Abstract
Background: Narratives have been used by a number of researchers to investigate the language of children with specific language impairment (SLI). While a number of explanations for SLI have been proposed, there is now mounting evidence that children with SLI have limited memory resources. Phonological memory has been the focus of the research on memory with this population. However, the use of narrative tasks to investigate memory limitations in SLI has not previously been undertaken.
Aims: The aims of the research were to investigate the narrative and memory abilities of 6‐year‐old children with SLI and the association between narrative skills and memory.
Methods & Procedures: Two studies were conducted. In Study 1 the performance of the children with SLI was compared with that of their peers with typical language development (AM), and to that of a younger group (LM) matched on expressive language (about 2 years younger). Children were asked to recall (‘story recall’) and show comprehension of a narrative they had been told, and also to tell a story based on a series of pictures (‘story generation’), to recall their story and to answer comprehension questions about it. In Study 2 the children with SLI from Study 1 and the AM children from Study 1 were tested on four working memory tasks: word and digit span, the Recalling Sentences task from the Clinical Evaluation of Language Fundamentals — Third Edition, and a dual‐processing task. An inhibition task and an attention task were also included. It was predicted that children with SLI would perform more poorly than the AM group on measures of narrative telling, but comparably with the LM group. It was also predicted that the SLI group would perform more poorly than the AM group on both literal and inferencing comprehension questions but comparably with the LM group. The third prediction was that there would be significant associations between children's performance on the narrative tasks and the tasks measuring working memory.
Outcomes & Results: In Study 1 the children with SLI performed more like the younger group on recall of a narrative they had heard, but more like their peers when asked to generate and recall their own narrative based on a series of pictures. The children with SLI had difficulty with inferencing questions. In Study 2 the impaired group performed significantly worse on measures of memory, showing a lower working memory capacity. The children with SLI made more errors on the attention task, but no group differences were found on the inhibition task. In comparing results from Studies 1 and 2, significant correlations were found between performance on the narrative tasks and memory tasks, but the contribution of memory to the narrative task scores differed for the two stories. The Recalling Sentences memory task was found to be the best overall measure for predicting variance on story comprehension and recall.
Conclusions: In support of previous research, the children with SLI showed problems with inferencing, linking directly observed or stated information to likely outcomes. They also showed a limited working memory capacity, and they were more likely to make errors in attention. A main finding was that the narrative abilities of the 6‐year‐old children were linked to their verbal working memory. The information the SLI group heard was harder to access than information they had been able to generate themselves based on a series of pictures. The findings suggest that children with SLI are likely to be at a disadvantage in classroom situations, particularly for information presented aurally and if the information is complex. The use of pictorial aids may help them encode the information. They would also benefit from having information broken into manageable units.
Notes
1. None of the children had been previously identified as SLI. Only two of the children identified as SLI in the study were known to be receiving any intervention from a speech pathologist.
2. Culatta et al. (Citation1983) list 24 propositions. However, there is some discrepancy, e.g. ‘He was disappointed because he didn't get a puppy’ is counted as one, but ‘Tommy was really happy because he got the present he wanted’ is counted as two. A total of 22 propositions were identified. A third inferencing question was also added to allow for more variability in scores. In addition, ‘baseball bat’ was changed to ‘cricket bat’ to be more culturally appropriate for the children tested.
3. Transcription and scoring for the narrative tasks were completed before the scoring for the screening tasks, that is before the SLI participants were identified A second scorer blind to the language status of the children checked the coding for the narrative tasks; there was no disagreement between coders.
4. Raw scores were used in the analysis to allow for comparison between the SLI and the younger LM children since ERRNI standard scores are not provided for 4–5 year olds.
5. For partial η2, 0.01 is considered to be a small effect size, 0.06 a medium effect size, and 0.140 a large effect size (Cohen Citation1988).
6. Attention Deficit Hyperactivity Disorder (ADHD) has been frequently found to be co‐morbid with language problems; as reported by Cohen et al. (Citation2000), working memory problems may link the two disorders.