Friday, March 25, 2011

Giving everyone an A is a terrible idea.

A few days ago, Cherish the Scientist tweeted out an article (@mareserinitatis) on how Everyone Should Get an A. David MacKay brings up a few interesting points in the article. These are largely centered on two themes. The first is the rate at which a student learns in conjunction with their starting position. If everything is linear, this is the classic Y = mX + b where Y is the level you’re trying to obtain, m is the rate at which you learn over time, X is the time it takes, and b is your offset starting position. (Yes, whenever you use an equation, you should explain the variables…). The second theme is there should be a minimum obtainable threshold that students should achieve. Thus, if students are “forced” to achieve an A to graduate, they should be allowed to take longer. The first issue is persuasive but has a problem when applying it to real life. The second argument, on the other hand, has some serious holes.

I don’t have any hard evidence for this but the thought process should fit match most anecdotal preconceived, prejudiced notions. Your b value when you begin in college is a very important aspect. If students took more AP classes, generally they’re going to start with a higher b. The same could be said if your parents were highly educated and/or come from an upper class economical situation. Students who aren’t fortunate to be in those situations often have much lower b values and generally have other issues to contend with once they get to college. Working a job on the side to support themselves comes to mind.

In principle, if you could test a student’s knowledge in the beginning and in the end of a course, and come up with a slope-o-meter, that would be a good to obtain an accurate picture on how someone learns. That would be very good for assessing the effectiveness of a teacher and identifying the potential of the student. This, in theory, could go a long way towards helping the educational system improve how it is educating its students.

There is one slight, 800 lb gorilla sitting in the corner of the room. Life doesn’t work like that.

Figure 2 in MacKay’s paper shows a prime example of this. Let’s assume you have some miners trapped in a coal mine and instead of it being exam-time, it’s actually people-are-going-to-die-time. Who do you want deciding where to drill the hole? A, B, or C? If it’s my life, I’m going with A because they, at this given point in time, know more than the other two. Sure, eventually B and C will surpass A. But now you’ve got a bunch of explaining to do to a family on why their loved one died.

Look, it sucks that B and C started out in a more disadvantaged position. It can certainly be argued that is a combination of social, economical, and educational factors. And yes, society as a whole should strive to correct those. But if we’re talking about university and graduate level education, you have to draw the line at some point. This brings the second theme, allowing indefinite time to finally get your A.

If you’re allowed indefinite time to get an A (or whatever you want to call the achievable mark), you’ve just given no incentive to learning how to get stuff done. This is a huge problem at OldEuropeU where the students are allowed to take a class as many times as they want to get a passing grade (6 out of 10 BTW, not even an 8 or 9). In an educational system that is nominally free for the students, that’s a huge overall tax payer drain. I have colleagues that took 10+ years (in a 5 year program) to get a BS and a MS and they spent that whole time in school. They didn’t take time off for work or family or whatever. They just spent too much time partying. If they were in the US system, they would not have been engineers after year two. They would have failed more than the allowable amount of classes, had an insanely low GPA, and would have been kicked out of any ABET engineering program. I’m sorry but not everyone should be an engineer just because they want to.

In the real world, you have to deal with meeting deadlines. It sucks. Everyone hates it. And it’s a shame that we don’t have a culture of saying “No, we can’t move on because this thing isn’t ready yet. We haven’t learned enough to make this product fully do everything we wanted.” Everything from software (security flaws), to baby strollers (safety recalls), to cars (stick accelerators and faulty air bags), to food (E. coli and salmonella poisonings) has their issues because of deadlines. But, unless you’re going to change how the world works, people cannot be afforded indefinite time to do everything. Nothing would get done then.

So no, I don’t think everyone should get an A.


  1. Oh, there's so so much wrong with Mr. MacKay's proposal. Gears, you hit a big one, the the real world works on deadlines. I'm thinking a little more about how students learn. I think it is very assumption to think that students will learn if just given the time. Another interesting one is that the only reason students do bad on test is stress? And the big one, that better grades indicated better mastery of the material or that better grades indicate better engineers in the workforce.

    See, I don't think there is a difference in mastery between a B and A. I think B's get you 90% of the way there and a A is just the window dressing.

    I'm still stuck on his idea that all you need if more time and less stress and the student will learn everything... Really? I just don't think everyone can learn everything. As you say, Gears, this is the Old Europe model, and I don't like it.

  2. I agree completely with the author of this article. Mr MacKay's theory is very weak and it is not something you can apply in the real world.