I think it's important to realize that not all kids who score at very high levels on such tests have been test prepped or formally taught higher level material. To me, that belief feels a bit like what the schools always tell me -- that it can't be possible for my kids to know what they know unless I'm some crazy nut making my kids do math 3 hours a day.
Given internet access, the ability to type searches, and really good reading ability, there's not much a kid can't learn if motivated. Some kids are. For a kid who is curious, one thing leads to another and the seeking of information in one area leads to finding information in another area, which leads to new terms to search, etc. Some kids do their own test prep by self-teaching all the time, for the joy of it, not because they are supposed to or are taught a higher level curriculum. For some kids, this can happen quite early. I stopped understanding most of what my son talks about before he got out of elementary.
My son took the ACT very early while in elem and did well. I can understand scoring high in English, reading, and science because those are mostly reason and reading based tests and he's good at reading and reasoning. I was surprised at the percentile scored in math, which was broken down for topics and included high % for courses where he had no exposure that I knew about. When I asked him about the math, he said that none of it was hard and even not having geometry or trig, he could figure almost all the questions out if he was given enough time. He had to rush a bit and didn't figure them all out, but certainly managed most of them. My experience with PG kids is that they often know many things they "shouldn't" know based on exposure or formal teaching. It shouldn't be surprising that a kid who can self-teach reading can also self-teach math.
I confess that I always found the Explore scores taken as DYS levels to be low. I suspect that the group taking such tests are not in the top 5% as expected (most talent searches require parent agreement that their children had some subscore in the top 5%), but rather something like top 1/3 or 1/4 of ability tests. It's not unusual for a kid with multiple tests given in ability and achievement over time to have some subscore at 95%. Many schools offer GT screening for all kids, so you wouldn't even have to be seeking out testing options to have such results. Even if parents are always telling the truth, the sample is likely not really a GT sample. And if that's true, then how can we figure out what level would really be 99% or 99.9?
