How and why do number-space associations co-vary in implicit and explicit magnitude processing tasks?

Abstract

Evidence for number-space associations in implicit and explicit magnitude processing tasks comes from the parity and magnitude SNARC effect respectively. Different spatial accounts were suggested to underlie these spatial-numerical associations (SNAs) with some inconsistencies in the literature. To determine whether the parity and magnitude SNAs arise from a single predominant account or taskdependent coding mechanisms, we adopted an individual differences approach to study their correlation and the extent of their association with arithmetic performance, spatial visualization ability and visualization profile. Additionally, we performed moderation analyses to determine whether the relation between these SNAs depended on individual differences in those cognitive factors. The parity and magnitude SNAs did not correlate and were differentially predicted by arithmetic performance and visualization profile respectively. These variables, however, also moderated the relation between the SNAs. While positive correlations were observed in object-visualizers with lower arithmetic performances, correlations were negative in spatial-visualizers with higher arithmetic performances. This suggests the predominance of a single account for both implicit and explicit SNAs in the two types of visualizers. However, the spatial nature of the account differs between object-and spatial-visualizers. No relation occurred in mixed-visualizers, indicating the activation of taskdependent coding mechanisms. Individual differences in arithmetic performance and visualization profile thus determined whether SNAs in implicit and explicit tasks co-varied and supposedly relied on similar or unrelated spatial coding mechanisms. This explains some inconsistencies in the literature regarding SNAs and highlights the usefulness of moderation analyses for understanding how the relation between different numerical concepts varies between individuals.

The influence of semantic processing and response latency on the SNARC effect

“…The relationship between the slope distributions of the different tasks were further investigated with Pearson's correlation coefficient (Table 3 and Figure 3). In line with previous studies (ρ = 0.08 in Georges, Hoffmann, & Schiltz, 2017 and ρ = 0.2 in Georges, Hoffmann, & Schiltz, 2018), the magnitude classification and the parity judgement tasks did not correlate. The only significant positive correlation emerged between the magnitude classification and phoneme detection tasks.…”

The SNARC and MARC effects measured online: Large-scale assessment methods in flexible cognitive effects

“…1, respectively, for unstandardized slopes). However, in some in-lab studies reliabilities have been comparable to those from the online study (e.g., .55 in Georges et al, 2017).…”

“…Although many studies on interindividual differences in the SNARC have been conducted, the studies on those differences were frequently not replicated, in line with current general concerns about reproducibility in psychology (in line with current general concerns about reproducibility in psychology; e.g., Open Science Collaboration, 2015), or were largely inconclusive. Possible reasons include that the reported correlations or between-group differences have mostly been small to moderate (in most cases between .25 to .44; e.g., Hoffmann, Pigat, & Schiltz, 2014b); in several instances, the reliability of the SNARC was relatively low (e.g., Georges, Hoffmann, & Schiltz, 2017); and the tested samples have been relatively small (yielding insufficient power). Namely, the sample sizes in published SNARC studies typically have not exceeded 100 participants.…”

“…The reliability of standardized slopes, calculated according to the method used here, is reported only in Cipora et al (2016), and it was .75. The reliability of standardized slopes, calculated in a slightly different way, as reported by Georges et al (2017) equaled .78.…”

On the Multitude of Mathematics Skills: Spatial-Numerical Associations and Geometry Skill?

“…Namely, a stronger effect of mental rotation (i.e., larger increase in the time to decide on object identity depending on the degree to which it is rotated) relates to more pronounced SNARC (Viarouge, Hubbard, & McCandliss, 2014). In a more recent study it was shown that an individual's visualization profile moderates the relationship between magnitude and parity SNARC (Georges, Hoffmann, & Schiltz, 2017a). Importantly, visuospatial skills themselves are also correlated with different types of math skill at several stages of development (Mix et al, 2016).…”