Poster #33 - Genetic and Environmental Influences on the Triple Code Model of Mathematics

• In Event: 2-163 - Poster Session 08
  In Poster Session: PS 08 Section - Cognitive Processes

Integrative Statement

Interest in how individuals acquire math skills and conceptualize numbers has risen due, in part, to the economic and social implications of math achievement. An important model for math cognition is the Triple Code Model (TCM; Dehaene, 1992), which posits that people process numbers through three different systems: auditory verbal word frame (numbers as words), visual Arabic number form (numbers in Arabic format), and analog magnitude representation (abstract representation of numbers). The purpose of the present study is to investigate the TCM using a quantitative genetic twin study approach to identify 1) whether tests of math cognition corresponding to the TCM fit into the predicted components of the model and 2) whether the variation in achievement of the three factors of the TCM are due to genetic or environmental factors.

Participants (n=380) from the Western Reserve Reading and Math Project in early adolescence (age M = 12.3, SD = 1.25) were administered math tests whose demands corresponded with one of the three components of the TCM (see Table 1). Confirmatory factor analysis was conducted to assess the correspondence of the hypothesized factors. Then the contribution of genetics, shared environment, and non-shared environment to the variation in the individual component and to the covariance between the components of the TCM were assessed using a univariate and multivariate approach, respectively.

When the factors were arranged as described above and a covariance was added between fluency and rapid digit naming, the confirmatory factor analysis showed acceptable model fit (CFI=.971, TLI:.954, RMSEA = .058). The standardized estimates for the factor loadings ranged for auditory verbal word form (.18-.76), visual Arabic number form (.51-.80), and analog magnitude representation (.42-.83). The heritability statistics were moderate, and did not show significant differences between the three factors (see Table 1). In the multivariate analysis, the factors were entered in the order presented in Table 1. There was overlap in the genetic (range of a = .34-.58) and shared environmental etiology (range of c = .22-.48) of the three TCM factors. No additional variance was accounted for by an independent genetic nor shared environment component for analog magnitude representation. The nonshared environment component did not show significant overlap.

The genetic findings are unsurprising given the previous literature, but of interest is that the shared environment component accounted for a significant portion of the variance in the TCM factors; this is a unique finding compared to the reading literature, which does not show significant shared environment effects after children enter school. In addition, the significant genetic and shared environment overlap between the three factors emphasizes the common etiology between these different aspects of math performance. There appear to be aspects of the environments that twins share that account for the variance in performance of different aspects of numerical processing, though the specific aspects of the environment have yet to be identified. This study not only lends further credibility to the growing literature on the TCM but also emphasizes the genetic and shared environment links between its components.

Authors

• Susan Gross, The Ohio State University
  Presenting Author

• Carol Gross, Case Western Reserve University
  Non-Presenting Author

• Lee Thompson, Case Western Reserve University
  Non-Presenting Author

• Stephen Petrill, The Ohio State University
  Non-Presenting Author