Metacognitive Support for Mathematics Learning in Children with Autism Spectrum Disorder (ASD)

Background:

Metacognition is awareness of one’s own cognition, sometimes termed ‘learning how to learn’. One of the best predictors of mathematical achievement is metacognition, predicting outcomes better than IQ (Iacalano et al., 2014). There are different aspects of metacognition that underpin learning, such as monitoring when or where mistakes are made (metacognitive monitoring) and adjusting learning strategies accordingly (metacognitive regulation).  Metacognitive tuition enhances mathematics learning among individuals within and below the ‘normal’ range of mathematical ability (Maxwell et al., 2014).

On average, mathematics ability is substantially lower among people with Autism Spectrum Disorder (ASD) than would be expected on the basis of IQ (Mayes & Calhoun, 2003; 2006). Previous research has shown that those with ASD are likely to make two distinct metacognitive monitoring errors when learning mathematical skills. Firstly, they are more likely than children without ASD to think that they have an answer correct when in fact they have answered it wrongly. Secondly, when informed that they have made an error, those with ASD are more likely to report that they meant to make the error (Brosnan et al., 2015; see also Williams & Happé., 2010). This indicates that learners with ASD need specific support for metacognitive monitoring (error awareness, judgments, intentions) in order to effectively adapt their learning strategies (metacognitive regulation).

Objectives:

The aim of the current research was to test computer-based metacognitive support for learners with ASD.

Methods:

Participants: 40 participants with ASD (mean age 13.3 years) and age- and ability-matched typical comparisons.

We developed a “maths challenge” computer program whereby, to maximise points won, participants needed to monitor their performance and adapt their strategy accordingly. The program comprised 7 levels of difficulty and each question answered correctly was worth points commensurate to that level. After each question, participants were asked metacognitive monitoring questions concerning whether they thought they had got the answer correct and whether they had meant to get the answer correct. Participants answered 4 blocks of 3 questions, beginning at level 4. After each block, participants were given the choice of whether to stay at the same level, move up a level to harder questions (worth more points if answered correctly), or move down a level (worth fewer points but easier) for the next bock until they had answered 12 questions in total. In the metacognitive support condition, continuous feedback was provided concerning the accuracy of each answer, and reinforcing the goal. This was absent in the ‘no support’ condition.

Results:

Participants with ASD demonstrated diminished metacognitive monitoring and they made more post-test intention errors (i.e., incorrectly attributing errors as intended). However, when provided with feedback about their performance and goal reminders, learners with ASD won more points by the end of the program than when no metacognitive support was provided (p <.05, d =3.36).

Conclusions:

Children with ASD employed more effective learning strategies when provided with the metacognitive support condition of the Maths Challenge program. Findings highlight the potential for metacognitive support to improve mathematics learning in classrooms.