When and How Children Understand Same/Different Relations

• In Event: 1-198 - Acquisition of Same/Different Relations by Human and Nonhuman Groups

Integrative Statement

The ability to detect sameness and difference is fundamental to relational cognition. This has led to much interest in the development of insight into the basic same and different relations (e.g. Ferry, Hespos, & Gentner, 2015; Hochmann et al., 2017). A key task used to assess relational insight is the RMTS task: Given AA, choose XX over YZ; and given AB, choose YZ over XX. Previous research has found that 4-year-olds pass the same-only RMTS without practice trials or feedback and that comparison processes are instrumental in acquiring abstract representations of these relations (Christie & Gentner, 2014). However, to claim that 4-year-olds have same/different representations, we must address (1) whether children can pass the same/different-mixed RMTS, and (2) whether children’s performance on RMTS is based on symmetry detection.

In Experiment 1, we tested 4-year-olds’ ability to match both same and different in the RMTS (n=75) (Figure 1a & b). Children performed above chance on same-different mixed (M=.69, p = .001), same-only (M=.63, p = .04), and different-only (M=.65, p < .001) versions of the RMTS, with no significant differences between groups.

In Experiment 2, we tested children’s understanding of symmetry. We gave 4-, 5-, 6-, and 8- 9-year-olds (n=80) a symmetry match-to-sample (SMTS) task analogous to the same-only RMTS. On each trial, children saw a standard card depicting two shapes that were symmetrical along the vertical axis and two alternatives– one that was also symmetrical and one that was not (Figure 1c).

If 4-year-olds’ success on the RMTS is based on symmetry, then they should perform equally well on the SMTS. However, we found that the SMTS was strikingly difficult for children. Both 5- and 6-year-olds were at chance, M=.60, p = .10 and M=.60, p = .15, respectively, while 8-and 9-year-olds performed significantly above chance, M=.68, p = .01. Interestingly, so did 4-year-olds, M=.63, p = .04. However, a one-way ANOVA found no differences between the age groups.

In Experiment 3 we ask whether comparison-based learning processes that facilitate children’s understanding of same/different also promote understanding of symmetry. Prior research found that early experience with easy-to-align stimuli facilitated later performance with hard-to-align stimuli (Lowenstein & Gentner, 2012). Here, we first presented children with four easy trials where the alternatives showed the same objects in either symmetrical or non-symmetrical configurations (Figure 1d). This was done to downplay potential surface similarities between the cards and encourage comparing the alternatives, thus highlighting the relational contrast. We then gave children eight trials of the standard SMTS (as in Experiment 1). We tested a new sample of 4-and 5-year-olds (n=36) on the modified task. An ANOVA found a main effect of Experiment--children in Experiment 3 performed significantly better than the 4-and 5-year-olds in Experiment 1 (p = .02). There was no main effect of age (4 vs 5) nor interaction between Experiment and Age.

In summary, 4-year-olds perform equally well on same and/or different-RMTS. Their high performance is not based on symmetry. However, as with same/different, comparison across exemplars supports symmetry processing.

Authors

• Ruxue Shao, Northwestern University
  Presenting Author

• Christian Hoyos, Exponent
  Non-Presenting Author

• Dedre Gentner, Northwestern University
  Non-Presenting Author