<p>
</p>
<p>There are huge differences in grading standards within and across departments. At my institution, we did a simple exercise to show this. For each department, we calculated the difference between what their majors got inside the major and outside the major. So, if Psychology majors average 3.45 in their Psychology classes, and 3.38 in their other classes, they would get 3.38 - 3.45 = -0.07. (This is just the simplest way to present differences across departments. One could go further and estimate course and student fixed effects.) There were figures ranging from +0.20 to -0.80. So, there are monstrous differences across departments in grading standards. But even within departments there are differences. </p>
<p>Another point is that we throw away so much information by using the current letter grade system that lumps together so many people. The brilliant student who is head and shoulders above the rest of the class gets the same A as the plodder who is 5th in the class of 30. (Letters of recommendation may distinguish them, but they might not.) Returning to a numerical scale (e.g., 0-100), as Yale has proposed doing, would help. </p>
<p>
</p>
<p>I have no problem with using GRE or GMAT scores instead of the SAT, but the question is not whether college GPA is a more valid measure than the SAT; rather, the question is whether the SAT provides additional context for interpreting the GPA. Clearly, many employers feel that both have value. </p>