Poster #43 - Directional Space-Number Mappings on the Vertical Axis

• In Event: 1-169 - Poster Session 04
  In Poster Session: PS 04 Section - Cognitive Processes

Integrative Statement

Children and adults organize numbers along the horizontal axis, mapping smaller numbers to the left side and larger numbers to the right (Dehaene, Bossini, & Giraux, 1993; Opfer & Furlong, 2011). But adults also align numbers on a vertical axis, with larger numbers represented higher in space than smaller numbers (Ito & Hatta, 2004; Schwarz & Keus, 2004). As discussed in Holmes & Lourenco (2012), the directional mapping of numbers onto vertical space may arise from asymmetries due to upright body position and gravity. Increases in quantity also co-occur with greater vertical extension, and MORE IS UP metaphorical expressions link quantity and vertical orientation in language, e.g., “the prices went down” (Lakoff, 1987). Given these perceptual and linguistic factors, we hypothesized that a directional space-number mapping on the vertical axis develops early in children. We employed a nonsymbolic numerosity comparison task (Patro & Haman, 2012) to investigate vertical space-number mappings in preschoolers, who lack formal mathematical instruction or extensive reading experience.

Thirty-two 4-5-year-olds were shown pairs of dot arrays displayed on the upper and lower regions of a tablet display. Children viewed 54 test trials. Children were instructed to quickly touch ‘‘the plate with more cookies (or less cookies, for half the participants)” using their dominant hand. An image of the sun was displayed after correct answers, and a rain cloud was displayed after incorrect ones. Association of vertical direction with number was hypothesized to lead to a spatial-numerical congruity effect: we predicted greater accuracy and shorter reaction times for displays with “less” targets at the bottom and “more” targets at the top compared to display with “less” targets at the top and “more” targets at the bottom.

Accuracy and reaction time were analyzed using mixed-effects regression models with instruction (“more”, “less”) and position (‘more is UP’, ‘more is DOWN’) as fixed effects and item and subject as random effects. Likelihood-ratio tests were conducted to compare nested models differing in the presence of a fixed main or interaction term. There were no main effects of instruction (p=0.55) or position (p=0.31) in accuracy scores; however there was a significant interaction (χ² (1) = 19.28, p < 0.001). Children were more accurate when the “more” target appeared on the top half of the screen (β=–0.92, SE=0.25, p<0.001), whereas accuracy scores for the “less” target did not vary by position (β=0.35, SE=0.33, p=0.28) (Figure 1). Children were faster to respond when indicating “more” targets (χ² (1) = 5.21, p < 0.05) versus “less” targets but the predicted main effect of position failed to materialize (p=0.40). Reaction times are not shorter when “more” targets appear at the top and “less” targets appear at the bottom (Figure 2).

Our study provides partial support for a “more is UP” mapping in preschoolers. They are more accurate–if not faster–in detecting larger magnitudes in upper space. There is no evidence that “less is DOWN” for children. Our findings demonstrate that, in some contexts, children can flexibly organize numerical magnitudes in space at an early age.

Authors

• Bhuvana Narasimhan, University of Colorado Boulder
  Presenting Author

• Norielle Adricula, University of Colorado Boulder
  Non-Presenting Author

• Mara Strother, University of Colorado Boulder
  Non-Presenting Author

• Eileen Ford, University of Colorado Boulder
  Non-Presenting Author