<p>kluge alludes to the excellent point that there are multiple kinds of grade inflation.</p>
<p>What do we want when a school issues grades? We’re trying to get an assessment of where candidates stand – who had the better college career? To do that, we need two primary things.</p>
<p>(1) We need consistency. We need to know that a 3.8 from one place means roughly the same thing as a 3.8 another place.</p>
<p>(2) We need meaningful separation. If the best students in the class get a 3.65 while the worst ones get a 3.6, then a lot of what determines a “good student” is going to come down to a single B or something like that – something that could easily be determined by luck. In general, then, we’ll prefer grading systems that use the entire spectrum in a non-arbitrary way. Kids who get patterns of C’s really are probably worse than kids who get patterns of A’s. But if everybody’s getting a mix of (say) A’s and A-'s, then it’s really hard to tell what’s excellence and what’s just luck.</p>
<hr>
<p>Objection #1: Grade inflation over time. For example, Princeton English majors now get much higher grades than Princeton English majors used to.</p>
<p>This is the phenomenon with which Gradeinflation.com is chiefly concerned. It may be that students now are just smarter than they used to be, but (a) I doubt it, and (b) even if that’s true, it doesn’t prove that this is the best way to issue grades now.</p>
<p>(1) Is it consistent? This isn’t a problem, since nobody really needs to compare a Princeton grad from 1970 with a Princeton grad from 2000.</p>
<p>(2) Does it give us good separation? Ah, this is the problem. As grades rise and rise, we’re using less and less of the spectrum – and more and more grades are being determined by arbitrary factors.</p>
<hr>
<p>Objection #2: Some institutions are more inflated than other institutions.</p>
<p>Again, I think this is true – even if the conventional wisdom is precisely backwards. Actually, most folks complain about inflation at Ivies, but really it turns out that Ivies and their ilk are deflated, not inflated.</p>
<p>In any case, though, it’s still a problem.</p>
<p>(1) Consistency is a problem because we would like GPAs across institutions to mean roughly the same thing. Ideally, a 3.5 from UMich would be roughly the same as a 3.5 from Princeton. Right now, it’s not.</p>
<p>(2) If you want to use the entire spectrum, then the solution is to push down Princeton grades, while pushing Michigan grades farther down.</p>
<hr>
<p>Objection #3: Some majors are inflated relative to each other. Again I think this is a serious concern, because students are naturally drawn to fields where they will do well. This results in a dearth of the hard science, engineering, and technology majors that our nation needs so badly.</p>
<hr>
<p>EDIT: Hopefully that’s clearer.</p>