Our district administers CogAT at one level higher than "typical"grade level as part of the assessment for the GT program. I understand that this is the publisher's recommendation due to ceilings at the higher end of the bell curve; however, I am having trouble finding information to clarify this. Does this mean that the Standard Age Scores on that end may be inflated? That percentiles are not accurate?
Also, the district uses grade rank percentiles. As I understand it, the raw score is used to find the Universal Scale Score. The USS is then
used to find the grade percentile. Age/grade is not taken into account as it is in SAS and age percentile rank. Wouldn't this mean that the GPR's show how students scored in relation to students in the NEXT grade (e.g. 2nd graders take LEVEL A and GPRs reflect performance relative to 3rd graders)? Any help os greatly appreciated!

Maybe I should narrow this down a bit. How many districts use above level CogAT for identification? If so, do they use Age Rank Percentiles or Grade Rank Percentiles? Thanks!

Ours uses above level CogAT. So DD took the CogAT 6A (third grade) a couple months after her 7th birthday. Her verbal score was 130, or 97th percentile for her age. But 95th percentile for the grade score. Since she was accelerated, and technically should have still been in first grade, she would have been one of the youngest kids taking that level of test. I'm assuming they are comparing her to other second graders who took the test, not third graders. On the report it says she is in Grade 2 so obviously someone entered that in the system. It also says Norms: Midyear 2005. Not sure what that means. She took the test Jan. 2013.
Not sure if this helps to answer your question. In terms of which score they use, they only care about the composite score, or standard score (whatever it's called), so they have cut-offs like 132 or 139. I think those are based on the child's age, not grade.

Our district uses the CogAT two grades up, so the 3rd graders take the CogAT intended for 5th graders. The reason for this is that when a kid takes an on level test, there is essentially no room for error at the 98th percentile (which is the cutoff for the gifted program). The kid has to get all of the questions right. When you go two grades up, it makes it so that a few dumb mistakes don't disqualify someone from the gifted program who should actually be there. (Of course, the CogAT is notorious for not identifying HG+ kids as gifted, but that's another issue.)

Anyway, in our district, they have the scores reported as compared to the kid's actual age/grade. There is a place on the answer sheet where that is entered.