Pattern understanding is a predictor of early reading and arithmetic skills

• In Event: 3-080 - Beyond Numeracy: Varied Predictors of Mathematics Knowledge

Integrative Statement

Measures of pattern understanding correlate concurrently with measures of arithmetic (e.g. Lee et al., 2011, 2012) and reading (e.g. Pasnak et al., 2016). Moreover, training children to identify patterns appears to support reading and arithmetic development (e.g. Kidd et al., 2014). However much of this evidence is complicated by the use of pattern tasks which rely on alphanumeric knowledge. Furthermore, only one study provides longitudinal evidence that patterning predicts later arithmetic (Rittle-Johnson et al., 2016). In this study, we examine pattern understanding as a longitudinal predictor of arithmetic and reading beyond the contribution of a wide range of theoretically critical cognitive skills. We predicted that pattern understanding would predict both arithmetic and reading, but the relationship would be stronger for alphanumeric (i.e. letters and numbers) than non-alphanumeric patterns. We also predicted that this relationship might largely be explained by other predictors such as EF.

Methods
A large sample of children (N=569) from 11 primary schools in Brisbane, Australia, were assessed three times at approximately six-monthly intervals from the outset of formal schooling (Mage = 5:03). Children were assessed on a wide range of numeracy (counting, number knowledge, magnitude comparison, and early arithmetic) and reading (word reading, letter-sound knowledge, phonological awareness and RAN) measures, tests of EF (working memory, attention, inhibition and visual short-term memory) and patterning. Three patterning tasks asked children to select one of four answers to complete patterns consisting of letters, numbers, and colours, objects, and shapes.

Results
Latent variable path models were constructed to examine a wide range of putative cognitive foundations for arithmetic and reading. In these models, alphanumeric patterning (letters and numbers) was separated from non-alphanumeric patterning (objects, colours and shapes). When both patterning variables were included in the model (Figure 2.1), only alphanumeric pattern understanding predicted both reading and arithmetic. To identify whether non-alphanumeric patterning was an important predictor of reading and arithmetic on its own we dropped alphanumeric patterning from the model (Figure 2.2). In this model, reading is predicted by early word reading, phonological awareness, RAN colours and non-alphanumeric pattern understanding. Arithmetic is predicted by addition, number knowledge and non-alphanumeric pattern understanding. Though EF has no direct effect on arithmetic or on reading there are significant indirect effects (EF to non-alphanumeric pattern understanding to arithmetic/reading). Overall, the model in Figure 2.2 accounts for 65% of the variance in reading and 72% of the variance in arithmetic.

Conclusion
These models show that pattern understanding is a predictor of both arithmetic and reading ability after accounting for a range of other potential predictors. This relationship cannot be explained by alphanumeric knowledge since non-alphanumeric pattern understanding is a significant predictor of reading and arithmetic on its own. The finding that pattern understanding is predicted by EF, and that EF indirectly predicts reading and arithmetic, suggests the relationships between pattern understanding, reading and arithmetic may be at least partially explained by EF.

Authors

• Kelly Burgoyne, University of Manchester, Australian Catholic University
  Presenting Author

• Stephanie Malone, Australian Catholic University
  Non-Presenting Author

• Arne Lervag, University of Oslo
  Non-Presenting Author

• Charles Hulme, University of Oxford
  Non-Presenting Author