The Psychology of Grading:
So here's a puzzle about the psychology of grading. Harvard and Stanford Law schools have recently announced moving from a letter grade system with pluses and minuses to a High/Pass/Low-Pass/Fail system. My sense is that most students like the change: Students perceive that it takes pressure off them.
But imagine a slight change. Imagine that instead of adopting the High/Pass/Low-Pass/Fail system, the schools kept the letter system and simply dropped pluses and minuses and the "D" grade. In other words, the possible grades became just A, B, C, and F.
My sense is that students would object strongly to such a system. They would object that it was too arbitrary and unfair, because a student who earned a very high B or a very high A would get no credit for it: They would just get the flat grade that didn't reflect their achievement. Indeed, I suspect some students would say that removing pluses and minuses would increase the pressure on students by giving students a single bar to hit rather than more of a sliding scale.
Why is this a puzzle? Well, the two systems are the same in a functional sense. High is just a new name for an A, Pass is the new name for a B, and Low Pass is the new name for a C. But my sense is that students don't see it that way. My best sense of why is that the experience of having received letter grades for almost 20 years of schooling before law school gives those letters tremendous meaning that new words like "high" and "pass" don't have. A switch to a new grading system makes the new grades feel different, even if the switch is mostly just a label.
But imagine a slight change. Imagine that instead of adopting the High/Pass/Low-Pass/Fail system, the schools kept the letter system and simply dropped pluses and minuses and the "D" grade. In other words, the possible grades became just A, B, C, and F.
My sense is that students would object strongly to such a system. They would object that it was too arbitrary and unfair, because a student who earned a very high B or a very high A would get no credit for it: They would just get the flat grade that didn't reflect their achievement. Indeed, I suspect some students would say that removing pluses and minuses would increase the pressure on students by giving students a single bar to hit rather than more of a sliding scale.
Why is this a puzzle? Well, the two systems are the same in a functional sense. High is just a new name for an A, Pass is the new name for a B, and Low Pass is the new name for a C. But my sense is that students don't see it that way. My best sense of why is that the experience of having received letter grades for almost 20 years of schooling before law school gives those letters tremendous meaning that new words like "high" and "pass" don't have. A switch to a new grading system makes the new grades feel different, even if the switch is mostly just a label.
[OK Comments: Perhaps. Although I should say that my sense was based on my three years as a student at Harvard Law from 1994-1997. We debated this exact question quite frequently, and the widely held opinion among the students I spoke with was that we would be very angry if the school abolished the +/- system but very happy if it adopted Yale's system.]
Where I teach we use a letter system with 12 possibilities. A student can get a straight A or an A- but not an A+. B's, C's, and D's all have pluses and minuses, and F has none. This is too many. As an instructor I can't justify the difference between a B and B+ to an inquiring student. Four or five possibilities at most would improve the situation, and two or three would be even better.
High Distinction
Distinction
Credit
Pass
Fail.
Just imagine the reverse: If Harvard Law only gave out a few A's (sort of like A+'s in some colleges), would many students be happy saying they had a B average and having to explain to people that 95% of the class got B's or C's?
So the new system is equally uninformative but without negative connotations.
Yes, only ~50% of the class would get an A or B. Hard grading, eh? For example, the Dean's list simply required a 3.0 average, with no D's or F's. Yet only about 35% of the students each semester made it. And only 2 of 750 members of my class had 4.0's, and that was abnormally high. Usually it was 0 or 1.
So far as the use of grades by graduate programs or employers in evaluating applicants, I suspect that a lot of them by now have realized that class rank is a more informative statistic.
First, I think that some students may be under the mistaken impression that Honors will represent A+'s and maybe A's. Everything between A- and B- would be a pass. One of the most maddening things as a (former) law student was the A-/B+ gap and how arbitrary that cut off felt. A system that eliminates all grades below A-’s would 1) eliminate the arbitrariness of the A-/B+ distinction and 2) make life much easier for the B/B+ student, which is what the majority of students are at HLS (due to the curve, of course). Conversely if honors went to a larger portion (say all A level grades) many people might feel like they are stuck in that same A-/B+ gap again.
Second, as an analytical matter the level of distinctions between grades and the discussion about the number demonstrates an overall problem with law school grading. The issue is iterations. Take for example a system that only has 2 real grades (not counting failing): Honors and Pass. It is comparable to baseball's batting average. In baseball, even though there are only two possible outcomes, a hit or an out, over enough iterations, one can determine the difference between players hitting abilities. (Incidentally the difference between 2/10 and 3/10 is incredibly wide). Even with only 24 courses an H or P system might separate people effectively. It would be even more effective depending on where the H cut off is. If 10% is the cut off, the difference between someone with 12 H's and someone with 2 H's is readily apparent.
Indeed given what the perceived student quality at schools like Harvard, Yale, and Stanford, with the exception of some very top firms and highly prestigious clerkships, it probably won’t hurt anyone in recruiting. Now if there were reported midterms or other evaluations to change the number of iterations the accuracy of a 2-grade system would increase. And this gets back to the problem with all law school grading. Even with all letter grades and pluses and minuses, it is arguable that the law school GPA is arbitrary given how few chances there are. Back in baseball terms there is a reason there is a minimum number of at bats to qualify for the batting championship. (Finally for those interested in overly long sports analogies, one could consider the slugging percentage, a weighted measure of hit power, as an analog to the standard grading systems.)
Finally, anecdotally, friends of mine that went to Yale said that the lack of grades changes the dynamic, but doesn’t change the stress. It simply forces you to become active in a journal and fight fiercely for professor face time. Resume lines and recommendations can be the only way to stand out.
Sorry this was so long.
also, according to most of my friends at schools with A+s, they have never heard of anyone getting them no matter how well they did and most of these students are envious of our system.
So it doesn't reflect the usual A/B/C breakdown, unless the amount of As and Cs given has changed drastically from my college years.
Unless you happen to be in probably the top 20% of the class, your grades aren't going to help you. If you go to Harvard, they probably aren't going to have any effect at all so long as you pass.
I'm sure someone has done studies that correlate law school grades to success as a lawyer down the line and I'm guessing it shows some sort of mild correlation at best, probably because there are some students that do well or do poorly because of something fundamental to the practice of law like writing skill or reading ability.
At least here, bar passage rates reflect this- the stop 90 percent of the class has an 80-90 percent bar passage rate. The rate does down to 50 percent for the bottom 10 percent of the class.
How do law professors separate out the 40th percentile from the 70th percentile? My hunch is that they really don't have any good way of doing it.
I strongly disagree, Jim at FSU. For example, most law students roughly get about 50% of the issues on an exam: I have never graded a law school exam, either at Chicago or GW, in which a student actually got every issue. (There's a very large gap between students and lawyers in this regard.) Also, I have never seen a law school exam in which regurgitation was helpful: If you regurgitate, you're not going to get a good grade.
The same dynamic is probably at play here. Students will know that showing a Pass on their transcript will not elicit thought that they could have worked a little bit harder to get a Pass and a half, whereas a people who see a B on the transcript will believe it was possible to get a B+
I'm assuming that under the current system the letter grades can be translated into numbers and used to calculate grade point averages. In most schools, an A is a 4 and a C is 2, so a C is worth only half as much as an A. I'm assuming that the shift to High pass etc. is signaling a move away from that system. Obviously, one could assign numerical values to those grades, but as I understand it, Yale doesn't do that.
From a student's perspective, I think that's a huge difference. A low pass might look bad on your transcript, but it doesn't bring down your average.
Now, I concede there are material differences in the two situations (such as midterms in the business school), but I think this would at least be a countervailing effect.
That's an excellent idea, honestly. I see far too many people for whom basic math is treated as some sort of cargo cult magic. I wish more elementary schools would devote greater attention to calculating in bases other than ten. I've seen schools that give more time to roman numerals (often doing so incorrectly to boot) than to bases.
I'm assuming that under the current system the letter grades can be translated into numbers and used to calculate grade point averages. In most schools, an A is a 4 and a C is 2, so a C is worth only half as much as an A. I'm assuming that the shift to High pass etc. is signaling a move away from that system. Obviously, one could assign numerical values to those grades, but as I understand it, Yale doesn't do that.
From a student's perspective, I think that's a huge difference. A low pass might look bad on your transcript, but it doesn't bring down your average.
Yeah but if you're a prospective employer you can ask the student for a transcript and evaluate it that way, if you so desire- or simply look for how many "high passes" you can find compared to others. "Low passes" will look bad on a transcript. The people who really benefit from this will be those in the middle and lower-middle of the class. The star performers and the washouts will still be identifiable, but the ones who are just getting by will be harder to distinguish from those who are nearly among the top grade-getters.
e.g., C- through A become clustered as 'Pass'. A+ becomes 'high' and D-/+ becomes low-pass.
This IS quite different.
But when the increments get bigger -- when there are just 3 options, Pass, High Pass, and Fail -- the chance of getting screwed by the variance goes up. (So does the chance of getting helped by the variance. But that didn't particularly concern me, since I thought that I did pretty well under a true assessment. Nor should it be a positive feature for law schools, since their goal should be true assessments -- not a few lucky winners.)
Distinction (Not generally awarded. Given to a student who demonstrates unusually good ability in the area)
Good (Top 15-20% of class)
Pass (Acceptable, but nothing remarkable. Most of the class will fit into this category)
Fail (Unacceptable)
One problem with the system was the lack of differentiation below Honors. Students who didn't expect to get the higher grades could coast and still pull a Pass. I remember one test where Distinguished was a raw score of 110 or more, Honors was 100 to 109 and Pass was from 99 down to 40 or so. Students with a 97 got the same grade as someone with a 45. The school went back to letter grades later.
If A/B/C/F do not map to H/P/L/F perfectly, and there is the perception that P's are better than B's, then the solution is perfectly obvious. Change the four-tiered letter grade system to A+/A/A-/F instead.
They, seem however, to believe they'll all be winners under the Yale-type High/Pass/Low-Pass/Fail system.
I think this is because they believe (without any empirical evidence?) that no one will Fail and almost no one will get a Low-Pass. They believe they can become experts at "gaming" the grade inflation potential of the new system.
FWIW, as a student I was always under a rigorous A-F system. I worked very hard for my "A" average and I was incensed whenever anyone talked about "reforming" the system to give others the benefits I worked so hard for because their self-esteem needed it or they'd be drafted otherwise (showing my age). Honors should recognize merit which usually comes from hard work.
Oh, BTW, the real world brutally selects the meritorious irrespective of the worker's need for self esteem. If you get beat all the time, you'll lose clients. If you lose clients, ... .
In my AP classes, most students end up with A's, those few who don't do well on my tests get A-'s, and those that don't do any of the homework (reading assignments) get E's. Removed of most of the pressure of grades, students instead focus on learning the material. My AP students have a very high rate of passing the AP exam.
Does this work for regular stduents? Less so. Not so much because the theory isn't right- it's just that regular students are so trained in the system that they can't break out of the box and actually learn. I think that's what elites want- a large population of trained gerbils who follow the letters to the cookie, but don't know sqaut about squat.
Anyways, I write about some of this stuff occasionally on my website- check it out and look at some of my education-related posts.
nah too much work
The change to H/P/L/F is a good thing, and more law schools should adopt this. Students will be less likely to be cutthroat when hard work and competence will guarantee a Pass and not an arbitrary grade-point letter between A- and B-. The quality of the school doesn't change, and recruiters will understand that a Pass at a top 20 means something different than a High Pass at a 3rd tier. I don't consider forcing law firm recruiters to look deeper into law students a bad thing, even if it does frighten those who go straight from undergrad to law school and don't have "real world" experience on their resume.
Law students are adults (or should be), it's okay to treat them like adults and not depend on grades to judge their value as people.
You don't sound like a conservative teacher (for that I'd go to Harvey C minus Mansfield). Your theory may be right for really really bright students (like the kind that get into Harvard and Yale Law). But for ordinary folks in ordinary classrooms (not where everyone in the room can score perfectly on the SAT) grades do mean something (if the teachers are effective graders). They are not meaningless. They reflect learning the material the teacher wants the students to learn and separate the more effective learners from the less effective learners.
So, we'll see if the HLS students end up finding the new system less stressful. Fewer distinctions will make the boundaries seem more arbitrary. Creating a culture of intellectual exchange driven by intrinsic as well as extrinsic factors is a noble pursuit, but such a culture can't take root if the students are focused on the competition for competition's sake, and that probably has more to do with the students and professors than the grading system. Bowdoin was pretty laid back (premeds less so), with some individuals putting pressure on themselves without getting smarmy about it. The switch to A/B/C/D/F had no meaningful impact on that culture, as feared, while probably giving profs the tool they needed to raise the minimum up.
That is really a function of a lack of + and -. Ultimately, I didn't like Bowdoin's system because I realized that I was a B+ student who ended up with a lot of Bs and a few As instead of a mix of mostly A- and B+ grades. (Grades weren't that inflated; my de facto 3.2 was in the top third or so of the class, based on cum laude designations).
The HS district where I teach shows + and - on transscripts but computes GPA with all As as 4.0 and all Bs as 3.0, which I find maddening. Kids with a B+ have huge incentives to cheat or be smarmy, and at exam time, folks with averages in the middle of a range have little incentive to study hard. Unlike college, we can't have finals count for huge %s of the semester grade in the first place for several reasonable reasons. But I hate seeing solid B students tail off at the end of semesters when a bit of healthy review would lock down their knowledge for the long term. I understand this well, having gone into finals at Bowdoin with a B+ and facing the diminishing marginal utility of studying your butt off to probably end up with a B in any case. Aren't law school finals so much of the grade that finessing grades like this is impossible anyway?
Orin makes a good point -- yet students will end up judging the system based on how it operates in practice and not in theory. Their reactions to the announcement of the change may be based on a gut reaction that this change takes pressure off, but their evaluation of the change itself will come from how the incentives are shaped and perceived within the school community. Given that, I doubt that this will have much material impact on the education as delivered and received, and will be as the schools intended, just an attempt to declare themselves peers with Yale.
I did not mind the system but many of my fellow students hated it because employers could not intelligently compare us to students at other schools — and we did not have the clout of Yale for that not to matter. So, two years after I graduated, Columbia reverted to a traditional A/B/C/D/F scale, exactly the opposite of what Harvard and Stanford apparently have now done.
Also, between screening clerkship and law firm applicants, I have seen dozens of Yale transcripts. I did not know until reading these blog posts that Yale had a Low Pass because I have never — NEVER — seen one on a transcript!
But I'm also often surprised by how differently faculty and grad students view grades. I'd guess that objectification of the form of a grade ("But I got a B, even if was a B-") outweighs placing the grade on the scale of possibles for a substantial proportion of students.
Of course, PhD programs have much higher failure-to-earn-degree rates than most professional programs too (before someone launches as screed about those softy humanists): I know that almost all medical students graduate, and suspect that a high proportion of law students do, whereas arts and sciences PhD completion rates run from 30% to maybe 70% at the very best programs (the better programs give much more support along the way, and also recruit the strongest students. Many students drop out for other than academic reasons, but there is often an academic subtext even when the official reason is "financial" or "personal").
I teach (part time) in a Master's Degree program that grades along these very lines. And I know exactly what this means. A "B" (which I rarely give) is the bottom of the class. Anything below that means you don't belong in the class. Below B- is failing.
Yet, in my full time academic job (a community college) a "B-" ain't so bad so to speak. A "C" transfers to 4 year colleges, so anything below that (A "D" is the next grade below) means something is wrong.
(My High School used numeric grading with 65 being the lowest passing grade. One quarter in my second year of Spanish, I got a 66. I don't remember passing a single test, but I showed up and tried so I think the teacher was being kind.)
In high school, we had an 8 point scale (A+, A, B+, etc...) This got averaged into an eight point scale. I was in a class of appox. 800 people, and after 3.5 years, there was a 6 way tie, to the 10,000th of a point. And there was a diference of less than .02 between the top forty students. These distinctions, of course, were statistically meaningless, and yet they still made a difference in class rank, and thus mattered for admissions. Pretending that grading can achieve this kind of precision is madness, and led to some truly absurd behavior in the race among the top six to become valedictorian in the last semester.
College (Yale) had A,B,C,D and F. Failing was extremely rare. I don't know of anyone who ever got a D. C was basically a failing grade, and happened once in a blue moon. So basically, it was about 35-40% As, and almost all the rest B's. There was no class rank, no GPA, and honors and distinction in the major came through the number of As one got. (I knew one person who ended up taking 11 classes in his last two semesters just to get the number of As to get Magna). So, even though this system had the appearance of letter grades, it was basically High Pass, Pass and Fail and that's how almost everyone treated it.
Columbia Film School was Pass/Fail. No-one in Hollywood cared what kind of grades you got in Grad school. The idea was absurd. The students were there to learn (or to kill time). Either way, grades didn't matter to the students, and we would have laughed at the idea. The reason for having a failing grade was so the faculty could kick out students who weren't cutting it. From an educational standpoint, this system was ideal. But it needs students who genuinely want to learn, otherwise it would be a catastrophe.
Yale Law School had the H/P/LP/F system. Some classes were just P/F, and were identified as such. That made some difference. I was in one class where the teacher (Deutsch) said that he wasn't going to give any honors from the outset. That was fine with me, but I then asked him to let us have the option of taking it as a P/F only class. I explained to him that I didn't want some employer thinking that I had failed to get an H. He saw the point and made the P/F grade possible. For me, the Law School grading was essentially the same as my undergraduate grading. What mattered was not any "average'. Rather, it was the number of H's on the transcript. And the distribution between H and P was about the same as between As and Bs in college.
So, my experience doesn't jibe with Orin's ideas. For me, the A B C F and the Yale system really were functionally, and psychologically the same. Of course, I was in the enviable position of never going to any school after high school where grades were a major concern.
Even if you get credit for a D class, it's a horrible grade so I don't know why you'd want it on your transcript to begin with... To me it makes sense not to give D's credit. If you can't get an average (C), I don't think you should get credit for a class. But I find not having pluses and minuses is more egalitarian in terms of GPA... I don't have to worry about someone looking at an A- and going, oh, that's not really an A, it's just as good as a B+. Getting rid of pluses and minuses puts a clear distinction between the grades.
I'm curious, what if you get a B+? Are you relieved that you don't have to worry about someone looking at a B+ and going oh, that's not really a B, it's just as good as an A-?
Reducing the gradations provides more incentives for a student to distinguish herself via non-grade signaling, such as journal participation, internships, or trying to publish articles. I have learned more through internships and other external activities than from any one class, and most other students I know feel the same. Encouraging students to spend more time on practical training, rather than excess studying to go from an A- to an A, probably would be better for everyone in the long run. Plus, it is a better way to signal preparedness for working in the "real world."
That suggests a huge gap in either the "teaching" or "testing" end of the teach - learn - test - grade process. If the suggestion is that most lawyers would get close to 100%, when law students get 50%, then the problem would have to be with teaching.