Bilingual Language Assessment: A Meta-Analysis of Diagnostic Accuracy
Journal of Speech, Language & Hearing Research
Dollaghan, C. A., & Horner, E. A. (2011).
Journal of Speech, Language & Hearing Research, 54(4), 1077-1088.
This meta-analysis investigates the diagnostic accuracy of assessment tools for diagnosing language impairment in bilingual Spanish-English children based in the United States.
Not stated
From 1949-1950 through July 2009
Not specified
9
<div>Findings demonstrated the positive likelihood ratios (LR+) for most bilingual assessment measures were suggestive (LR+= 4.12; 95% CI=2.94, 5.78), indicating that additional testing would be necessary to confidently confirm the presence or absence of a disorder. Only finite verb marking did not meet or exceed the level to be deemed as suggestive based on LR+. Based of LR+, six measures (i.e., formal word definition, Spanish Morphosyntax Test ages 4 and 5, invented morpheme, parent report plus grammatical errors per T-unit, and parent report plus grammatical errors per T-unit and positive family history) met or exceeded the level deemed clinically informative, meaning a person who scores in the affected range is very likely to have the disorder. Negative likelihood ratios (LR-) were noted to be more variable, with heterogeneity making the validity of the overall pooled LR- doubtful (LR-= 0 0.22, 95% CI= 0.11 to 0.46). Overall, no measure stands out as the most effective method for identifying language impairment and typical language in bilingual Spanish-English children. Further research is indicated.</div>
