Degree Name

Doctor of Philosophy


Faculty of Education


Children with specific language impairment are a group of children who experience severe difficulty learning and using language, despite the absence of any intellectual, hearing, vision or emotional difficulties (Beverly & Williams, 2004). Compared to children with normally developing language, children with specific language impairment have been shown to experience difficulty learning grammar, with many also believed to experience working memory deficits (Archibald & Gathercole, 2006; Gathercole & Baddeley, 1990; Montgomery, 2004). Cognitive load theory (CLT – Sweller, 2010a) is based on the assumption that human cognitive architecture plays a major part in learning, in particular when instructional design takes into account our limited working memory. A feature of cognitive load theory is the split-attention effect, which examines the increase in cognitive load experienced when two sources of information, which are necessary for learning to take place, are physically or temporally separated. Learners are, therefore, required to mentally integrate these sources of information in order to understand what is being taught. This mental integration places an increased load on the capacity of an already limited working memory, to such an extent that schema construction and automation may be impeded and learning becomes more difficult.

The three studies reported in this thesis were developed based on cognitive load theory and how this theory applies to children with specific language impairment in regular school settings. Traditionally, most students in Australian primary schools use a conventional, split-source format during grammar instruction rather than an integrated format. These experiments investigated whether children with specific language impairment could improve their identification of verbs found in a factual text by using an integrated format designed to reduce the split-attention effect. No time limit restrictions was given to students in Study 1 to complete the test phase but a time limit of fifteen minutes and eight minutes was imposed on both groups of students during the test phase for Studies 2 and 3 respectively. A mental effort rating scale was introduced to students in Studies 2 and 3 in an attempt to measure the mental effort of the students in both the conventional format and integrated format.

It was hypothesised that students in the integrated format would be more accurate in finding present-tense verbs than students in the conventional, split-source format for the test phase and they would respond more accurately in the post-test phase. Furthermore, in Study 2 and Study 3, it was hypothesised that the students in the conventional, split-source format would nominate a higher mental effort score than the students in the integrated format due to the effects of the split attention on a fivepoint subjective mental effort scale.

A one-way ANOVA design was used to analyse the results of the three studies. Study 1 (n = 18) showed no significant difference (p = 0.12) between the conventional (M = 47.78, SD = 21.67) and integrated formats (M = 66.67, SD = 27.39) in the test phase when there was no time limit imposed. Study 2 (n = 16) also showed no significant difference (p = - 0.0 9) between the two formats when a time limit of fifteen minutes was imposed (conventional: M = 36.25, SD = 29.25 and integrated: M = 60.0, SD = 23.30). A significant difference (p = 0.02, d = 0.49), however, was found in Study 3 (n = 20) when the time limit was restricted to eight minutes with integrated format (M = 44.0, SD = 33.39) significantly more accurate (p = 0.02, d = 1.22) than the conventional format (M = 14.0, SD = 15.78). The results for mental effort rating in Studies 1 and 2 were not so positive with students in neither condition showing significantly higher mental rating scores.

Overall, the results of these studies showed that cognitive load theory can make a difference to children with specific language impairment and can support their learning of grammar. The effect of split attention was not as large with children with SLI as might have been expected with normally developing children, especially when students were able to self-pace their responses. A more clear effect of split attention may have been observed with larger sample sizes and more rigorous selection criteria. Future research into CLT and children with SLI could examine other cognitive load effects in a range of learning domains.