Investigating Malleable Factors During Spatial Training: Transfer to Preschoolers’ Spatial Skills but Not Math Outcomes

• In Event: Examining the Roles of Patterning Knowledge, Spatial Assembly, and Analogic Reasoning in Early Mathematics Development

Abstract

Spatial skills, or the ability to mentally or physically manipulate objects and spaces in our environment, are ubiquitous and predict later STEM achievement (Mix & Cheng, 2012; Wai et al., 2009). Though spatial skills are malleable (Uttal et al., 2013), little is known about specific malleable factors that may be most beneficial for spatial and math learning during preschool interventions. For example, language (Jordan, 2007; LeFevre et al., 2009) and gesture (Ping et al., 2011; Ehrlich et al., 2006) play an important role in children’s understanding of spatial and mathematical constructs. Digital spatial interventions may be beneficial platforms for spatial learning.

We explored the effectiveness of a spatial puzzle assembly intervention across concrete versus digital delivery. Three-year-olds (N=331; Mage=42.98 months, SD=3.24; 49% high-SES) were pre- and post-tested in preschools on the 2D- and 3D-Tests of Spatial Assembly (TOSA) (Verdine et al., 2017), Spatial Language Comprehension (Bower et al., 2020), Beery Visual Motor Integration (VMI), shape identification, and math (WJ: Applied Problems, TEMA). The spatial training (2D-TOSA) taught children to piece together different colored geometric forms into designs shown in a model image. An adult trainer (concrete-training) or character in the app (digital-training) corrected wrongly-placed puzzle pieces using (1) modeling and feedback (MF) or (2) condition-specific feedback: gesture feedback (GF) or spatial language feedback (SLF). Modeling and feedback in GF and SLF conditions incorporated gesture or spatial language techniques. Children were randomly assigned to one of 6 training regimens (see Table 1) administered 1x/week over 5 weeks or a no-training control.

We hypothesized that (1) children who received training, regardless of format, would increase their spatial (2D TOSA) and math skills more than control group children; and (2) children with condition-specific training (GF, SLF) would be more likely to transfer learning than those in the control group or MF.

Findings indicated that both concrete (B=.22) and digital (B=.21) training worked equally well to enhance learning. Collapsed across conditions—all training increased posttest performance on the 2D TOSA, p’s<.001 relative to the control group. However, there was no transfer to other spatial or math skills from either format’s training. Condition-specific training improved performance on some of the spatial tasks, in concrete and digital training (see Table 2). As predicted, Concrete-SLF had a significant effect on Shape Identification compared to control (B = .11, p = .014). Digital-GF had a significant effect on 3D TOSA (B = .13, p = .021). Surprisingly, concrete-MF had a significant effect on Spatial Language Comprehension compared to the control (B = .11, p = .049).

We observed significant differences in performance on transfer spatial tasks as a function of specific feedback during training. However, because children’s transfer between 2D-representations and 3D-objects is challenging (Barr, 2010), it is not surprising that transfer from either training to some spatial and math tasks was not observed. Thus, specific malleable factors, especially spatial language, hold promise for promoting spatial skills in young children, but more work is needed to identify ways of facilitating transfer to a broader range of mathematical contexts.

Authors

• Laura J Zimmermann, Tech Play Collaborative
  Presenting Author

• Corinne A. Bower, California State University - Los Angeles
  Non-Presenting Author

• Brian Nicholas Verdine
  Non-Presenting Author

• Tamara Spiewak Toub, Temple University
  Non-Presenting Author

• Kathy A Hirsh-Pasek, Temple University
  Non-Presenting Author

• Roberta M Golinkoff, University of Delaware
  Non-Presenting Author