The Gap is Everywhere!

Of late, I have been writing about problems that differential grading causes, both in terms of employers and graduate schools picking the “wrong” students, and in terms of dissuading students from studying certain subjects.

I must admit that my arguments and analysis of different grading schemes in different disciplines was based on one crucial assumption:

The intellectual quality of science students is no worse than that of

students in the humanities, arts, or social sciences.

I readily admit that, if science students tend to be less intelligent than humanities students, then they deserve poorer outcomes. However, there seems to be little evidence to this effect so I shall not belabor this point.

More importantly, I have recently run across several studies, which support the idea of grade normalization across subjects

In a landmark study done by Durham University in the UK, which tracked over a million students taking A-levels and GSCE exams, researchers found that students of equal ability received approximately a full letter grade lower in the sciences than in subjects like drama and English. The study was commissioned by the Royal Society (the world’s oldest professional body for the sciences, whose membership includes the likes of Newton and Hawking) in order to provide evidence that differential grading schemes may be contributing to a decrease in the number of students pursuing science degrees in the UK.

While my analysis in my last post was solely regarding grades at Dartmouth, it appears that the same differential grading practices take place all across the US as well. An online magazine article describes studies done at Cornell, comments by a Stanford Business School professor, and several other pieces of evidence which all point to the idea that lower grades in sciences can dissuade students from pursuing science degrees.

It is also evident that these grading differences between the sciences and humanities have been there for many decades. The question thus arises, why did they arise and why has no one corrected them, despite the easily available evidence as to their existence?

How Wide is the Grading Gap?

Introduction

While most people concede there is a grading gap between the sciences and the humanities, few are aware of big it is. In order to determine precisely how wide the gap in grading is between various areas of study, I analyzed the grades at my alma mater. Dartmouth is one of the few colleges that publically makes grading information available on the web. And there are fewer still colleges that publish course level data, such as Dartmouth and Cornell do.

In particular, in an attempt to combat grade inflation, Dartmouth started placing median grades next to the student’s grade on his or her transcript for each course with an enrollment of 10 students or above. Additionally, Dartmouth College publishes a list of all courses that fit this enrollment criterion along with their median grades on its registrar’s website.

Methodology

In order to analyze the grading gap between different broad disciplines, I obtained the median grades of all available courses on the website, totaling to 3166 courses taken across the last two years. I then placed each course into one of four categories:
1. Arts
2. Humanities
3. Soft or Social Sciences
4. Hard Sciences.

After classification of each course, I calculated the mean of all the median grades in each discipline. You can view the results here.

In looking at the data, there are some useful things to know. At Dartmouth, an “A” is a 4.0, A- is 3.66, B+ is 3.33 and so on. Furthermore, the median student at the time of graduation has approximately a 3.3 GPA. The cut-off cumulative GPAs to be in the top 5%, 10% and 15% of students at the time of graduation are 3.89, 3.77, and 3.59, respectively.

Results

In light of these reference points, the results are quite shocking. A science student who achieved the median grade in each of his or her courses would have a GPA in the middle of his or her graduating class, whereas a humanities student achieving the same feat in his or her courses will receive honors for being in the top 10% or 15% of all graduating students.

There is a 0.43 GPA gap between arts and hard sciences students and a 0.23 GPA gap between humanities and hard sciences students, which is needless to say, significant. A difference in 0.23 grade points, ceteris paribus, makes the difference between getting into a top professional school or not, between getting a prized job or not, between being awarded honors at graduation, or not.

Differential Grading – Altering Intellectual Paths

During my analysis of the disincentives to study science in the US, I described how lower average grades may dissuade students from studying the sciences for a variety of reasons. Differential grading across majors and subject areas poses a much broader threat to our education system and its aims than simply lowering the interest in studying sciences.

One of the fundamental goals of education, at any level and at any age, is to enable an individual to pursue his or her intellectual interests. Admittedly, this is rarely possible in the purest sense of the word freedom, because of economic constraints, parental pressures, and a variety of other factors. However, it is still desirable to grant students as much freedom as possible to pursue what they are passionate about.

Unfortunately, different grading schemes across subjects and across courses within the same subject tend to further incentivize students not to take certain courses or disciplines. College students are desirous of maintaining a high GPA, to impress their parents, their friends and for admissions into graduate schools and for employment purposes. Consequently, in deciding their majors, college students often consider what the grades are like in that major. Often, the average grades are a significant deciding factor when students pick college majors (I know this from my own experience in advising underclassmen in picking a major).

Even within a particular discipline, students often pick easier classes. “Easier” in this context means that the course has a good grade-difficult payoff. In other words, students are often unwilling to put in the extra effort for a challenging upper level course if it does not have a relatively higher average grade as compared to lower level courses.

(For the purpose of simplifying my examples, I have restricted them to those that would occur at the college level. However, many of the same arguments apply to high school students. For example, a high school student would want to maintain a high GPA for college admissions as opposed to for employment.)

Our education system aims to both prepare students for the world and allow for the student’s personal intellectual growth. A prerequisite for the latter is the ability of the student to pursue his or her academic passions. However, if the student is penalized heavily for doing so, in the form of a poorer grade, her or she may decide simply to ignore those passions and study something “easier” (here easy again means a higher grade-difficult payoff).

Inflation Part 2 - Inflation at the Top

In my last post, I only described grade inflation at Dartmouth, but the problem is just as bad amongst the other Ivies. Nearly half of all Harvard students receive A’s in their courses, and approximately 80% have honors conferred upon them at graduation (see: Ivy League Grade Inflation). At UPenn 51% of grades are in the A range; at Columbia 22.3% are A’s (not including A-'s); at Cornell A range grades make up 41% of awarded grades. At Princeton, which has instituted grade deflation programs, by limiting the number of A range grades that could be awarded, 42.9% of grades are still in the A range. More statistics can be found at Gradeinflation.com.

These statistics reveal an unfortunate trend at some of America's top institutions. The very institutions who have world-class brands and don't need to inflate grades still do. A range grades constitute nearly half of all grades awarded, even at Princeton which has attempted to institute a deflationary policy (unfortunately, with little success).

Many Princeton students who are against grade deflation cite that their quota system results in extreme competition between students, as there are a limited number of As to go around and that this competition may be to the detriment of learning. I feel there are three strong responses to this point. Firstly, those who truly care about learning will still learn. Secondly, a B does not destroy their career opportunities because it is widely known that Princeton is attempting to institute this policy. Thirdly, while learning is certainly important for education, I do believe that competition is a necessary part of college and in fact life. Indeed, it is very rare when one is judged on absolute scales; more common is the judging of one relative to ones peers. Quota systems such the one at Princeton, if done effectively, normalize grading. An “A” means you are much better than your peers and a B or B+ says you are average.

While I commend Princeton on taking the first step amongst the Ivies, I do wish it had taken a much stronger stance. For example, officially only 35% of students are supposed to get in the A range, but, in fact, about 43% do, thereby showing that the policy is not strictly enforced. And 35% of students being classified as "excellent" or "outstanding" still seems a little high.

In finance, printing money causes inflation. In academics, printing A’s causes inflation. Let our academic institutions not go the way of Zimbabwe.

Inflation Part 1 - A Rant

Inflation's become a buzz word recently, especially in the context of economics. With interest rates hovering around zero, inflation is rearing its ugly head. Central banks, the IMF, and investors around the world are scared at what many see to be an inevitable period of hyperinflation, when prices will skyrocket for both consumer and industrial goods.

There are many reasons that inflation troubles anyone who has to deal with money (which includes you if you’ve ever bought, sold, lent or borrowed anything in your life). Inflation, in its simplest form, is defined as an increase in prices (or put another way, a decrease in the value of currency). Two of the key problems with inflation are that it causes devaluation of money and that it creates financial instability in markets, because it becomes difficult to value things, as one has to account for an often unobserved level of inflation.

Inflation is hardly a problem isolated to finance. It is just as pressing, and nearly as concerning, a problem for our education systems, where grades are being inflated constantly. Just as with money, when grades are inflated it devalues them.

How bad is it?
To see how bad the problem really is, I performed an analysis of how bad the problem really is at Dartmouth, my alma mater, which places the median grades for all courses online (see: Median Grades at Dartmouth). Out of 450 classes analyzed the unweighted average was 3.33 and the average weighted by class size was 3.18. This problem is seen in further light when you note that out of 450 classes, 268 of them (i.e. over half) had A- medians or higher, with 48 (10.6% of the total) having A medians.

Consequences
Because grade inflation is so common, employers, who often lack the time to do an in-depth analysis of grades, may simply ignore the grade that a student has worked hard to earn, even if the student’s school does not grade inflate. This is especially hurtful to students graduating from non top-tier institutions, as employers assume that the institution is inflating grades, so an A from that institution will mean nothing. Hence, a student graduating from a good state school with excellent grades may be passed over in favor of a Stanford graduate, as the employer is unable to compare grades and must resort to using educational pedigrees in discriminating who to hire.

While graduate schools are often more savvy in terms of separating good students from bad, they too find it difficult to separate between a “good A” and an “excellent A.”

A Final Word
It is interesting that we are so concerned about inflation in economics as it hurts our investments, and yet, we are so unconcerned and unwilling to act on protecting our grades from devaluation. Those same grades, which we have spent hundreds of hours of studying for, tens of thousands of dollars investing in, and nearly two decades of in school, to obtain. A significant investment, indeed, but not one protected from the devaluating effects of education.