Very little is known about diagnosing specific language impairment (SLI) in children who are exposed daily to a dialect (community language) and a standard variety (school instruction). The research reported here examines the specificity and sensitivity of language tests used so far to evaluate language performance in the context of diglossia (Cyprus). Sixteen children with SLI aged 5-9 years and 22 age-matched typically developing children were examined on a range of language tests modified to include dialectal differences. Properties of each test were evaluated through logistic regression analysis in order to identify children with SLI. The analysis revealed that many of the tests used are sufficiently accurate concerning sensitivity and specificity levels. Furthermore, a combination of tests is proposed as a good tool for diagnostic purposes. Speech and language therapists as well as researchers can now rely on an accurate diagnostic procedure within a practice-based evidence framework.
Keywords: Bilectalism; Cyprus; Greek; SLI; diagnosis; diglossia.