Association of Cognitive Factors and Anxiety with Math Achievement in Adolescents with ASD

Friday, May 15, 2015: 2:09 PM
Grand Ballroom A (Grand America Hotel)
J. Beck1, T. Oswald1, A. M. Iosif1, J. C. Matter1, L. Gilhooly1 and M. Solomon2, (1)UC Davis MIND Institute, Sacramento, CA, (2)Department of Psychiatry & Behavioral Sciences, University of California-Davis, Sacramento, CA, Sacramento, CA
Background: Mathematics achievement in adolescents with ASD has been understudied, with most research focusing on young children and arithmetic calculation. The paucity of research on this topic may be partly explained by the widely-held belief that most individuals with ASD are mathematically gifted, despite evidence to the contrary (Dickerson, Mayes & Calhoun, 2008). Solving applied math problems helps develop quantitative reasoning skills which are associated with both academic and everyday problem-solving abilities. Thus, it is critical to examine the ability of students with ASD to solve applied math problems and to identify those factors that influence achievement on such problems.

Objectives: (1) Determine proportions of the sample demonstrating giftedness or disability on applied math problems. (2) Examine which aspects of cognition (i.e., fluid intelligence, verbal ability, working memory) and anxiety (i.e., test anxiety) best predict achievement on applied math problems in ASD relative to a typically-developing control group.

Methods: The sample consisted of 27 high-functioning (FSIQ > 80) adolescents with ASD and 27 age- and FSIQ-matched typically-developing controls (combined-group means: age 14.8 yr.; FSIQ 102.8). Math achievement was evaluated by the untimed WIAT-III Math Problem Solving subtest.  We utilized 1.5 SD above/below the mean cutoffs to classify children as having performance indicative of mathematical giftedness or a math disability.  To explore factors that uniquely predict math achievement in ASD, we entered cognitive and anxiety variables as predictors in a linear regression model of grade-normed math achievement. The predictors were fluid intelligence and verbal ability, working memory, and test anxiety, as measured by the WASI-II, WRAML2, and BASC-2 respectively. Regression commonality analysis was used to illuminate how predictors shared variance.

Results: 22% of the ASD group was classified as evidencing a math disability, with only 4% showing giftedness.  Compared to the typically-developing group, the ASD group contained a significantly greater proportion of participants whose math achievement indicated a math disability (χ² = 6.75, p ≤ .01). The final most parsimonious linear regression model (R2adj = .54) showed that the strongest predictor of math achievement was fluid intelligence (β = .39), followed by verbal ability (β = .30), test anxiety (β = -.29), and diagnosis of ASD (β= -.26). All interaction terms between diagnosis and significant predictors were non-significant. This model suggests that, even after accounting for other significant factors, a diagnosis of ASD was associated with an 8.3 point decrease in math achievement standard score.

Conclusions: While a majority of our sample of high-functioning adolescents with ASD demonstrated average math achievement on applied problems, a significant minority had achievement indicative of a math disability. This finding adds to the body of research suggesting that math is not an area of strength for the majority of students with ASD. In fact, an ASD diagnosis significantly predicted poorer math achievement; although, cognitive abilities and text anxiety were even stronger predictors of math achievement. These results inform our theories of math ability in ASD and highlight possible targets of intervention (i.e. test anxiety reduction) for students with ASD struggling with mathematics.