Using Nonword Repetition to Identify Developmental Language Disorder in Monolingual and Bilingual Children: A Systematic Review and Meta-Analysis
Journal of Speech, Language, and Hearing Research
Schwob, S., Eddé, L., et al. (2021).
Journal of Speech, Language, and Hearing Research, 64(9), 3578-3593.
This systematic review and meta-analysis investigates the sensitivity and specificity of using nonword repetition tasks as a clinical indicator of developmental language disorders in children, younger than 9 years old, with developmental language disorder or children with late language emergence.
Not stated
2000 to June 2020
English, French, or German publications that included data; excluding systematic reviews and meta-analyses
46 studies included in the systematic review; 35 studies included in the meta-analysis
<div>Studies investigating the discriminative value of nonword repetition tasks (NWRT) for identifying developmental language disorder (DLD) in children report sensitivity ranging from 40% to 100% and specificity from 48% to 100%. While the majority of studies find NWRT to be diagnostically accurate with sensitivity and specificity between 80% and 100%, 14 studies reported at least one diagnostically inaccurate result. This review found that, in general, NWRT performance correlates to other oral assessment measures (e.g., vocabulary, morphosyntax, early developmental indices); however, "no firm generalizations can be made regarding the link between NWRT and other verbal and nonverbal measures" (p. 3588). The authors conclude that "combining NWRT with other tasks can be an effective means to increase diagnostic accuracy, especially in situations where the discriminative value NWRT is not sufficient on its own" (p. 3588).</div>
<div>Studies investigating the nonword repetition task (NWRT) scoring method used found that both percentage of items correct (PIC) and percentage of phonemes correct (PPC) methods discriminated between children with DLD and children with typical language development; however, PIC demonstrated better sensitivity (71% to 100%) compared to PPC (63% to 94%). PIC scoring was found to be more accurate for monolingual children, while PIC and PPC demonstrated equal accuracy for bilingual children.</div>
<div>For bilingual children, the sensitivity and specificity of nonword repetition tasks (NWRT) for discriminating developmental language disorder can improve depending on the language used when conducting NWRT. The authors found that it is "not possible to conclude whether it is better to use an NWRT in the language of the society and/or family.... In order to obtain the best diagnostic accuracy, it, therefore, seems appropriate to carry out NWRTs in both of the child’s languages or to use a quasi-universal task" (pp. 3587-3588).</div>
