Wednesday, April 07, 2021

Giving Math Tests In A Hybrid Model

Each morning I see anywhere from 1/8 to 1/3 of my students per class in person, while the rest continue taking the course via Zoom in the afternoon.  If I choose to give a test, how do I ensure that the at-home students don't get advantages (except, of course, for easier cheating) that the in-class students don't get?  If I had to defend my test-giving against charges of a lack of equity, how would I do so?

Tomorrow I'm giving a stats test, and next Thursday I'm giving a pre-calculus test.  Here's how I'm handling each of them.

Tomorrow's statistics class must be done during class time.  In-class students will get one version, on paper, and will do the test during our 50-minute class.  At-home students will get a different version, and it will show up on their Google Classroom page 5 minutes before class starts.  They have until 5 minutes after class to submit it.  This gives them plenty of time to print the test out (for those who do that), to scan their tests when done, and submit them through Google Classroom.  As I cannot prevent the at-home students from looking at notes, the book, etc., I will allow the in-class students to do that as well.

That isn't very creative, though.  With next week's pre-calculus class, though, another teacher and I are working together to get a bit more creative.

This test will be multiple choice, which is something I never do.  In-class students will get a paper copy and an answer sheet, which I can scan by document camera as soon as they turn it in and immediately give them their scores.  At-home students will take the test on their computers, and thus will need no extra time to submit their tests, but I am going to require them to submit their work so that I can give credit only for correct answers that are supported by work they did (an anti-cheating measure).  The online test administration can be set so that the students can take the test only during the test window I designate.  Also, the multiple choice answers will be written so as to prevent "process-of-elimination guessing" and other such issues.  And of course, there will be different versions of this test.

My tests were designed during pre-'rona "real school", where classes were 50-55 minutes each.  Thus, they're designed to be completed in 50 minutes.  However, due to the 'rona, I've been much more lax about time constraints--that might have been a mistake on my part, and not just because of cheating.  Now that I've had students back in class, I see that they've learned even less than I'd thought.  Too many haven't "engaged" in their learning, they have at most listened to what I've said and watched what I've done, and then during tests, they try to follow examples in the book or their notes.  Too many of them don't "own" the knowledge they're trying to demonstrate.  I shall endeavor to correct that.


Ellen K said...

Online quizzes became one of my biggest nightmares. Yes, they were easier to adminnister and grade, but the amount of cheating between the two classes of AP even when I did multiple versions (at one point I was up to 8) was appalling. I had to resort to doing hybrid tests with 30% multiple choice and the rest essay. Of course that's unworkable for math and most science calculations. Testing has become a real problem with online classes. I remember finding one of my quizzes on Twitter.

Anonymous said...

I predict the in class students will get screwed because all of the remote ones will cheat. It's insane how much cheating is going on these days; I don't envy you the task of figuring out which is which.

Even if you let them all use notes, the online ones can call each other or even just use one of the many problem solvers out there--all of which will generally "show steps."

Now, IF (a big if) you had the same students both in school and remotely, you could probably gin up some stats and tests to figure out which ones were cheating--for example, including the same problem (with different #s, etc.) You should look into that, perhaps.

Darren said...

You can bet that my fellow math teachers and I are very much attuned to the cheating situation, and take proactive steps to eliminate the effectiveness of cheating.

What I've found is that *most*--yes, *most*--students haven't learned very much at home. They've been passive learners, listening only, not doing or questioning. Those that have done well on tests have done so because I gave large amounts of time to do tests, meaning they could look things up and figure out how to do something on the fly without really grokking what they're doing. One way of limiting cheating is to require strict adherence to timing--the test will not appear on the web site until this time, you have 50 minutes to take the test, and it must be scanned/submitted within 10 minutes after that. Or else. Given the way I write test problems, that doesn't allow much time for looking things up.

There's not much I can do when the smart kid takes a picture of his test and sends it to friends. But I've caught people who've done that, and everyone knows the penalty is pretty severe. I plant my flag on that hill and do *not* waver. I've had conferences and threats of going to the school board, etc, but when I give a 0 on a test for cheating, you can be sure that the evidence is overwhelming. Every one of those zeroes has stood.

Anonymous said...

yes, timing helps. So does on the spot stuff:
If a student can't manage to factor x^2 - y^2 when called on, it's probably fair to question their tests which show successful factoring of much harder polynomials. But then again I suspect you probably don't have the ability to force folks to turn on cameras, much less to force suspects to come in and take a proctored in-person test to prove they weren't cheating.

It's a fascinating field, really. I assume you've probably read Freakonomics but there was a whole chapter about how they caught some cheating teachers.

It sounds like you're generally on the right track. I'm a lawyer and catching cheating (aka lying) is a lot of what we try to do.

It's mostly about repetition: Someone can fake knowing X once, but it's amazingly hard to remember what, precisely, you claimed to know (and why) so that you can consistently lie with complexity.

If you can, post your results! It's an interesting read.

Darren said...

All of your assumptions are correct!

PeggyU said...

I was going to suggest turning on cameras. Why is it you are not allowed to do that? I know that with my sons' online courses they used a proctoring service. Son had to provide valid photo ID to use it (gasp!), and proctor also asked to see the immediate environment to make sure there weren't cheat sheets or other prompts available.

Darren said...

It's a privacy thing--it might bother some people to have others see the insides of their houses.

Yes, I'm serious.