Language Correlates of Achievement in Children with Math Difficulties

Many factors are believed to predict math performance. One of such factors is language, and yet the relations between certain relevant language variables and specific math outcomes have not been fully studied. This project analyzed the differential significance of vocabulary and verbal fluency on math and word problem solving in 19 school-aged children with and without spina bifida. It was expected that verbal fluency would be a stronger unique predictor than vocabulary of math fluency, and conversely, that vocabulary would be a stronger unique predictor than verbal fluency of word problem solving. Both these relations were expected to hold even in the context of known strong predictors of math performance (e.g., number line estimation). Preliminary analysis included an evaluation of variable distributions, group structure, and demographics. For analyzing hypotheses, we used SAS (Copyright ©2012 SAS Institute Inc.) to find zero order correlations and partial correlations controlling for age and number line estimation as a measure of unique effect size. Although the predictors had weak to moderate zero-order effects on the outcomes (with r values between 0.01 and 0.37, p < 0.966), these relations lost significance in the context of partial correlations (with r values between -0.20 and 0.24, p < 0.531). Results suggest that word knowledge and the ability to recall verbal facts from memory are of little importance in predicting math performance in children, whether in terms of purely arithmetic or word problems, while accounting for the effect of strong known covariates of these outcomes.