Is there any way to avoid having to write all-new assignments every term for the classes I teach?

Ok readers, I give you plenty of advice, so now it’s payback time! 🙂 I need your help, with a problem I suspect many of you have: is there any good way to avoid writing all-new assignments every term for the classes I teach?

In order to save on paper and printing costs, I now distribute assignments online. I do the same for answer keys, for the assignments for which I make answer keys available (usually every assignment except the midterm and final exam). Which means that past assignments and their answer keys probably are circulating among some students in electronic form. So in order to minimize the opportunity for cheating, I have to write all-new assignments every term, which is a huge time suck. I’d really like to reduce or eliminate that time suck.

Years ago, I used to deal with this by only posting answer keys in a display case outside the classroom. I think some instructors here at Calgary may still do that, I’m not sure. But I don’t do it any more because students can just take pictures of the answer keys with their camera phones. So this approach doesn’t prevent answer keys from going into electronic circulation.*

Some possible solutions, along with their (often major) drawbacks:

  • Write new questions every term, but do so by modifying old questions just enough so that students can’t just copy old answers. This is more or less what I do, but it’s still a time suck.
  • Only ask questions drawn from the question bank that came with whatever textbook you’re using, or drawn from some other source. Obviously, this only works if there’s a big question bank available, and if you like those questions. I find I usually don’t, though there are exceptions.
  • Go old school: hand out all assignments in paper form, and don’t distribute or post answer keys (online or otherwise). You still go over the answers with students, but it’s up to them to take their own notes on that. Draw your assignments from a moderately-sized question bank (so that you can go, say, 2-4 terms without reusing any questions). Cross your fingers that students won’t compile answer keys from their notes and make those keys available to future students. And that if they do, those keys will likely be rather brief or garbled, and so of only limited help to future students.
  • Go to project-based assignments, with the projects being the sort of thing that can’t be copied from previous students. I have taught project-based courses. But I don’t think it’s feasible to teach, say, intro biostats to 125 students/term as a project-based course. And even if it was it would require a massive up-front investment of prep.
  • Never give out or even go over the answers. Just throwing this out there for the sake of completeness, I think this is a bad idea pedagogically plus it would really annoy the students.
  • Don’t worry about it. Reuse assignments as seems convenient, and rely on the honor system. Ok, I can hear you all laughing at me over the intertubes. But the idealist** in me wonders if maybe there are a few places where you can do this because there’s a very strong honor code. For instance, when I was an undergrad Williams College had a lot of unproctored final exams (as I recall; it’s been, um, a while). You were on your honor not to cheat, and if some students did, it was rare enough that I never heard about it.

I’ve asked around at bit at Calgary, and everyone seems to do as I do: write all-new questions every term, or in some cases draw from a massive question bank. But what do you do? Help please!

*Note that the issue is more serious for take-home assignments than for other sorts. Even if a student has access to the answers for past in-class assignments or exams, they still have to memorize and understand those answers in order to make use of them. Which from what I can tell is just as difficult for most students as learning and understanding the course material in the usual way. The issue is more serious if students can consult and copy from past answers on their own time.

**for idealist read: naive idiot?

26 thoughts on “Is there any way to avoid having to write all-new assignments every term for the classes I teach?

  1. In my experience, providing students with model answers makes no difference to the average quality of their submissions. Those students who would bother to read the model answers would do well anyway; those who tend to do poorly don’t bother to look at the model answers. Of course that’s a hugely cynical over-generalisation 🙂

    • I actually think that’s mostly true. But part of the issue here is that students might not think it’s true. I had an incident several years ago in which some students complained because I’d asked an exam question that had been used (and the answer posted) a couple of years previously. The students complained because a couple of other students apparently had seen that old exam and answer key. When I looked at the students’ answers, I couldn’t see any sign that anyone had benefited from seeing the old answer key. But obviously I couldn’t be sure, and so I had to throw the question out.

  2. You could also only rewrite some of the questions and if students do vastly better on the old ones vs the new ones, that might give reason to suspect cheating.

    • When I was a grad student TA at Rutgers, the profs used to catch cheaters on intro bio labs by making small changes to the assignments every year. For instance, having students grow corn seedlings instead of bean seedlings in lab. In their lab reports, the students would of course have to include the Latin name of whatever kind of plant they grew. Every year a few students would copy an old lab report–including the Latin name of the wrong organism.

      Like Jeff, in my experience most students who are smart and industrious enough to cheat and not get caught are smart and industrious enough not to need to cheat or want to cheat.

  3. …I don’t think it’s feasible to teach, say, intro biostats to 125 students/term as a project-based course. And even if it was it would require a massive up-front investment of prep.

    One element of the coursework for my 1st year biostats course is based on the students coming up with their own data. I spend a (very) little bit of time introducing them to databases, e.g., Dryad, during a computer practical session and then tell them they have to find their “own” data for one part of the class assignment, either from a project they previously carried out (e.g., at high school, or earlier that year in the biology course) or from a peer reviewed article (see Dryad). During another computer lab session, I chat with each student for a couple of minutes about the data and approach they’re planning to use (~140 students last year) to make sure they’re on the right track.

    Otherwise, my prep time for this question is 0 and marking time is relatively brief – all I have to do is take a quick look at their data, ensure they are applying the appropriate statistical analysis for that sort of data (choice of 3, which may increase to 5 next year), then check their working, presentation of results and interpretations (marks awarded for each of these elements). For the other questions in this assignment, I’m building up a bank. We have to supply a repeat assignment for any students who failed, which is allowing me to build up the bank quickly, but the investment for these other questions is certainly non-zero!

    I’ve run the course twice now (so my bank = 4 possibilities for each question). Some students have even repeated the entire course (around 30 I think), some of whom did worse the 2nd time round. I changed the rest of the content, but the “analyse your own data” section could have been repeated exactly by those students without any problems. They apparently didn’t bother to download/consult their notes/handouts and/or the feedback on the assignment they received from the previous year, which leads me to believe you (or at least, I) can relax about reusing questions from a bank!

  4. Not in ecology, so I don’t know what kinds of questions you need to ask, or whether this kind of approach applies, however, my solution is to ask questions that don’t have “an” answer, and for which _no_ answer is ever likely to be complete.

    My questions tend to be “synthesize and hypothesize based on the class material” type questions, and I end up using identical questions quite frequently. I grade based on a list of topics and ideas that I expect to see addressed, and the coherence of the argument or discussion the ideas have been used to construct. There are always significantly more ideas that I hope that they’ll hit, than any of them ever do, so a 100% score on an answer is somewhat less than 100% “right” on an absolute scale for the question, and any two 100% score answers are unlikely to be particularly similar.

    I use a really simple text-search comparison to catch obvious “just copied someone else’s answer” situations. I used to be some custom Perl scripts, now there are a bunch of online tools for this, but realistically the handful of students that I’ve caught who were dumb enough to try to cheat, were so blatantly obvious about it that I knew their answers were copied before I ran the check.

    This does, of course, run the risk that the grading is a bit subjective, but I haven’t found that to be a problem in general. It’s rare that the distribution of good, bad, and in-between answers doesn’t look essentially normal, and I’ve only had one or two instances in some 12 years of doing this where an answer was simply orthogonal to all the others, and couldn’t be reasonably ranked on fairly objective criteria.

    The thing that I rather like about this strategy, in addition to the ease of application (pushes more difficulty to the grading, obviously) is that even if they try to “cheat” by synthesizing from other exam answers, they’re _still_ doing some of the learning that they should have done in class, so heck, if I can trick them into learning by trying to coherently collate from other exam answers, I’ll take it!

    (and yeah – the biology lab experiments with “slight changes”… We caught more than “a few” that way. Whole quarter lab experiment, growing flowers under different light intensities – final exam question “What color were the flowers you used in your lab experiment?”. A full 30% of the students answered “Salt”. The previous quarter they grew plants in different concentrations of salt water… Some people do not deserve to graduate.)

    • Yes, for some courses or assignments asking open ended questions that test the students’ ability to synthesize a sophisticated answer is the way to go. Unfortunately it’s not a complete solution, I don’t think, since there are some times when it’s pedagogically valuable (for both you and the students) to ask simpler, less open-ended questions.

      I don’t recall if the slight changes to our labs at Rutgers ever “tricked” any students into doing something as stupid as answering “salt” when they really should’ve answered “yellow” or whatever. 🙂

  5. Hi Jeremy, last term I was faced with the same problem, and decided to use for a stats home exam uncorrected exercises from the textbook, Whitlock and Schluter’s Analysis of Biological Data (because I thought these were pretty good).

    It turned out that the “uncorrected” exercises (correction available teachers only) could actually be found online, digging quite a bit (there are robots that search for weaknesses in intranets, and somebody posted the answer keys after I tried to google it….). Fortunately, the answers were way less detailed than what I asked. Still that was pretty annoying.

    The next class I taught, I did not use a question bank and wrote the questions myself. Like you I found it extremely time-consuming. However, on the plus side, if you do it every year, it eventually makes a nice pile of exercises that the students can use as training for the exam (and I am sure one gets better at creating exercises).

    I don’t know how it is at Calgary, but in many universities we now have a student server where all the course material is supposed to be available: that makes the old school approach difficult.

      • Disappointing, isn’t it?

        W.r.t. the “don’t worry about it” strategy, doesn’t it depend more on the evaluation system than honor? If you have a continuous student evaluation where home assignements count, re-use is a problem, but if all the grade is based on a final exam, only the final exam needs to be somewhat new.

      • Re: the evaluation system, in the courses I teach it’s a mixture of take home assignments plus a midterm and final exam. And I wouldn’t really want to go solely to in-class exams just for the sake of minimizing the opportunity to cheat.

  6. Just a thought (that I stole from a publication that came out a month or so ago, actually), but what about having students create their own exam questions? You would have to review them to make sure that the questions were worthwhile, but that approach has the benefits of 1) reduced time suck compared to making your own, and 2) learning reinforcement by the students for whatever topic they choose. And an easy way to create small homework assignments, if such a thing matters.

    While I will admit to spending a lot of time on it, when I TAed chemical oceanography at Rutgers, most of my questions were of the involved, multi-part style that required a lot of concepts to be synthesized into one package. It doesn’t mean that the answers couldn’t be copied, but it would definitely take some thought to be able to put all of the pieces together. Depending on the question, you might not even need an answer key ahead of time, since students might take different approaches to answering a sufficiently open-ended question.

    • “what about having students create their own exam questions? You would have to review them to make sure that the questions were worthwhile”

      Have you ever done it, or known anyone who has? Because that sounds to me like it would be way *more* work than writing your own exam questions (though as you say it could have some compensating pedagogical benefits). But perhaps I’m just not sufficiently familiar with the idea to judge the time commitment required.

      • I’ve tried this in high school and it was definitely much more work than just doing my own questions. But it was done for the pedagogical benefits, so the time savings weren’t really a problem.

      • I tried this last year in an Intro. Bio. class of 30 students and it worked very well. At exam time, each student posted 2-3 questions to an online forum, one on each major subject I’d covered, so that everyone could study from those questions. Note: I did have to offer ‘points’ for posting. I privately collated the questions and transformed them into short answer, mult. choice, fill-in-blank, etc. The best part of this method is that students seem to study more earnestly with the idea that they are controlling what they learn and how they’re evaluated. I worried about the breadth of questions not being broad enough, but because I five, small exams rather than three big ones, this was not an issue. I also liked this method as a form of self-evaluation (for me!): I wanted to know, ‘What are they going to REALLY get excited about and take away from this class? Can I piggy back off those subjects to teach the less intuitive topics?’ and so-on. This method also is an opportunity to teach through evaluation. I was able to preface some test questions with a statement (say, about evolution, for example) so that students could say to themselves, ‘Well, if this preface statement is true, then such-and-such should follow’. Now I have written too much here! Apologies!

  7. You can either count me in the “hopelessly naive” camp, or in the “teaches at a small school and doesn’t know the realities of life in a large university” pool, or, maybe, “overly optimistic”. My approach is a combination of having students write own exam questions, doing most work in groups, posting last year’s exams, creating assignments and exams that require integration and interpretation, and probably most deeply and philosophically, being upfront about the “you get out of this what you put into it” basis of my course. I do make a big deal all throughout the semester that ecology (and biology as a whole) is increasingly a group enterprise, that in their careers they will most probably not work alone on any given project, and that science, like life, doesn’t have any single “right” answer.

  8. Would it be possible to do a low-stakes quiz at the beginning of class that draws from the homework? I’m going to try that for the first time this semester (with a 30 student upper div ecology course) to try to get them to do the readings. I’ve got a set of 5 questions associated with each reading and then I’ll ask them one randomly at the beginning of the discussion portion of class. If they don’t do the homework (or only looked at an answer key), they’ll probably do poorly on the quiz. We’ll see how it works out.

  9. I don’t have any useful suggestions, but did notice that some of the things Jeremy suggested would favor certain types of students. Evidently fraternities and sororities maintain exam banks. Thus something like the “old school” option could benefit students with easier access to people who previously took the class, if students are encouraged to contribute notes on exam answers.

    • Hmm. If by “favoring certain types of students” you mean “giving some students more of a chance to cheat than others”, then yes, I suppose so. But that’s not what I’d ordinarily think of as “favoring” some students. Perhaps I’m misunderstanding?

      • Yes, that’s basically what I meant (except technically it would not be cheating since they are just sharing notes they took in class). I was using “favoring” to refer to the method, not the professor.

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.