<p>First, I want to be clear about terminology. I believe that there are very few, if any, bad students at any of the colleges under discussion. A high school performance that would place one in the bottom quarter at Berkeley might still earn one a merit scholarship at many fine colleges. So I am NOT talking about "bad" students. Getting into Berkeley is quite an accomplishment, and should not be dismissed. However, it is even harder to get into Stanford, and other elite privates, and the grade distributions reflect the strength of the student bodies.</p>
<p>So the question is whether, given the same performance, one would get a different grade at Stanford than at Berkeley. Since no one actually takes the same course at both places, one cannot answer this question directly. It must be inferred from related data. We know that the distribution of grades at both colleges is remarkably similar, but that the distribution of high school qualifications is not similar, but rather much higher at Stanford. I take this to indicate easier grading at UCB. Sakky says that his interest is what happens at the bottom of the distribution. For those who end up at the bottom, are they more likely to flunk out at UCB or at Stanford? </p>
<p>I don't have grade distribution data for Stanford, so I have been using Princeton as a proxy. At P, D's and F's account for ~1.5% of all grades. At UCB "less than 5% were D's or F's". So one can be somewhat higher in the overall ranking and still get a D or F at UCB than at S (again assuming S is similar to P). Now, does this small difference in percent D and F at an elite private vs an elite public make up for the large difference in the composition of the bottom of the class? To me, the answer would be "no, there are so many students at UCB who would not be admitted to S (or P) that one would have expected a much larger portion of D's and F's at UCB if they gave that grade to students who performed at the bottom of the S or P distribution". </p>
<p>I illustrated one very crude pass at the question: take away enough UCB students to eliminate that large cohort who would not be admitted to S or P, then look at the grades of those who are left. But to do this one would have to eliminate more than the bottom quarter. Since few colleges publish the distribution of high school qualifications in enough detail to attempt to fully equalize the student bodies, I started with the bottom quarter at UCB. Even this incomplete adjustment is revealing. Just eliminating the bottom quarter of grades gets rid every grade below B.</p>
<p>One could argue that truly terrible performance at UCB earns an F, but that comparable performance would be rewarded with a C at Stanford. Given the rarity of D and F grades at both places, and the much lower student distribution at UCB, I do not find that persuasive. Instead, it seems likely that there are lots of students at UCB who would be hard pressed to keep up at S or P who are not only passing, but are getting B's in many of their courses.</p>
<p>None of this is to deny that classes are smaller at S and P, certainly they are. Not to deny that the elite privates pay more attention to their undergrads, that there is more contact with professors, more academic support, and an overall better academic experience- at the prices they charge they had better add value to the UCB offering.</p>
<p>But the data do not support the idea that the same student, turning in the same performance, would get a higher grade at S or P, anywhere along the distribution from the top to the bottom. In fact, given very small differences in grade distribution, extremely low rates of D and F grades, but large differences in high school preparation, the grading seems a tad easier at B than at these elite privates.</p>