Introduction: High maths anxious people may suffer from a reduction in working memory (WM) due to ruminations and disturbing thoughts (Processing efficiency theory, PET; Eysenck & Calvo, 1992). Studies by Ramirez et al.(2013) and Vukovic et al.(2013) supported this hypothesis, but found that a negative association between maths anxiety (MA) and maths achievement is present only for the high WM group. Miller & Bichsel (2004) showed that maths anxiety primarily impacts visual-spatial component of working memory (VSWM). However, the effect of VSWM on the link between math anxiety and mathematical performance has never been tested. The current study examined the mediation and moderation effects of VSWM on the association between MA and mathematical performance.
Methods: 136 people (63 pairs of twins and 10 singletons) participated in the study. Mathematical performance was measured with Problem Verification Task (PVT, Murphy & Mazzocco, 2008) and Number Series Test (NST, Smith, Fernandes, Strand 2001). Spatial ability was measured with Mental rotation task (MRT, Vandenberg & Kuse, 1978). Maths anxiety was measured with sMARS questionnaire (Alexander & Martray, 1989) and visual spatial working memory (VSWM) was assessed with Corsi block tapping test (Pagulayan, K.F. et al., 2006). Intelligence (general cognitive ability) was measured with Raven’s Matrices test (Raven, Raven, & Court, 1998). Hierarchical multilevel regression analyses was used in order to account for the biased standard errors arising from twins nonindependence (Atkins, 2005; Carlin et al., 2005). We examined moderation and mediation effect of VSWM on relations between maths anxiety and maths abilities, after accounting for general cognitive ability and Spatial ability.
Results: VSWM did not mediate the association between MA and mathematical performance. However, VSWM moderated the association between them and reduced the negative effect of MA on mathematical performance. In the model with interaction effect the main effect of MA remained significant and negative for NST (B=-.22, z=-3.24, p<0.01) and non-significant for PVT. The interaction between MA and VSWM was significant and positive in both NST and PVT(B=.14, z=2.38 , p<0.05 and B=.13, z=2.18, p<0.05, respectively). Simple slope analysis showed that higher MA was associated with reduced accuracy in NST, but only in people with low and moderate VSWM. The effect wasn’t present for people with high VSWM. For PVT, higher MA was associated with decreased accuracy in this test for people with low VSWM only. The effect was non-significant for people with moderate and high VSWM.
Conclusion: Results are discussed in the framework of processing efficiency theory (PET; Eysenck & Calvo, 1992) and Attentional control theory (ACT; Eysenck et al., 2007). Contrary to the results by Ramirez et al. (2013) and Vulovic et al. (2013), who found the significant effect of MA on mathematical performance for higher WM, our data shows the effects of MA only for lower VSWM group. However, in the present study the visual spatial component of WM was studied, while previous studies examined other WM components. Differences in the math tasks may also be the reason for inconsistent results across studies.
Education , Cognition and Attention , Group differences
