Tall, Dark, and Mysterious

1/17/2005

Grade inflation

File under: Those Who Can't, Home And Native Land, I Read The News Today, Oh Boy. Posted by Moebius Stripper at 2:33 pm.

Education Wonk wants to hear from Canadian educators about grade inflation. I’ve got some stuff to say about that, mostly anecdotal, but the statistics support me:

A recent National Post survey of 25 universities including SFU found that only 14 of 81 faculties handed out lower grades than they did five or 10 years ago. It also found that the average grade in most faculties is a B.

SFU, the article mentions, seems to be immune to this trend, but the article doesn’t give much insight as to why that is. Pity.

Some subjects suffer more than others. I think I read the print version of this article when it first came out:

Using information on first-year university grades from across Ontario, we examine whether or not there has been grade inflation by discipline. In a survey of seven universities for the periods 1973-74 and 1993-94, we find significant grade inflation in various Arts and Science programs. The rate of inflation is not uniform. Some subjects, such as Mathematics experienced little or no change in average grades at most universities, while English and Biology experienced significant grade inflation.

A former camper of mine from Toronto told me about one local high school - I don’t remember the name - in which students routinely graduated with averages of 100%. One by one, they would flunk out of college, and eventually it got to the point that some universities would no longer accept graduates of this school. Those high marks meant nothing.

There’s a lot of talk about grade inflation, but measuring it is not easy. Addressing it is even more difficult.

For instance, EdWonk’s query was inspired by an article about a British math(s) exam, linked here, in which a grade of 17% translated into a B. Having not seen the test, I can’t say if this is an example of grade inflation: it’s possible that the exam was preposterously difficult, and that B students in the 1950’s would have received grades of 17% on it as well. What is clear is that if that’s the case, then the exam was unreasonable. Kimberly Swygert puts it well:

…it’s possible that a 17% on the higher-level paper really is B work…However, it doesn’t seem very useful to have an exam in which even the A scorers get half the questions wrong, because all those additional items go to waste. It would make far more sense to assemble the exam to have many more of the B level items, which would both raise the percent-passing level for a B (thus satisfying the public) and better discriminate among B and A level students. There can still be a few impossible items on there to sort out the A from the A* kids, but there don’t need to be many of such items, if they’re well-chosen.

This reminds me of a second-year physics exam I took as an undergrad: I scored an A+ on the midterm; the class average was a D. Most of us anticipated that the final exam would be a cakewalk, so as to bring up the average grade. We couldn’t have been more wrong. There were seven questions on the final; I (of the A+ midterm) understood what three of them were asking. I spent the entire three hours in the exam room anyway, figuring that I might as well make the prof suffer when he graded my paper. It paid off, and my final grade in the course was an A. I should qualify that: it paid off, provided the only thing I cared about was my letter grade in the course. I’d like to say that I also cared about what I learned in the course, but I’m not so sure I did at the time.

And that’s a big part of the problem when it comes to analysing these things: judgements about the suitability of a grade or a test are typically made based on little other than the actual numerical or letter grade received on various tests. If a 17% translates into a B, then grades are being inflated. If a class average was 57%, then the test was too hard. If students are averaging 75%’s in a math class, then the tests are too easy.

At the two universities where I’ve taught, the only supervision neophyte teachers received from the senior faculty came in the form of making sure that the marks - the average marks we gave in our classes were okay. Every few weeks, I’d be asked how it was going, and how my students were doing.

“They wrote their second test the other day,” I’d say.

“How did they do?”

“Class average was 62%.”

“Not bad - 65% would be better, but 62% is fine. Next time aim for a bit higher.”

That’s me being addressed in that last quote: the implication is that the marks (the average marks, anyway) that my students receive are entirely under my control. No wonder grades are being inflated.

(ETA - Outside the Beltway has more along those lines.)

In truth, a mark that is higher than it “should” be can mean any number of things. It could mean that my test was too easy, as has certainly been the case at times. It could mean that my students are smarter than last year’s. It could mean that my students worked extremely hard - this is what happened in my discrete math class last term: I gave difficult tests, and the kids blew me away. It could mean that I am the most effective teacher the school has ever hired. It could mean that one student hacked into my computer and distributed copies of the test to their classmates ahead of time. Simply designing a more difficult test would be the right course of action in the first of these situations, but not in the others.

It’s hard for a new teacher to know why a class has performed better or worse than they “should”. I consider myself a pretty decent teacher, but I am an inexperienced one. There is no class that I’ve taught more than three times. I would wager that a teacher would have to teach a course ten time or more before they could get a reliable idea of how one group of students should be doing relative to another, and whether, for instance, it was justified that the class of ‘02 had an average of 68% while the class of ‘97 had an average of 63%. At the university level, I suspect that relatively few instructors have taught a single class that many times. University instructors are an intellectually restless bunch, who grow bored teaching the same material over and over again. In other words: it’s very difficult for individual teachers, even highly talented ones, to ensure that their students are receiving the grades they deserve, the grades that their parents would have received with comparable understanding of the material.

For this reason, among others, I support standardized testing, particularly in math - cross-referencing standardized test marks against class marks over the years can paint a decent picture of whether students are receiving higher or lower marks than they would have received several years ago. However, this too is fallible: for obvious reasons, it’s unwise to simply give the same standardized test to students year after year, as both students and teachers will respond by learning how to answer the specific questions on the test and not the actual subject matter. Altering the standardized test significantly from year to year gets around this problem, but only by relaxing control over the samples.

(Do not misread these previous paragraphs to say that I think that uncertainty exists and that therefore, one can never really know anything about things in general and grade inflation in particular; I do teach statistics, after all. All I’m saying is that accurately measuring - never mind correcting - grade inflation is a tricky affair, one that requires attentiveness to any number of variables. I gather Kimberly and her ilk may have more to say about the specifics of this problem.)

However, the difficulty of applying consistent standards from year to year does not explain why, by most accounts, grades measured over long periods of time are increasing rather than decreasing or regressing to the mean. I’d be interested in what others have to say about this; I can speculate, though.

At the risk of waxing nostalgic over an era during which I was but a twinkle in my parents’ eye: the last several decades have given rise to two changes in attitude, which are almost certainly related: the notion that a university education is necessary in order to ensure success in life; and the idea that a university education is a commodity to be purchased in exchange for eventual employment. (I’ve discussed the tension between the goals of the academy with the perceived (and actual) importance of university education at Jobs are for the little people.)

Students are told - and told correctly, to a large extent - that they’ll be terminally unemployed and unemployable if they don’t go to university. University, for them, is not an option; it’s a requirement. Since decent marks are needed in order to get into university, they too are a requirement. Grade inflation is merely the right-hand side of that equation.

Every university instructor will tell you stories about the clueless student who indignantly proclaims that the B they received on that essay or the C they received on that test wasn’t good enough, and would the teacher please reconsider. I’ve read several such stories on blogs, and I’ve told a number of them here. The fact that my students can’t add fractions is no reason, apparently, for me not to give them at least B’s. After all, they’re trying. And they need these marks.

When these stories appear on my elders’ blogs, they’re almost invariably accompanied by commentary of the form well I never - “Back when I was a student,” the bloggers write, “I’d never have even considered complaining about my marks…”

In the most extreme cases, students have sued, or attempted to sue, professors. A fellow grad student of mine once taught a calculus class in which one student hadn’t achieved the grade he needed to keep his scholarship. The student complained to my classmate. My classmate held his ground. The student complained to the course coordinator. The course coordinator replied that he wasn’t going to change the grade unless Grad Student did. At this point, the student’s father called Grad Student, who, while agitated, didn’t buckle. Finally, Dad called Course Coordinator, and while I don’t know exactly what transpired during that phone call, Dad must have been pretty pursuasive, because the next thing anyone knew, Course Coordinator was backpedalling furiously and demanded that Grad Student alter his pupil’s grade. When Grad Student argued, Course Coordinator pointed out that he was under no obligation to rehire disobedient graduate students as instructors for the following term.

(This never happened to me, by the way, but at one point I actually hoped it would. If I were ever told that keeping my job was contingent upon me dishing out high grades on demand, I would have gone straight to the media with my story. Whose job is on the line now, huh, Course Coordinator?)

While every marginally principled instructor will decry the attitudes of Kids (and their parents) These DaysTM, the fact of the matter is, it’s stressful to be continually under attack by students whose inflated grades are surpassed only by their senses of entitlement. There’s also the pressure of teaching alongside teachers who buckle under the pressure: should we standardize our grades to be in line with theirs, or should we stand firm and give out the marks we think our students deserve? My students were competing for scholarships and such with the students of an instructor who gave B’s for attending class and A’s for doing a modicum of homework. The committees assessing our students aren’t administering math tests to measure their abilities; the assumption is that we’ve already done this, and that the grades we gave reflect the students’ relative standings. The committees don’t know that the Nice Teacher and I could have wildly different standards. I ended up scaling grades at the end of the term to bring mine in line with my colleague’s, but I didn’t enjoy having to do so. Again - this is certainly an argument for some level of standardization.

In terms of preparing students for future academic programs, the job market, and such, inflating grades is a prisoner’s dilemma of sorts. If the other teachers are inflating grades, you should, too, so as not to place your students at a disadvantage. If the other teachers aren’t inflating grades and you do, then your students have an edge in admissions and scholarship committees; not only that, but they’ll also like you more. But it’s best for everyone if no one inflates grades.

And as any student of the prisoner’s dilemma knows, we generally can’t trust multiple parties to cooperate; someone’s likely to defect and inflate their students’ marks, and I believe the data bear this out. But are we instructors willing to relinquish our autonomy and allow a third party to verify that our students’ grades are fair? And if we do, is there any assurance that that will help?

I don’t know - do you?

(I can’t close this post without linking to one site I found in searching on the terms “Canada grade inflation”. The irony is truly a thing to behold.)

10 Comments

  1. One of the local private schools here is well-known for having their students marked out of 120, so if you get, say, a 90/120, you show up as having a 90%. Luckily here your high school marks count for very little (there are a few — but not many — selective Cegep programs).

    I had a physics professor who did the same thing: exams out of 105 (for my “honours” class) or 200 (for a friend’s not-honours). I was unimpressed.

    And the times I’ve had a class with curiously low midterm marks, I’ve had both very hard (physical chem) and very easy (astrophysics) finals. Actually, the astrophysics final was so basic I asked the professor if I was missing half the exam after skimming it.

    - wolfangel — 1/17/2005 @ 3:24 pm

  2. Do not misread these previous paragraphs to say that I think that uncertainty exists and that therefore, one can never really know anything about things in general and grade inflation in particular.

    Off topic, but a lot of geotechinal engineering is particularly illustrative of how plenty of useful information can be gleaned from relatively small amounts of data that one is fairly uncertain of.

    - Jen — 1/17/2005 @ 3:31 pm

  3. wolfangel, regarding your profs who gave unreasonably difficult exams and then scaled grades - here’s what I posted on Kimberly Swygert’s blog:

    Now that I’m a teacher, I understand why teachers would do that. By the time the final exam rolls around, I have a pretty good idea of how much my students know. If I design a test that’s unreasonably easy, then a very weak student might earn a passing grade or even a C, even if they were two standard deviations below the average. ON the other hand, if my test is unreasonably difficult, then *I* control the grades at the end of the day. If the highest mark was a 60% and I think that 75% of my students deserved B-’s or above, I can cook up the marks to reflect that.

    That said, I’ve been quite successful at designing reasonable tests; I tend to be able to engineer a class average within 5% or so of what I’d like it to be. In designing finals, I prefer to err on the side of too hard rather than on the side of too easy, but there’s a difference between making the test a bit harder than it should have been, and giving a test in which no one could answer more than half the questions.

    Jen - it might be off-topic, but I’d love to hear more. Any interesting references?

    - Moebius Stripper — 1/17/2005 @ 10:54 pm

  4. *Wow* Outstanding post. It’ll be included on the next Extra Credit.

    - EdWonk — 1/18/2005 @ 1:40 am

  5. Actually, I think the first exam (in both cases) was quite reasonable: I have no idea what happened, and nor, quite clearly, did the prof. I’ve had midterms which were deliberately very hard so that the profs could figure out which students were the very best, too, or ones where it was a first time teaching the course, so the timing hadn’t been added up quite right — but also ones where the marks just fell short of what would have been expected. (Students, too, can tell the tenor of a class.)

    I know I had some sort of point when I first posted this, but I can’t imagine what it might have been.

    - wolfangel — 1/18/2005 @ 6:42 am

  6. here you go:
    http://www.rocscience.com/hoek/pdf/Chapter%207%20of%20Rock%20Engineering.pdf

    - Jen — 1/18/2005 @ 1:02 pm

  7. oh, by the way, the case study discussed in the pdf I linked, is about a very quickly done analysis, the preliminary of which was done under emergency conditions to decide whether or not to evacuate immeditately. I forget if it says in the text that the preliminary analysis was more or less done over a long weekend.

    - Jen — 1/18/2005 @ 1:11 pm

  8. I’ve linked to this post to the latest Extra Credit, but wordpress doesn’t like Haloscan’s trackback.:) Here is the URL: http://educationwonk.blogspot.com/2005/01/extra-credit-assignment-great-reading_19.html

    - EdWonk — 1/19/2005 @ 12:41 am

  9. It’s not Wordpress, it’s MY Wordpress, which doesn’t accept trackbacking, period. I have no idea what’s wrong, argh.

    Thanks for the link.

    - Moebius Stripper — 1/19/2005 @ 1:03 am

  10. The Carnival Of Education: Week 1
    We are delighted to be hosting the First Edition of The Carnival Of Education. What we have done is assemble a variety of interesting and informative posts from around the EduSphere (and a few from the Larger Sphere) that have

    - The Education Wonks — 2/9/2005 @ 2:02 am

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.