The SNARC effect – Does it depend on the level of mathematical training?

Behavioural studies show a relation between numbers and space (DeHevia et al., 2008). One instance of this link is the SNARC (Spatial Numerical Association of Response Codes) effect, consisting in faster reaction times responding to small/large digits with the left/right hand respectively (Dehaene et al. 1993). The SNARC effect has often been replicated, but it is also characterized by high inter-subject variability (Wood et al. 2006a,b). Although differences in mathematical skills are an obvious candidate source for SNARC variability, this variable has not yet been explored systematically. While in their seminal study Dehaene and colleagues had included two groups of participants differing in their field of study (science vs. literature), the differences in SNARC effect strengths indicated by the result figures were not tested formally (see also Fischer and Rottmann (2005)). The present study aims to assess the influence of mathematical training, as instantiated by the enrollment in study fields characterized by distinct levels of mathematical requirements, on the SNARC effect. We only included students of either a subject with a strong mathematical focus (e.g. mathematics, computer sciences), or no mathematical requirements at all (e.g. literature, philosophy). Results (N=36; 18/group) yielded a significant overall SNARC effect, i.e. interaction between digit magnitude and response side [F(1,35)=9.51; p=0.004], but no influence of study group on this interaction [F(1,35)=0.92; p=0.345]. Interestingly though, including gender in the ANOVA yielded a significant four-way interaction [F(1,35)=6.34; p=0.017]. These results replicate previous findings in children (Schweiter et al. 2005), interpreted as gender-related differences in mental strategy use.