Year: 2007
Type: article
Abstract: Integer Comparison and the Inverse Symbolic Distance Effect Sashank Varma ([email protected]) Daniel L. Schwartz ([email protected]) Robb Lindgren ([email protected]) Janet Go ([email protected]) Stanford Center for Innovations in Learning, Stanford University 450 Serra Mall, Building 160, Stanford, CA 94305-2055 Surprisingly, mixed comparisons showed an inverse SDE, with far comparisons slower than near comparisons (t(20)=4.94, p .93). A natural interpretation is that participants used rules (e.g., “positives are greater than zero”), not magnitude processing. To test this interpretation, we conducted a multivariate analysis with two repeated measures, valence (i.e., the sign of the non-zero number) and distance. The interaction was reliable (F(1,20)=6.57, p<.02), as shown in Figure 2. Positive-valence comparisons show an SDE and negative-valence comparisons an inverse SDE. Keywords: Symbolic distance effect; mental representation. Introduction The positive integers constitute the most well-studied number class. An important finding is the symbolic distance effect (SDE): the greater the distance between two positive integers, the faster they are compared (Moyer & Landauer, 1967). The SDE is important because it indicates that numerical symbols are understood in part as magnitudes, i.e., using a mental number line. The current study investigated whether the SDE holds for all integers – positive, negative, and zero. Method 21 participants were recruited from the Stanford University community. Two repeated measures were varied orthogonally. Distance had two levels: far and near. Comparison type had four levels: positive (x vs. y), negative (-x vs. -y), mixed (x vs. -y), and zero (x vs. 0 (positive valence) and -y vs. 0 (negative valence)). The dependent variable was response time on correct trials Figure 2: Valence effect. Discussion Results This study investigated the mental representation of integers. The results suggest that integers partition mentally into two classes, non-negative and negative. Comparisons within the same class show an SDE. This is consistent with conventional magnitude processing, i.e., a conventional mental number line stretching from –∞ to ∞. By contrast, comparisons across classes (i.e., a negative integer to either a positive integer or zero) show an inverse SDE. This is inconsistent with conventional magnitude processing. We are developing a new (and unconventional) mathematical model to account for these results. A multivariate analysis revealed reliable main effects of comparison type (F(3,18)=69.79, p<.001) and distance (F(1,18)=10.51, p<.005) and a reliable interaction (F(3,18)=15.13, p<.001). Means, standard errors, and sample comparisons are shown in Figure 1. Positive comparisons showed an SDE (t(20)=4.33, p<.001), suggesting use of magnitude processing (i.e., a mental number line). Negative comparisons showed an SDE (t(20)= 4.43, p<.001), also suggesting use of magnitude processing. Acknowledgments This research was supported by the National Science Foundation under grants REC-0337715 and SLC-0354453. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation Reference Moyer, R. S., & Landauer, T. K. (1967). The time required for judgments of numerical inequality. Nature, 215, 1519. Figure 1: Symbolic distance effects.
Language: English
Subfield: Developmental and Educational Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
Funder: National Science Foundation
OpenAlex ID: W2767983781
