I have reason to believe the extended norms are less accurate for 6 and 7 year-olds.
In the case of my DS7, his extended norms were capped by the limits of the subtests. For example, he completed the coding test with a perfect score in half the time (62 out of 120 seconds).
Although he clearly had a lot of spare capacity, he hit the limit of the subtest and his score was artificially capped, even under the extended norms.
The version of the coding test used for 8+ year-olds has a much higher ceiling. If he would have taken that version of the test, it's likely that his processing speed (and thus his FSIQ) would be significantly higher.
Because the FSIQ scores are normalized, the fact that my son's FSIQ score was artificially capped suggests that other 6-7 year-old FSIQ scores are artificially inflated.
