Source URL: View this document on the ASHA website

Authors conducted a meta-analysis of diagnostic studies for language impairment in bilingual children. They found a serious lack of necessary psychometric measures in the vast majority of studies examined.

Background: The purpose of this study was to describe quality indicators to be used in determining the value of studies of diagnostic accuracy. Additionally, the authors sought to conduct a meta-analysis of a variety of diagnostic tools available for diagnosing language impairment (LI) in bilingual Spanish-English U.S. children.

Hypotheses: N/A

Methodology and Participants: A collection of peer-reviewed English-language studies was compiled and analyzed for a description of appropriate psychometric measures and accuracy in differentiating between bilingual typically developing (TD) (n=109) and language impaired children (n=100).

Conclusion: Dollaghan and Horner found that most publications did not include information about the reference standards or procedures, making it difficult to rate the studies. Results of analysis revealed that the diagnostic tools investigated in the studies produced likelihood ratios that were at most “suggestive,” and none were determined to be diagnostically informative. In conclusion, the authors suggested improvements in the design and implementation of the studies, including ensuring that measures are taken to prevent subjective bias during testing.

Relevance to the Field: This article exposed the dearth of useful research being published about appropriate assessment for language impairment in bilingual children. The vast majority of diagnostic studies reviewed contained serious methodological or other flaws preventing them from being very practical to the clinician. As law and policy requires the use of assessments that are “valid, reliable and free of bias” (IDEA, 2004), it is important that the field of speech language pathology produce more quality research to support evidence based assessments and interventions. Clinicians should be familiar with the language of validity (e.g., reference standard, likelihood ratios, specificity and sensitivity, accuracy) in order to be able to examine the assessments they use and ensure valid and ethical practice for the individuals with whom they work.

Dollaghan, C. A., & Horner, E. A. (2011). Bilingual language assessment: A meta-analysis of diagnostic accuracy. Journal of Speech, Language, and Hearing Research54, 1077-1088.