Maybe they're just all /really/ smart? edit: actually as an addendum to this (don't want to hide dumb statements). I realize my problem is not the high proport

Maybe they're just all /really/ smart? edit: actually as an addendum to this (don't want to hide dumb statements). I realize my problem is not the high proportion of high grades, or the high average GPA, however you look at it. If you have a group of student who for various reasons are very capable - there's lots of nonsense with Ivy's, but they do have entry grade filtering - you should expect more of them to get higher grades. The problem is, I don't think there's alternate evidence that each year is inherently more capable than the prior year. Assuming same entry restrictions, and so presumable capability, you should expect on the whole the average GPA to hover at a consistent level, sometimes going a little up, sometimes a little down, but generally hovering around the same place. In that situation it would still be completely possible for one school to have a significantly higher average GPA than a /different/ school. The idea of grade is to measure capability, and for the sake of argument lets pretend every university/college used the same end of year tests for math, physics, law, whathaveyou. In such a case if you compared the average GPA from two schools where one was a community college in a remote rural america, and the other was something like MIT, an Ivy, ICs, etc you would absolutely expect a significant gap in those averages. So the problem isn't what the average is, its that the average is moving in a consistent direction year after year, and that is an indication of grade inflation. Growing up in NZ the national standardized tests, and university courses, had this concept called "scaling". The goal of scaling was to recognize than from year to year there may simply be unintentional variations in assessment difficulty. So what happened is at the end of a year/course, the average % score of the students is calculated, and compared to the prior, if the difference is significant every % grade is essentially scaled by some factor of the difference. That has the effect of pulling this years average grade closer to prior years, and ideally compensating for assessment variation.