About this srcd poster session

---

Panel information
Panel 4. Cognitive Processes

Abstract

---

Mathematical word problems present significant challenges due to their cognitive demands, which extend beyond basic arithmetic skills (Fuchs et al., 2006). In 2nd grade, numerical magnitude increases problem difficulty (Daroczy et al., 2015). This study investigates the relationship between verbal working memory (VWM), visual-spatial working memory (VSWM), and spatial visualization (SV), and the types of errors second graders make when solving math word problems. While previous research has emphasized the relation of each of these cognitive skills to problem-solving performance separately (e.g. Swanson & Beebe-Frankenberger, 2004; Zheng et al., 2011; Mix et al., 2016), here we examine their collective impact on errors made in problem-solving.

214 second graders (118 boys, 96 girls) from diverse Chicago schools completed standardized cognitive assessments for each cognitive skill, and solved seven math word problems, including small and large number problems, single-step and multi-step problems, and problems with or without extraneous numbers. Overall, 44.3% of answered questions and 85.1% of participating children had at least one error, with an average error proportion of 0.448 (SD=0.301) across the seven questions. A coding scheme was developed to categorize errors, with the most common being: missing number errors (17.4%), operation errors (15.9%), computation errors (12.7%), and irrelevant number errors (6.5%).

Higher VSWM and VWM predicted fewer operation errors (p=.012 and p=.003 respectively), computation errors (p=0.048 and p<.001 respectively), and irrelevant number errors (p=.007 and p<.001 respectively). Additionally, higher VSWM and SV predicted fewer missing number errors (p=.035 and p<.001 respectively). To further investigate these relationships, we used K-means cluster analysis that identified three distinct cognitive profiles: a low cluster (43.5%), which had the lowest VSWM and SV scores (both ps<.001), a high cluster (31.8%) with high overall scores, and a spatial cluster (24.7%) with the highest VSWM (both ps<.01) but the lowest VWM (both ps<.001) (Figure 1). Kruskal-Wallis and post-hoc Dunn tests indicated that the low cluster made more errors overall than both the spatial and high clusters, and the spatial cluster also made more errors overall than the high cluster (all ps<.001) (Figure 2). Children in the low cluster also made more errors of each type than those in the high cluster (all ps<.001) and more missing number errors than those in the spatial cluster (p=.008). Children in the spatial cluster made more operation (p=.041), computation (p=.012) and irrelevant number (p=.040) errors, but did not make more missing number errors (p=.199) than those in the high cluster.

These findings extend previous research by showing that spatial skills—VSWM and SV—relate to missing number errors, and working memory skills—VSWM and VWM—relate to the other three common error types. This suggests that instructional strategies may be differentially effective depending on students’ cognitive skills (Swanson et al., 2015). Future research should focus on developing targeted interventions that address common errors made by learners with diverse cognitive skills to better support word problem-solving.

Author information

Author	Role
Tania Dhaliwal, University of Chicago	Presenting author
Dania Carr, University of Chicago	Non-presenting author
Kelly S. Mix, University of Maryland	Non-presenting author
Susan C. Levine, University of Chicago	Non-presenting author

⇦ Back to session