Keeping in mind that this is just my opinion...
According to ACT, the growth rate from 8th to 9th grade in students on-target for college readiness was about 2 score points. This suggests that the lower limit for growth in a GT child taking Explore as an out-of-level would be about 2. If we take my guesstimate of ceiling as accurate, then those scoring 20 or above would be at the point where subsequent assessment with Explore is likely not to capture their full range of achievement.
PSAT 8/9 is, of course, a different instrument, though designed for the same grade range, so one may or may not be able to extrapolate a similar truncation effect. I don't believe there is sufficient longitudinal data on the PSAT 8/9 to speak confidently about growth rates on the combined PSAT89/PSAT10/SAT score scale.
My inclination would be to move any Explore taker who appears likely to go off the end of the Explore/PSAT89 on their next administration to a higher level of the SAT or ACT suites. So probably those scoring in the 20s and maybe upper teens, and those who have already had math beyond algebra I.
