Can it be too late?

Usually, my applied work focuses on limiting OVER-identification (false positives) of language impairment in children who speak English as a second language. But, there’s another side to this too and that’s UNDER-identification of language impairment in this same population. I see these patterns in some of the national schools data. From year to year sometimes it looks like preschool English language learners (ELLs) are less likely than average to be identified with language impairment while school-age ELL children are more likely to be diagnosed with language impairment.