Yes, no, maybe, it depends.
When we do bilingual testing to determine whether there is a language impairment or speech disorder we are trying to get at a child’s capacity for speech and language learning. So, if we test in a bilingual child’s weaker language we may underestimate their language ability and unnecessarily identify them with a speech or language impairment.
What we’ve found in our work with preschool age children and school age children is that they often have mixed dominance. Kids in the “middle” of the bilingual continuum– that is kids who have at least 30% exposure to each language (30-70, 40-60, 50-50) will have mixed performance. They may do better in English vocabulary but not English grammar. On grammar they may be better in the home language. In these cases it is absolutely critical that we test in both languages.
Sometimes we really don’t know the pattern of children’s language exposure. They may not be very good at judging their own expertise. If you can’t talk to both the parents and teachers to gain this information, it’s best to test in both and then compare the two. We have found that often teachers and parents focus on different aspects of language. And that makes sense, this is why if possible it’s important to have information from both to put it all together.