Concurrent and Longitudinal Relations Between Math Anxiety, Working Memory, and Math Achievement During Childhood

• In Event: 2-023 - Identifying Important Concurrent and Longitudinal Correlates of Math Anxiety in Elementary School Children

Integrative Statement

Greater math anxiety is associated with poorer math achievement during childhood (Ganley & McGraw, 2016; Gunderson et al., 2018). In parallel to findings on adults (Beilock, 2008), the relation between math anxiety and math achievement appears to be particularly pronounced for children with high working memory (Ramirez et al., 2013; Ramirez et al., 2016), suggesting that math anxiety hinders math performance by disrupting children’s working memory processes. However, some studies have found no moderating effects of working memory in children (Lauer, Esposito, & Bauer, 2018) or have reported that working memory moderates the math anxiety-achievement link when it is measured longitudinally, but not concurrently (Vukovic et al., 2013). A potential source of these discrepancies is the variable use of verbal and spatial working measures to evaluate working memory across studies. The present research therefore sought to elucidate the specific roles of verbal and spatial working memory as moderators of the math anxiety-achievement relation when measured both concurrently and longitudinally across two time points in elementary school.

In Year 1 of the study, 272 children (152 girls) in first- through fourth-grade completed self-report measures of math, spatial, and verbal anxiety (Lauer et al., 2018; Ramirez et al., 2012; Ramirez et al., 2013) and standardized math, spatial, and reading ability assessments. One year later, children again completed the anxiety scales and standardized assessments from Year 1 as well as spatial and verbal working memory tasks. Gender, grade, spatial anxiety and ability, and verbal anxiety and ability were included as covariates in all analyses to control for domain-general influences on the anxiety-achievement link.

We first examined concurrent relations at each time point, finding that math anxiety negatively correlated with math achievement in both Years 1 & 2 (prs<-.15, ps<.02) and that this concurrent relation was moderated by verbal working memory (p=.026), but not spatial working memory (p=.326; Figure 1). We next conducted a longitudinal path analysis using the “lavaan” package in R (see Figure 2), finding that lower math achievement in Year 1 predicted greater math anxiety in Year 2, but math anxiety in Year 1 was not associated with math achievement in Year 2 (Figure 2). Importantly, however, moderator analyses demonstrated that higher math anxiety in Year 1 indeed predicted lower math achievement in Year 2 for children with greater verbal working memory (b=-.11, p<.05). Spatial working memory had no moderating effect on this longitudinal relation (b=.08, p>.20; see Figure 1).

The present research is the first to examine the respective roles of verbal and spatial working memory in the association between children’s math anxiety and their math achievement. Verbal but not spatial working memory was found to moderate the math anxiety-performance relation both concurrently and longitudinally. These findings support previous contentions that math anxiety impedes math performance by limiting the use of working memory-demanding problem-solving strategies (e.g., Ramirez et al., 2016), leaving high-working memory children more susceptible to the detrimental effects of math anxiety over time and suggesting such children may particularly benefit from interventions designed to reduce math anxiety.

Authors

• Jillian E Lauer, New York University
  Presenting Author

• Alena G Esposito, Clark University
  Non-Presenting Author

• Patricia Bauer, Emory University
  Non-Presenting Author