Abstract
Purpose
We examined indices of narrative microstructure as metrics of language development and impairment in Arabic-speaking children. We examined their age sensitivity, correlations with standardized measures, and ability to differentiate children with average language and language impairment.
Method
We collected story narratives from 177 children (54.2% boys) between 3.08 and 10.92 years old (
M
= 6.25,
SD
= 1.67) divided into six age bands. Each child also received standardized measures of spoken language (Receptive and Expressive Vocabulary, Sentence Imitation, and Pseudoword Repetition). Several narrative indices of microstructure were examined in each age band. Children were divided into (suspected) developmental language disorder and typical language groups using the standardized test scores and compared on the narrative indicators. Sensitivity and specificity of the narrative indicators that showed group differences were calculated.
Results
The measures that showed age sensitivity included subject omission error rate, number of object clitics, correct use of subject–verb agreement, and mean length of utterance in words. The developmental language disorder group scored higher on subject omission errors (Cohen's
d
= 0.55) and lower on correct use of subject–verb agreement (Cohen's
d
= 0.48) than the typical language group. The threshold for impaired performance with the highest combination of specificity and sensitivity was 35th percentile.
Conclusions
Several indices of narrative microstructure appear to be valid metrics for documenting language development in children acquiring Gulf Arabic. Subject omission errors and correct use of subject–verb agreement differentiate children with typical and atypical levels of language development.