next »

[professor_pat]

I've reviewed about ten previous threads on take-home vs in-class exams, all of which have good suggestions. I haven't assigned a take-home exam in many years, though, so it'd be great to get your feedback on my current idea about how to do this.

This is a midterm exam in an interdisciplinary social-science course. Our class period is two hours. I'm leaning toward a test rather than a paper because (a) I'd like to have all students address the same questions for comparative purposes, and (b) students will already have a term paper and a couple of presentations to be working on, so a test seems more time-efficient.

Students will have turned in weekly response papers to the assigned readings, which has worked well in the past as an incentive to actually read the material, so they will presumably have already thought about each reading. The purpose of the exam will be to get them to synthesize the readings and consider relationships among them.

I'd like the exam to be open-book, open-notes whether I do it in class or as a take-home. So I'm thinking that I could give a list of eight potential questions a week before the test would be due, then post the actual questions (probably a choice of two out of three) 24 hours before the test is due.

Ideally I'd like to limit the time each student can spend on writing the test -- maybe 4 hours? -- in an attempt to equalize effort among students, but I can't see how to do this realistically. (I could state that each student has 4 hours to complete the test once they open the questions, and could verify that via our CMS, but of course the first student to open the questions could just pass them around to everyone else.)

Any comments on this overall concept? How many words do you think would be reasonable to expect students to write for this type of midterm? Any caveats or suggestions? (Aside from the plagiarism potential, of course, which I'll both caution students about and attempt to foil via narrowly worded questions.)

[antiphon1]

Does your CMS have an option to randomly select sets of questions from a pool?  You might be able to post the potential questions then let the CMS randomly select 2 - 3 questions from the pool for each student. 

[professor_pat]

Hmm. I'd thought about that and decided students might feel it was "unfaaaair" somehow, but now that you've suggested it I wonder whether that might actually make good sense. That would also help diminish collaboration on answers (although they still would have the overall set of questions, so could collaborate that way).

I assume our CMS has a way for me to tell which student pulled up which questions, and that the student can write on their computer and then copy/paste the answers into the CMS answer box - will have to check with our IT folks about how to do that. I'd still kind of like to offer students a choice of two out of three questions; maybe I could get the CMS to pull up three random questions, then tell students they'll only be graded on two.

[antiphon1]

I do something a bit similar with the short answers for online tests. The students have a list of the topics and types of questions but not necessarily the specific questions they might be asked to answer on each topic.  I then pair topics in compare/contrast, examine the relative significance of or which of the two topics offers a better explanation for _____ types of questions.  The CMS then randomly selects the set number of questions for each student some of which may be very similar but not exactly alike.  I do get some complaints from the students who try to copy a buddy's answer to similar question onto their test.  Griping I can stand; cheating I can't abide. 

[2clueless]

Whether students who take the exam early share the questions may depend on your institutional culture. This type of exam was quite common at my undergraduate institution, except the exams were typically closed-book, came in sealed envelopes, and our professors didn't have the benefit of a CMS to see what time we opened the questions and what time we finished the exam (and yes, we walked to class barefoot in the snow, uphill both ways. Well, okay, it was uphill both ways from my dorm...). This institution has an Honor Code to which students adhered, in my experience, and I never heard of anyone sharing questions. OTOH, Colonel Mustard tried to give a similar type of exam at a different institution, one that does not have an Honor Code and has a culture of accepted "cheating" (e.g., frats keeping exam files, etc.) as a way to get ahead. When reading RMP, he discovered that students were cheating by taking the exams open-book and untimed.

Two suggestions:
1) Have students sign a contract in advance that they won't share the questions and then have them write a statement on their exam that they have not received any outside help and will not/did not share the exam questions, then sign it. I typically do this for assignments/exams that exclude outside help; in talking to students, they reported that it helped squash the temptation to cheat by increasing their guilt levels. YMMV, of course!
2) From personal experience as a student, I think it's a little unfair to give students only a 24 hour window to complete a four-hour exam. You know your institution, of course, but as a student, I often had/have set commitments on certain days of the week that would have made it difficult to find a four-hour window in a 24 hour period, such as a long work shift on a particular day, musical rehearsal, etc. In a typical week, I was able to plan my schoolwork around a 10-hour work-study shift on Tuesdays, but an unexpected 24-hour turnover that required four hours would have been very hard for me, at least without pulling an all-nighter, etc. Of course, giving students a longer window might increase the chances that students would share the questions...

You can warn them that they will be hurting themselves if they pass the questions to their classmates, of course, and point out that if the exams completed later are stronger than ones completed earlier, you'll be able to tell that they cheated. Personally, I'm not good enough/experienced enough to track patterns like later exams being stronger than earlier ones, but there's no reason to tell students that!

[professor_pat]

Many thanks for the responses. I wasn't clear about my rationale for the time-length of the exam, sorry. It would be the same exam as I'd give in a two-hour class session, but I would give students double that time so they can relax a bit and actually think. The 24-hour window to take the exam would include our scheduled class time, so that all students should be able to find a good window of up to four hours to take the 2-hour exam (if that makes sense). Given that new info, does the 24-hour window now make sense? Or is that still too short? Or should I limit the actual test time to just the 2 hours that include our class period?

Over the years, I've tended to move away from trying to strongarm those requirements that cannot be enforced, like collaboration on a take-home test. However, I do think that getting students to sign an honor pledge that they haven't shared questions and haven't collaborated on the answers probably makes sense.

This is a really small class (12), so it'll probably be pretty clear if students have worked together on their answers. The ethos in this year's group seems good, though, so hopefully this won't be an issue...though I'll surely watch for it.

[antiphon1]

A 24 hour window seems enough to me provided you are giving them the study materials at least a week in advance.  The four hour testing time sounds fine because of potential connectivity problems, too.  I'd do 3 hours, but that's me.  My sense of the online classroom is that students are more successful if forced to focus on the task at hand rather than being allowed to wander in and out of the material.  Again, that opinion is limited to my experience with the format not any in depth study.

[polly_mer]

As a student, I hated the 24-hour window for reasons mentioned upthread.  However, if students know well in advance that they need to find at least two hours (and class time is in that 24-hour period), then that seems fine to me.

I have used the pool-of-questions-with-a-random-selection-for-each-student technique with good results and practically no whining.

[burnie]

Hmm. I'd thought about that and decided students might feel it was "unfaaaair" somehow, but now that you've suggested it I wonder whether that might actually make good sense. That would also help diminish collaboration on answers (although they still would have the overall set of questions, so could collaborate that way).

I assume our CMS has a way for me to tell which student pulled up which questions, and that the student can write on their computer and then copy/paste the answers into the CMS answer box - will have to check with our IT folks about how to do that. I'd still kind of like to offer students a choice of two out of three questions; maybe I could get the CMS to pull up three random questions, then tell students they'll only be graded on two.

It won't discourage collaboration, necessarily, but it will mean that through collaboration they are exposed to and forced to consider more questions than the ones they're actually being tested on.  I actually like this method for just that reason.

[wet_blanket]

What is the disadvantage of letting people take 24 whole hours if they are so inclined?  Sure, some people may have only 2 hours and others 6 or 16 hours free in the window, but if it's a test they could take in two hours in class, the extra time isn't that much of an advantage.  Someone who needs 10 hours isn't going to score as well as someone who could do in 3 hours.

[mixedmetaphor]

Definitely repeat over and over that the reason why you have a narrow window is because this is something that you'd normally expect students to complete in the two-hour class period.  Let them know explicitly and frequently that you are allowing them to work in a timing window that works best for them, and that you are actually being very generous by giving them an extra hour or two that the in-person class meeting wouldn't afford.

I only urge this Jedi mind trickery and repetition because I did something similar in a literature course when a scheduled final exam timeslot was 7:30 am on the last day of exam week.  I gave the students 48 hours to complete the essay-style exam and told them that they should only need to take about two hours since that's as long as we'd take in class (and I'd used pretty much the same exam in an earlier semester with a less hideous timeslot, and not one student had a hard time completing it in two hours).  Even with all of this explanation, one student would NOT stop complaining about "only" having 48 hours.  I thought I'd explained myself and my rationale pretty clearly, but he STILL wouldn't/couldn't get it.  He even complained at length about it on his eval (completely misrepresenting the exercise, of course).

As an aside, he was also the most furious nose-picker.  And he wiped his fingers in his books.

[hiddendragon]

My last exam, I gave students 24 hours to do it.  They can do it anytime and use all available sources.  I stipulate that they can't get others input.  I'm sure some do not listen and consult with one another, but this stipulation makes it so they can't complain and say, "But my mom read it before I turned it in and she thought it was brilliant so how could I get a C?"  It also makes it so that students who write the same things--exact same answers--can be called into question.  Beyond that, I really can't control what they do and whether or not they collaborate.  But, if they have to collaborate, I don't think they will do well anyway.

But I did have two students complain that 24 hours was too little time for them to compose an answer because they have other exams and they work.  I was a bit surprised.  I thought anyone would rather have 24 hours than just 2 hours in class.  Even if they don't have the whole 24 hours to work on it, they would still compose a better answer at home in 2 hours (on a computer) than in class while cramming with pen/pencil on a blue book.  And, maybe they'll have 3 hours free, or 4, or 5.  AGain, still better than the mandatory scheduled two hours with pen and pencil.

[scampster]

But I did have two students complain that 24 hours was too little time for them to compose an answer because they have other exams and they work.  I was a bit surprised.  I thought anyone would rather have 24 hours than just 2 hours in class.  Even if they don't have the whole 24 hours to work on it, they would still compose a better answer at home in 2 hours (on a computer) than in class while cramming with pen/pencil on a blue book.  And, maybe they'll have 3 hours free, or 4, or 5.  AGain, still better than the mandatory scheduled two hours with pen and pencil.

If you grade absolutely, it shouldn't matter, but I think it is hard for students to get the idea out of their head the idea that "Sally got the whole 24 hours, while I had to work and take another exam, thus she will do better than me, and that will make me look bad, and I will get a lower grade!" Depending on the way you grade, they may have a point.

[professor_pat]

OP here. Bumping this thread with an update and a followup question.

When I asked the students at the beginning of the term whether they preferred a take-home midterm configured as discussed here, or an in-class test, they overwhelmingly voted for the take-home and were really excited about that option. The test opens this afternoon for 24 hours, and everyone still seems fine with the format.

It's interesting that according to the CMS, only about half of the students have even viewed the question bank I distributed a week ago, even given my caution in class that they were going to want time to prepare, since the questions are pretty complex.

Remember, the setup is that students were given a set of 8 questions a week ago, and the CMS will randomly choose three questions per student.

One student asked for guidance about length of answers. Since this test is like the one I gave two years ago using blue books in the classroom, I don't really have a feel for word count for typed versions of the answers. I had thought that since students have had a week to prepare their answers, they might be substantial/on the long side, but now I realize that at least half the students are probably planning to just open up the questions when they sit down for 2 hours during our regular class-period time, and to begin thinking about their answers in that time period.

Not that I should let students' lack of preparation guide my thoughts about length - but it'd be great to get your advice about what kind of word count might be reasonable to ask/expect of students in this situation.

[polly_mer]

Two hours with a week to think means I'd be expecting about a typed page per answer.

For perspective, I expect a handwritten half page on daily quizzes without seeing the questions first.