Evidence that students cheating on tests graded on a curve is stoopid.

Factual Questions
21

I noticed that when North American student joined my course in the Second year, the atmosphere changed somewhat.

We are not marked on a curve, and there is no limit to the number of first degree honours, or for that matter, fails. Your mark is your mark- therefore everyone swaps works, copies each other’s notes, and studies together, because helping someone else to do better has no effect on your own grade.

The North Americans had been used to being marked on a curve, and as a class we suddenly noticed that things which should have been passed around to everyone…weren’t. That trying to get notes from classmates if you had missed a lecture was very difficult if they were North Americans, and that they refused to join study groups, wouldn’t return much needed books to the library when asked, etc.

Once they realised that the curve wasn’t happening, and that the techniques they had learned to beat it no longer worked, things settled down to normal again.

I don’t think marking on a curve is particularly helpful to the atmosphere of the class. If you’re in a class with a genius, you can never get a good grade, if you’re in a class with idiots you’ll get a good grade you may not deserve, and it seems to enforce a very competitive, selfish and aggressive attitude.

Everyone should be able to do the best they can, and get the mark they truly deserve, it shouldn’t be relative. Learning is about doing the best to improve yourself, not about keeping others down so you can rise above them.

22

irishgirl, you were studying to be a doctor, right? You should know that that kind of cut-throat behaviour is limited almost exclusively to premed/medical students in the US.

I remember that in the physics library at my undergraduate alma mater, there were special arrangements for the notes and reference materials for the physics course that the premeds had to take. When they were checked out, they could only be taken to a special table that was about two feet from the circulation desk, so that the library workers could watch the students as they used them. Otherwise, homework solution sets would disappear and books would be vandalized.

You that kind of behaviour is almost unheard of—but it’s just expected from premeds.

Makes me feel reeeeallll great about the medical profession in this country. There’s either something wrong with the way we train doctors, or something wrong with the kind of people who want to become doctors.

23

Wow! That just…stuns me.
Here medical students have a reputation for having the best class spirit, the best parties, the best study groups and for being the best people to share notes and exam tips. As an example, within hours of final exams, every final med class compiles a list of the types of patients they got and questions they were asked, so the year below and those doing re-sits can profit from their experiences.

Another example: in the run-up to finals, each person in the class took a Consultant who was likely to set an exam question, and trawled through medical databases, looking for any research they had published in the last 20 years, then emailing the most relevant papers to the rest of the class. As it happens, one of the Respiratory Consultant’s pet topics came up on the final paper, but, thanks to this system, everyone knew what it was, and could quote his own paper in their answers!

I can’t think the people who want to be doctors are so different, so I suppose it has to be the method of teaching. If everything comes down to being in the top 10% in your class, and being top means stopping other people doing well, I can see why it would foster those kind of cut-throat attitudes.

24

Former pre-med here (long ago and far away). One of the reasons I switched out was that I was tired of the brutish competetiveness of it. My impression at the time was that there were so many people fighting for so few slots that the system had drifted into a hazing mentality to weed out those who weren’t “serious.” Overlooking, of course, that a club is a poor tool for trimming a hedge.

From speaking with doctors, this continues all the way through their training. Absurdly long hours in residency, for example. Maternity leaves measured in days, and single digit ones at that. Etc.

25

So that doesn’t happen in law schools, too?
The law students at my alma mater seemed pretty danged tense most of the time, so I had assumed they did that stuff too.

26

Yes, and the result is actually pretty obvious. It’s not going to be a Prisoner’s Dilemma because the end result is not necessarily sub-optimal. Imagine it pays to cheat; e.g., the odds and punishment of getting caught are minimal compared to the downside of a bad grade. If lower-grade students cheat up to higher grades, then there’s an incentive for better students to cheat as well. Because we have a curve that is effective, it necessarily follows that no more than 10%, let’s say, are getting perfect scores—otherwise, the test could not differentiate As from Bs.

If it is then worth it for the good students to cheat to regain their top spots, the poor students will be pushed back down into their previous positions. Ulitimately one of three things are going to happen, they will all cheat and just shift the curve upward—akin to every athelete in a sport doping: the advantage is washed out by universal drug use; they will push the grade up to the upper bound and the test will no longer differentiate students; or the system will be changed to prevent cheating. The tests in such a system would measure not only what a test is meant to measure, but also the morals and/or ability to cheat, which makes me wonder if there’s a lot of cheating in law school. A-hem. I apologize for that. Just a joke. I rely on attorneys every day and they really save my ass, and they do so honestly.

Now, suppose the cheating is limited to looking at the paper of someone around the cheater. Then there is no new knowledge being brought into the classroom, so in the limit, all students will have the same score as the best student. Again, there’s no differentiation, and what differentiation there is will probably come from ability to cheat and proximity to the smartest student.

(That would be in interesting exercise: determining who is the best student by the pattern of grades based on the links of communication from one student to another in the classroom. Maybe in that case, the best student will be hurt, because a student spatially far from the best who scores better than those around her will stand out.)

Stepping closer to reality, cheating also judges the “merits” of risk-taking behavior and lax ethical standards, so a student could benefit from cheating if nobody else does. Because we can’t really assume a real class of real students will really be identical in terms of risk & ethics, then we won’t have neutral results such as just shifting the whole curve upward. So, yeah, there can be a benefit because the person who cheats may boost herself above another student who does not cheat. Because of that, it may actually pay for an instructor to be very lax on cheating. In that case, a test sufficiently difficult should maintain some semblance of differentiation while providing cheating no net benefit. To improve the effect, submit several different tests of the same material, so that looking at a neighbor doesn’t pay. That is, have open-book tests.

27

Is this considered cheating? At my university, old exam papers from previous years were publically available in the library. They were a sensible tool for studying for the exams that pretty much everyone I know used. I used them a lot, they were useful for getting a feel for how well I knew the material and for the sort of tricky things I might see in the actual exam. It was up to the professors to ensure that they didn’t keep reusing the exact same questions.

28

Driver8- yeah, I’m with you on this. Old exam papers are posted on my college website within a few weeks of the exam (just the papers, not a marking guide or any answers).

If one professor has asked the same question for 20 years, that’s their problem, and if one professor chooses to put a worked example from a lecture on the paper, without changing any of the numbers (as happened to me once), well, that’s their problem too.

Surely someone, at some point, is going to repeat the class, or take the same class as an older friend or sibling, so if the Prof recycles questions, it’ll come out somehow, even without cheating.

My father-in-law is a Professor of Psychology, and his ploy is to work out what parts of the course his students think they can ignore, and then put that on the paper…that’s right, he uses reverse psychology!

I don’t think question-spotting is cheating, provided you don’t bitch and moan if you spotted wrongly and lost out as a result. If the Professor makes it very obvious what they’re going to ask, well, you’d be a fool not to focus the majority of your energy on that topic…if it turns out they were deliberately misdirecting you, it’s nobody’s fault but your own if you didn’t cover enough of the rest of the course to cobble a decent answer together.

29

Blasphemy! Cecil would somehow get 110 every time.

30

Could be. Prelaw students don’t all have to take physics, so I don’t have any direct experience with it, and I haven’t heard any stories of law students ripping pages out of a library book to ruin it for the next guy. But maybe they do.

I have personally taught lots and lots of physical therapy / occupational students. While they could also be described as extremely tense about their grades, they don’t have the same attitude of clawing their way to the top that the premeds do. They’re much more supportive of one another.

31

And this is a problem that needs to be fixed. If you and I do identical work and have identical understanding, we should receive identical grades. But what if I’m one semester ahead of you? In my class, there are lots of idiots, so the curve pushes my grade up. In your class, there are a lot of bright students, so the curve pushes your grade down. We have identical qualifications, and I get a better grade than you do. Why bother with grades in that case? Who cares how you perform with relation to the students who coincidentally end up in your classroom?

The purpose of grades is to determine whether you understand the material and have performed the required work. When I was teaching college, one of my classes required at least a “C” grade in one of the prerequisite courses before you could get in. I did that because the students coming into my class needed to understand the material presented in the prerequisite class. If that other class had been graded on a curve the number of students that could move on to my class would have been arbitrarily capped, potentially allowing students that didn’t understand the material to get into my class.

No, the GPA isn’t useless. It indicates, on an objective scale, how well the student learned the material. With a curve, it indicates how the student did against the random collection of students who happened to be in the same classes.

When I was hiring software engineers, I wanted to know if they’d taken courses in the languages and concepts I required and learned the material. Whether they’d scored well against other students that weren’t applying for the same job was immaterial.

32

That kind of competitiveness doesn’t show up in a class for physics majors, even if grades are curved. I suspect at least one reason is that physics major programs aren’t generally trying to weed people out. I think the weed-out mentality is the main problem, not grading on a curve.

Yes, one of you was obviously cheating off the other, so you should both fail :stuck_out_tongue: Seriously, if two students had done absolutely identical work when I was a TA, I would have thought it was cheating. Unless the entire class grade is based on multiple-choice or true-false tests, objectively identical work just isn’t going to happen unless there’s cheating.

There’s generally no reason to assume that there should be a significant difference between people taking a class from one term to another. What’s much more likely to vary between terms is the quality of teaching- if there are different professors, for example, or the professor spends more time on different material with different classes, or if the professor is just human and is more on top of his/her game one term than another. You couldn’t compare objectively between classes, because no two professors are going to teach exactly the same material in the same way. One professor might even teach different material from term to term (especially in a rapidly changing field, like some areas of astronomy or computer science).

It doesn’t really make sense to say that a 95 in Professor X’s planetary science class in 1994 (the year before extrasolar planets were discovered) is the same as a 95 in the same class taught by Professor Y in 2005. The students aren’t going to know the same material, because the field changed between those classes, plus they were taught by professors with different ideas about what’s important in the field.

33

A friend of mine wrote a paper (warning: pdf) on an empirical study of the effects of anti-plagiarism software.[sup]1[/sup] The discussion on page 836 of the students’ reactions to the effect of plagiarism detection on the scores of noncheaters was particularly interesting, I thought.

  1. Braumoeller, Bear F. and Gaines, Brian J. "Actions Do Speak Louder than Words: Deterring Plagiarism with the Use of Plagiarism-Detection Software " The Teacher, Dec. 2001
34

Okay, I apparently misunderstood how the grading system worked (I’ve been out of school a long time). If I now understand it correctly the curve system basically works that the instructor has a predetermined set of grades he will give out and then ranks his students based on their actual marks and assigns them the grades from the predetermined set. Is this correct? If so, then a cheater is definitely hurting many of the other students - he is, in a sense, cutting ahead in line.

35

Bingo.

36

Hey Anne Neville. I indeed meant that the problem was the weed-out mentality, not curving as such. Sorry if I was unclear.

As for pre-law v. pre-med, understand that there are many pre-law tracks, so neither the weed-out mentality nor any other form of competetiveness comes into play. Relatively few people in your particular track will be pre-law, i.e., most will be planning to use the major for some other purpose. And pre-laws in separate tracks don’t run into each other much.

In law school itself there is some competetiveness. In my school, at least, all classes were graded on a curve (except legal writing, which was pass-fail). But it wasn’t cutthroat and most students ended up in some study group or another. This may have been because there was little-or-no secret material to be hoarded. Everyone was working from the same course books and lectures. Also, first year classes (the grades for which are the most important) are large, so there’s relatively little direct A-bumps-B concern. You feel more like you’re struggling against the material than against each other.

37

This correctly describes the simplification many have been using in this thread. It gets the point across, but this system has nothing to do with a bell curve and requires no statistical analysis (it just involves ranking the scores from lowest to highest).

I believe a true curving scheme would be to fit the distribution of raw test scores to a normal curve, and assign grades based on position in the curve. So, rather than giving A’s to the top 10% of test scores, A’s are given to any test lying withing the top 10% of the bell curve (in terms of area under the curve).
NOTE: normal curve, bell curve and gaussian curve are all the same thing.

This brings up another question: Is it possible for all students to pass a curved test? In the scenario I’m thinking of, the distribution would have a maximum right at the lowest test score, and decrease as test scores go up. So essentially, the distrubution would resemble the right half of the normal curve. Since there are no tests below the midway point, the lowest grade should be around a C. Is this correct?

38

Irishgirl: In the US, both Law and Med programs are highly competitive, brutal affairs, where students learn highly competitive mores, reinforced by course which typically are viewed as “weed out/filter” course, rather than courses for their own sake. Law and Med students typically graduate with student debt of at least $150,00 USD. It is not a pleasant process, but some graduate with their humanity intact. These survivors, of course, are then further brutalised by badly paid internships and residencies. The lawyers, however, do tend to earn more earlier.

The purpose of a course is to learn the material. I leave the comparative brainiac ratings to the IQ fetishists.

In a pedagogical model, a grading rubric based solely on objective mastery would serve as an IQ sifter … without the fallacy of a relative ranking read as a meaningful rank.

In any course, the learning objectives vary in complexity and difficulty. Provided that you design the course in a way that spans the straightforward to the sublime to the downright bloody difficult, imposing the unblinking objective standards, you’ll get your Non-stoopidity Index(NSI).

In short, the high scorers with regards to NSI will consistently show superior objective mastery of the “hard bits”…

And all that without a stupid curve.

39

If they aren’t available to everyone, then yes. They were not generally available and the teacher (I can’t bring myself to call this hack a professor, although he probably was at least an associate prof) made a big deal of collecting up the actual exams afterwards.

40

Mathocist, you’re right… I starting talking about explicit curves, and then was trying to show that you still hurt other students even if it’s not curved… because there’s a larger implicit curve in that people making use of the grades will likely choose the top fraction. Very poorly phrased. I hope your rage has subsided.