Do Spatial Labels Facilitate Children’s 2D-to-3D Transfer of Relational Information in a Spatial Mapping Task?

• In Event: Poster Session 06
  In Poster Session: PS 06 Topic - Cognitive Processes

Abstract

Relational reasoning, as when children are asked to recognize the commonality of a type of spatial configuration between two objects, is a challenging skill for three-year-old children. One tool for overcoming this difficulty has been the use of relational labels, which have been shown to be powerful tools - capable of drawing attention to subsurface relational structure (Loewenstein & Gentner, 2005). As young as 18 months, children benefit from hearing a relational term, such as a spatial locative term (Casasola, 2005), shifting from encoding only the objects to instead attending to the type of spatial relation between those objects. Whether relational labels support relational reasoning and transfer from 2D (i.e. computer screen) to 3D (i.e. real world) environments remains an important question. There is a known difficulty for young children in transferring information from 2D to 3D sources (Barr, 2010), and while this difficulty is largely resolved by the third year of life, it has been found in older children in more challenging tasks (e.g. Tarasuik, Demaria, & Kaufman, 2017). Young children consume a significant amount of digital content in a normal, non-pandemic year (Rideout, 2017), and may rely on digital content as a part of remote learning, making this a worthwhile topic of investigation.

In this study, we examined the performance of 3-year-old children in a remote digital to analog mapping and transfer task conducted over the Zoom video conferencing platform. Children (N=15) were randomly assigned to either a spatial labels condition (n=6) or a deictic labels condition (n=9). They were introduced to Spot and Dot, two cartoon dogs who were described as loving to play hide-and-go-seek together, and always hiding the same way. The dogs were then shown hiding in one of three locations, either in an array of bowls or an array of mugs, with accompanying description. In the spatial labels condition, children were told “When Spot hides [on the left/on the right/in the middle], Dot hides [on the left/on the right/in the middle],” while children in the deictic labels condition were told “When Spot hides here, Dot hides here.” This was demonstrated twice in each hiding location. Following this training, participants received six on-screen mapping trials, in which one dog hid and participants were asked to point to where the other dog would hide. The on-screen array of bowls was then removed, and children were presented with six additional trials and asked to transfer the relation to real world bowls in front of them, provided by their parents. Performance was analyzed with a 2 (condition) x 2 (trial type) repeated measures ANOVA, revealing a significant main effect of trial type, F(1, 13) = 8.57, p =0.01. While participants across both groups performed significantly above chance in the on-screen mapping trials, t(14)= 2.98, p<0.01, they did not perform significantly above chance in the transfer trials to the real-world bowls, t(14)= 0.21, p=0.83. These results suggest that children at this age may struggle to transfer on-screen relations to real-world ones, even when relevant relational labels are provided.

Authors

• David Nicholas Tompkins, Cornell University
  Presenting Author

• Victoria Guevara, Cornell University
  Non-Presenting Author

• Max Portnoy, Cornell University
  Non-Presenting Author

• Tamar Kushnir, Duke Univeristy
  Non-Presenting Author

• Marianella Casasola, Cornell University
  Non-Presenting Author