Evaluation of the Crosslinguistic Nonword Repetition Test: Evidence From a Large and Diverse Secondary Data Set.

Abstract

Purpose: The aim of this study was to evaluate the crosslinguistic validity of the Crosslinguistic Nonword Repetition Test (CL-NWR) based on a large multicountry sample by investigating factors related to language ability, as well as potential confounds.

Method: The data consisted of CL-NWR scores from children aged 37-165 months, collected by 18 research teams across 15 countries. Item-level analysis was employed to examine any nondesirable effects of gender, socioeconomic status, bilingual status, and the amount of exposure to the test language, as well as desirable effects of age, item length, and clinical status (children categorized as typically developing [TD], with developmental language disorder [DLD], or with reported language concerns [LC], respectively). Subsamples were used to evaluate the consistency of findings across three time points and between different versions of the CL-NWR.

Results: Bayesian analysis provided strong evidence for the effects of age, item length, and clinical status on CL-NWR performance, as well as consistency across time points. In contrast, there was weak or no evidence for the effects of gender, socioeconomic status, bilingual status, amount of exposure, or test version. Additionally, there were two interactions between (a) item length and clinical status, suggesting that children with DLD found longer nonwords disproportionately more challenging than TD children, and (b) age and clinical status, with the gap between TD and LC groups narrowing with age.

Conclusions: The CL-NWR was unaffected by environmental and demographic factors that often influence language assessments, including some nonword repetition tests. Performance was driven by factors reflecting language abilities. This makes the CL-NWR a unique and valuable tool for language assessment contributing to the identification of DLD in diverse linguistic, social, and geographical contexts.