To be fair, in fact, the data are not all that reliable for young kids. Now, an EG/PG youngster is not going to score well on early testing and then suddenly plummet to mere high average two years later (absent TBI, major illness, or some other traumatic change affecting mental functioning), but there is enough instability in IQ scores, and competition for gifted slots, that it makes some sense to wait to identify for highly-selective programs until the scores settle down around age 9.
That's where these policies (prevalent in many school systems across NA) originate from. Unfortunately, they ought to be accompanied by generous support for academically-advanced students K-3, so that GT students don't have to just sit around twiddling their thumbs while they wait to be identified.
Just as, on the other end of the spectrum, we should be generously (in both cases, I am speaking both of intervention admission criteria and of the interventions themselves) intervening with the lower-functioning end of the academic spectrum during the primary years, instead of waiting for them to fail, and then identifying them for special education services.
We can't always pin down who the extremes are in the early years, but since plenty of children off the mean, but not in the tails, would benefit from additional support or challenge, it would make sense to service a bigger slice of each end in the mean time.
