Poster #42 - Young Children Don’t Notice the Odd One Out: the Development of Gist Representation

• In Event: 2-110 - Poster Session 07
  In Poster Session: PS 07 Section - Cognitive Processes

Integrative Statement

While counterintuitive, we know that children are less likely to make certain types of errors, specifically those related to more gist-based or abstract representations of the world (Anastasi & Rhodes, 2008; Reyna & Ellis, 1994). It has been posited that children are spared these effects due to a difficulty to form abstract representations. Yet, little is known about whether children can form abstract, gist representations in a novel learning task, and whether gist representations can be formed for dimensions that are relevant to what is being learned as well as dimensions that are not immediately relevant.

To address these questions, 60 adults and 60 children (5 to 8 years old) performed an A/B categorization task, followed by an item recognition test. Through response feedback, participants learned that category membership was determined by distortions of two prototypical dot patterns (the relevant feature for learning). Each item had a unique shape and colour (the irrelevant features for learning), but the shapes were all designed along the same shape space by morphing two shapes together. The recognition test consisted of seen and unseen items, and participants were asked to indicate which were old and new. Of the unseen items, half were created along the same shape space (same-space lures) and half were novel shapes (novel-space lures). Difference scores of correct rejections were calculated between the two types of lures. Memory precision (d’) was also calculated in three ways: (1) using all of the lures, (2) using only the same-space lures, and (3) using only the novel-space lures. If abstraction occurred for features that were not relevant to the category learning task (the shape space), we would expect superior rejection rates for the novel-space lures and a high difference score; if abstraction did not occur, we would expect comparable rejection rates across lures and a low difference score.

Indeed, children’s responses were much more consistent across lures. While children and adults both rejected novel-space lures more successfully than same-space lures, difference scores were significantly higher among the adults (Fig. 1). This suggests that children may not have developed a very well-defined representation of the shape space, while adults did. Further analysis found that there were also age effects among the children. Older children showed learning of the shape space, but younger children only showed a slight trend towards learning (Fig. 2). This is consistent with some studies finding that young children are more bound to items and verbatim information than their categories (Sloutsky & Fisher, 2004), although we cannot definitively claim that here given poor item memory across age groups. Nonetheless, these data may point to a general shift in representation from verbatim to gist across childhood, with young children showing little evidence of abstraction. Understanding how children’s representations may differ from adults in a learning context provides valuable insight into how they may see the world differently, as well as having valuable pedagogical implications.

Authors

• Erika Wharton-Shukster, University of Toronto
  Presenting Author

• Amy S Finn, University of Toronto
  Non-Presenting Author