f the (law school) curve

"If you don't value your time, neither will others. Stop giving away your time and talents–start charging for it." –Kim Garst

I once had a class curved the way you suggest. The problem with it was we had one very smart guy in the class. He always got the top score and people would always get mad at him when he did well for "raising the bar". So you create an environment where no one wants to be the highest scorer in the class because that means youre lowering everyone else's grade.

In Australia, there's no cold calling and no curve so everyone ends up getting their degrees pretty stress-free. But in terms of Australian "Big Law" jobs – these 6 firms have their own cut-off systems where your application (both for clerkships and graduate positions) is not even considered unless you've achieved a min GPA of 6 out of perfect 7, i.e. you're an A+/A student across almost all subjects. And it is incredibly hard, if not impossible, to get a BigLaw graduate role, unless you did a clerkship (summer/ winter internship) for that particular firm.

For #1: This strikes me as a correlation/causation fallacy. They take for granted that the inability to get all possible points on a test causes the curve, when it could be: (1) the curve that causes the inability to get all points on the test; (2) both are coincidental; and (3) an unknown 3rd factor that causes both the inability to get all points on the test and the curve.

For #2: This seems to be an attempt to point to #1's option (3) to explain the relationship. This will be true of any system because it is inherent in our human nature. Though, some other grading systems can mitigate this. I believe Harvard doesn't use the curve and have a pass/fail system. The difference in grading still exists, but I suspect it's much more negligible because most students score enough above passing that the difference in grading doesn't matter as much. So, while the difference in grading by the professors is still an issue, it's risk is mitigated much more. Given that most academia and professional training doesn't seek to eliminate this, it begs the question why law school is so unique in its aversion. Tests are anonymized to eliminate bias against the individual, but the unequal grading methods still manifest in other ways. Some professors may only give grades close to the curve, whereby less students do poorly and less do above average ("imperfect average approach"). Other professors are more of a spread in their distribution ("glass cannon approach"). In this way, unequal grading methods are still present in the system. In short, it doesn't seem that a desire to eliminate bias is the reason, and if it is, it doesn't accomplish what it seeks to do.

For #3: I think you nailed this.

My thoughts: Discussion around the curve tends to revolve around the utility of a SINGLE curve ("isolated curve"). However, throughout law school by virtue of taking many classes, we are placed on MULTIPLE curves ("convergence curve"). Building off the idea that a single curve is about averages, a student's law school career should also be one of averages. In theory, the amalgamation of all 3 years of law school grades, compared amongst all students, should be a sort of "grand curve."

However, I don't think this theory would be accurate, as the ability to choose ones classes creates a system bias where some classes are graded/curved differently, catered more to the student's natural interests, and may be manipulated by the datapoint (i.e. student) in the way they strategize classes. Consequently, the median GPA at the end of the "convergence curve" will be higher than the "isolated curve." If the curve in general exists for a purpose of uniformity of averages, one would not expect disparate results between the isolated curve and the convergence curve.

This issue can be eliminated by considering the convergence curve being limited as a data set to 1L year. There, everyone takes the same classes at random. So, the idea I present is that if a student performs randomly on isolated curves (sort of like the "glass canon" description in #2), they should be around median at the convergence curve (see "imperfect average approach" in #2). However, there is one thing that undermines the theory of averages. One thing may undermine this theory: High performing students and low performing students seem consistently placed on their extremes on both the isolated curve and the convergence curve. I don't think this quite undermines it. The extremes are more resistant to methods of averaging than students who are more scattered in their performance. As a consequence, both the isolated curve and the convergence curve seems to have the primary effect of separating the wheat from the chaff, and the secondary effect of just lumping everything else in the middle. It's an imperfect averaging mechanism.

I'm not fond of the system because it artificially emphasizes uniformity over merit; it's not a meritocracy. On the other hand, it greatly reduces the risk of flunking out and being on the hook for up to $200,000 in tuition. So, a question manifests as to whether the curve system does a disservice to future clients by pushing out less-than-capable lawyers? I think we can answer that with an empirical observation: no. Pulling back and expanding scope, I'm not even sure law school is necessary in its current form and its certainly not the best. I frequently hear about "3LOL". I recall even President Obama said law school should be 2 years, not 3 (https://archive.nytimes.com/dealbook.nytimes.com/2013/08/23/obama-says-law-school-should-be-two-years-not-three/)