CogAT is pretty decent for what it is: a quick and dirty group screener. It's reasonably accurate most of the time. BUT - - according to the authors themselves, it was never intended to be used for identifying giftedness by itself. It measures - call it "learned abilities". It's not quite an achievement test, but it's not an ability/ IQ test either. The authors explain that where the errors are mostly found is for:
(1) Kids above the 90th percentile. And the further they get from the norm, the more likely they are to provide correct but not standard answers - which are marked wrong. (In a WISC, the testing is done one-on-one, and the psych can take the time needed to probe for rationale, giving marks for well-justified but unusual/ divergent responses)
(2) Kids who do not come from an enriched early environment (since it mostly measures things that have been learned), and/ or English language and culture (since it is a written test)
(3) 2E kids (again, because it involves reading (after grade 2) and writing, in a (potentially distracting) group setting, without the rapport, and the kinds of interactions and supports that are inherently built-in to a one-on-one oral test.
So if the CogAT results make sense, great. But if your gut and the CogAT don't agree, trust your gut.
