Mind Games

[stoutfellow]

Classroom tests are supposed to measure students' knowledge, skills, and understanding of the given material. Unfortunately, they'll always try to game the system, looking for patterns that allow them to avoid having to think. For instance, if a test has six True-False questions, they'll expect three True and three False, and use that to leverage such knowledge as they have.

I refuse to let them get away with that. In my test yesterday, question two was in two parts. Part a) described a situation, then asked a question with two possible answers. Part b) modified the situation slightly, and asked the same question again. A lazy student would conclude that the two parts have different answers. A decent but hurried student would, perhaps, be deceived into thinking that the difference between the situations is larger than it actually is (and I worded it so that it looked that way). A good student... well, about fifteen minutes into the test, one of the better students raised his hand. I walked over, and he murmured, "Did you really mean to say 'y-axis'?" I said, "I said 'y-axis', yes." He shook his head and said, "O-kay..."

You can't pull this too often; the idea is to break up patterns, not establish new ones. Probably this could be analyzed game-theoretically to determine the best strategy, but I'm guessing that randomness is the way to go.

Comments

[toraks.livejournal.com]

So the answer is actually the same for part a and part b?

That's definitely tricky!! I assume I'd have caught it, but perhaps I'm overrating myself!! :-D

Though, in my defense, I do remember something of that sort being pulled on me before, and I do think I caught it.

[stoutfellow.livejournal.com]

So the answer is actually the same for part a and part b?

Yes. The thing is, if the two have different answers, and the students know they are guaranteed to have different answers, then there's no point to asking the second question. So I make sure, this way, that the students know there's no guarantee. (I could have outright said that there's no guarantee, but then we get into I-know-that-he-knows-that-I-know territory.)

[oilhistorian.livejournal.com]

Ah, the joy of test-writing. Back when my survey sections held 300 students, I used multiple-choice tests. My students would search in vain for a pattern to the key. Frankly, if 7 answers in a row were "C", I'd let it stand. I can't tell you how many Aggies were certain that one of those answers had to be wrong because "No professor would let 'C' be used that much."

[filkferengi]

You're definitely Cute When Sneaky. ;)

Heh.

[carbonelle.livejournal.com]

And I always tried to just Do The Math.

No wonder I stalled out after calculus.

[commodorified.livejournal.com]

Hee. I used to teach a small wandering lecture, by invitation, on multiple choice. It was not, mind you, on how to game the system, but rather how to cope with the basic format and the inevitable badly done multiple choice tests that do turn up.

Basic stuff: do the ones you know first.

Don't go back and outguess yourself. Unless there is a penalty for wrong answers, answer everything. If you are going to guess, read all the options carefully. One of them is probably really stupid. One more will be unlikely. THEN flip your coin.

Do the scantron sheet last, and concentrate on it.

Read the questions carefully, as if they were in a different language. In case they are.

Talking of languages, the University of Ottawa permits bilingual (translation) dictionaries during exams; use them.

Just because you are in theory being tested in Soc of Religion doesn't mean that you can't blow a question through not catching an awkward double negative or deliberate trick.

If there actually does not appear to be a correct option and you are quite certain you've read the question possibly, ask the proctor if there is a misprint. They happen, and sometimes they throw out things that look like sense and are not. Also, professors who deliberately insert obscure jokes, awkward constructions meant to mislead, or downright traps in their questions instead of just making the damned things test you on what you are supposed to know deserve to spend the exam period on their feet.

Grades usually went up when I did that little talk before the midterm... as much as I object to students trying not to think, there is a case for actually teaching the exam formats; *I* find multiple choice insanely easy, but I have come to know that many perfecly bright people who have in fact studied panic at the sight of them.