or transported, and was superior at test to that of non-choosing groups. Throughout training and testing, the ATM component of performance was superior in groups allowed free choice. WTM was more accurate than would be expected by chance in all subjects at all stages of the experiment, but did not differ between groups. The problem of comparing WTM scores in groups differing in ATM accuracy was discussed. It was concluded that the primary benefit of free choice in spatial memory tasks is that it promotes accurate environmental segmentation, and that WTM is little affected by training or environmental familiarity.