Academic journal article
By Spetch, Marcia L.; Parent, Marise B.
Psychonomic Bulletin & Review , Vol. 13, No. 5
Male and female children, 3, 4, and 5 years old, searched for a sticker that was hidden in 1 of 15 linearly aligned boxes. Two identical bear-shaped landmarks cued the sticker location, which was always in the middle of 3 boxes that separated the two landmarks. The absolute locations of the landmarks and sticker varied across training trials, but the distance in relation to each other remained constant Training continued until the child chose the correct box first for 3 consecutive trials or for a maximum of 20 trials. Striking age and sex differences emerged in acquisition: The percentage of children who reached criterion increased over age groups to 100% for the boys but stayed at approximately 20% for the girls. A landmark expansion test (with the landmarks moved farther apart) given to children who met criterion revealed that most of these children chose the middle location.
Using visual landmarks to find a hidden goal is common in both human and nonhuman animals, but the strategies used to derive spatial information from landmarks can vary across species (see Spetch & Kelly, 2006). One situation in which differences between species have been found is when a goal is hidden in the middle of an array of identical landmarks that maintain a constant distance from each other but are moved within the search space across training trials. Spetch, Cheng, and MacDonald (1996) and Spetch et al. (1997) found that adult humans responded differently than pigeons on "expansion" tests in which the landmarks were spread farther apart: In both computer touch-screen tasks and open-field tasks, adult humans consistently responded to the middle of the expanded array, suggesting a relational use of the landmarks, whereas pigeons searched at the trained distance from individual landmarks, suggesting an absolute strategy. Gerbils (Collett, Cartwright, & Smith, 1986) and nonhuman primates (MacDonald, Spetch, Kelly, & Cheng, 2004; Poti, Bartolommei, & Saporiti, 2005; Sutton, Olthof, & Roberts, 2000) also have not responded to the middle of the array on expansion tests. The difference between adult humans and these nonhuman species appears to reflect "preferred" strategy rather than "ability," because both pigeons and dark's nutcrackers can learn to respond in the middle of a landmark array if the training task cannot be solved on the basis of absolute distances (J. E. Jones, Antoniadis, Shettleworth, & Kamil, 2002; Kamil & J. E. Jones, 1997).
MacDonald et al. (2004) reported that human children also responded differently than human adults to landmark expansion tests. In one task, the search space consisted of an array of discrete hiding places, and the goal was in the middle of, and directly adjacent to, the four identical landmarks in training. Children (ages 3-7) and adults rapidly learned to choose the hiding place in the middle of the landmarks on training trials, but they responded very differently to expansion tests. Whereas all adults continued to choose the middle, only 1 out of 13 children chose the middle. The remaining children searched almost exclusively in hiding places adjacent to the landmarks, suggesting that they had used the landmarks as beacons. To determine how children would respond in a task that could not be solved with a beaconing strategy, MacDonald et al. tested children (ages 3-5) on another task in which the landmarks were farther from the goal. This search space was a tray filled with confetti, and the goal was a sticker hidden in the middle of, but a fixed distance away from, four identical landmarks. This task proved difficult for children to learn, and only 10 out of 19 children reached criterion within 20 trials. Of the 10 children who learned, 3 searched closest to the middle on the expansion test, suggesting a relational strategy. Most of the remaining children searched closer to the learned distance and direction from individual landmarks, suggesting an absolute strategy. …