Academia Stack Exchange is a question and answer site for academics and those enrolled in higher education. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I teach a writing course to different sections, across different days of the week. From the schedule, some students will take their mid-term exam early in the week, while others will take it later in the week. Recently, I found one student who was “just checking to see what the exam was like”, but was planning to go to another session to actually take the test. I realized there might be a variety of ways that students taking the exams on Friday would have an advantage over those taking it on a Monday.

I've taken the following steps in an attempt to address this problem:

  • I build multiple tests, e.g. "A", "B", "C", etc., which are of similar difficulty, have the same types of problems, but different subjects. "A" is given to one section, "B" to another, etc.
  • Provided a sample test that students can print and practice with and discuss with me in class.

Note that, there are several challenges to administering this:

  • The classes are large, and the school offers me no assistance in managing the exams.
  • Many students regularly attend my lectures during sections that they did not register for, so I will not easily recognize who belongs in which time.

Are there any additional steps I should consider to make sure the test is administered fairly?

share|improve this question
Do a real exam timetable, so that all students that take an exam does it at the same time/day regardless of the section that take the course in. It is not hard, the software to workout the exam timetable is on the open market. – Ian Jun 12 '14 at 15:09
@Ian that's not allowed at many universities. You have to get special permission to do that at my institution and it's really only given to certain freshman level courses for finals. Space is one of the hardest considerations. – guifa Oct 5 '15 at 23:37
@guifa, is it a space problem, or unwillness to buy good exam scheduleing software? UK universities cope, buy they tend to think that it is important to get that sort of thing right. – Ian Oct 6 '15 at 7:45

There are various ways of preventing students from showing up on the wrong day or showing up more than once. The first step would just be to tell them straightforwardly that it's not allowed. You could also assign seats for the exam. Assigning seats has the added benefit of reducing opportunities for copying from a neighbor's paper; they can't just choose to sit next to a friend or sit next to someone they know is doing much better in the class.

I teach a writing course [...] The classes are large, and the school offers me no assistance in managing the exams.

The enrollment of writing classes taught by a single instructor is usually limited because it's not practical for the instructor to grade large amounts of written work from a large number of students. At some schools, writing classes are large, but there are TAs who read papers. If you're a single instructor teaching a writing class to a large number of students, then that's a problem in and of itself; these issues with exam arrangements would then be a symptom of the more general problem.

share|improve this answer

Many students regularly attend my lectures during sections that they did not register for, so I will not easily recognize who belongs in which time.

This is handled as follows at my university:

  • Students entering the exam room go to the exam proctor before they sit down
  • Exam proctor has a list of students enrolled in the class
  • Student shows a photo ID to the proctor
  • Proctor notes on the list that the student attended the exam on that date, gives the student an exam book, and tells him where to sit.

This way, students can attend only one exam (and if you want, you can enforce that they attend the section they're assigned to).

I build multiple tests, e.g. "A", "B", "C", etc., which are of similar difficulty, have the same types of problems, but different subjects. "A" is given to one section, "B" to another, etc.

This is OK, although if students aren't sure what types of problems you're going to give they'll still gain an unfair advantage if they hear about the exam from earlier sections. However, you can avoid this if you...

Provided a sample test that students can print and practice with and discuss with me in class.

Giving plenty of sample exams ensures that students thoroughly understand the type of questions to expect. Then, if each section is given a different exams of the same type, but different problems, there is no advantage to hearing about the "type" of the exam from earlier sections.

Sometimes you can avoid having to build many exams by giving everyone about 80% of the exam "content" ahead of time. Then vary the remaining 20% between sections.

For example, in a writing exam where students have to read some text and write an essay in response to a "prompt": give everyone the text ahead of time, then give every section a different prompt to respond to.

(You can do something similar in problem-set type exams, such as those given in science and engineering, by giving everyone the problem scenarios ahead of time and then asking each section to solve something different about the scenario.)

Depending on the exam type, it may or may not be possible to design an exam of this type. However, if it is possible, it certainly reduces the likelihood that students will gain an advantage from hearing about an earlier section's exam.

share|improve this answer

I have faced with all of the above situations this semester alone! I teach two different sections spread over all 5 days of the week. How do I ensure heterogeneity of examinations? By administering different versions (different questions) of the examination.

Throw into this mix, my need to "accommodate" students with "needs" and different time frames that they require. The only way I managed this was have different versions of the exam and had to invigilate these examinations myself (no help from uni) at different times of the week and well into 6pm on a Friday. I ended up having 5 different versions with varying number if questions but potentially the same difficulty level.

Yes, some students cribbed at the end of the day about them not solving the same problem set as their peers but I had to (politely) ask them to deal with it. If they are thrust into a real world situation, outside the safe confines of a university setting, they may have to face non-homogeneous conditions on a regular basis with difficult time lines.

I self-tested these different versions and all of them took the same amount of time withing +/- 5 minutes for me. Yes, perhaps not the best metric but fair enough, I supposed. However, a lot of teaching may be subjective. You can try to use a grading rubric that you publish after the examination so that the students feel that it was a fair metric/yard-stick.

This was for a first year engineering class on Engineering Mechanics.

share|improve this answer
>If they are thrust into a real world situation... The University IS a real world situation and I wish more students would realize this. Unfortunately, it just so happens that Uni is typically the first time a student enters the 'real world' so the transition is difficult. – LordStryker Mar 9 '14 at 19:00
@LordStryker Well my uni is not quite real world. We seem to coddle students and give them second chances for some egregious offences. – drN Mar 9 '14 at 22:24

A few of my math profs really REALLY didn't want their questions being posted online as they reused them. Their strategy to make sure no one left with a test was usually a variation of the following.

1) When the student enters the exam room, he or she goes up to the prof at the front and turns in everything to the prof. The students name is marked off the list of students. Book bags are left at the front of the room where everyone can see.

2) The prof or TA hands the student the exam along with paper, pencil, calculator, etc. The student takes a seat and begins the exam. Each student receives the exact same thing, and is not allowed to bring outside paper, notes, calculators.

3) Leaving the room constitutes finishing the exam (i.e. no bathroom breaks). When the student is ready to leave the room he or she turns in ALL materials handed out at the beginning of the exam.

share|improve this answer

Disclaimer: I have no idea what exams for writing courses even look like, so this is a general answer.

Students will talk to each other and thus information about the contents of the first exam is going to leak, no matter what you do. Therefore your goal should be that the information about the content of the first exam does not give any advantages on the second exam (keeping both exams at the same difficulty).

There are basically two ways to achieve this (and you do not need to consistently follow one throughout the exam):

  • Make information that could be obtained from the first exam available to everybody beforehand, e.g., clearly state that there will be a question of a given type.
  • Make information that could be obtained from the first exam useless in the second exam, e.g., by not asking a similar question.

The main difficulty is now to recognise useful information. One pedantic example to illustrate this:

Suppose, there is a topic which is not central to your course and which is considered difficult or unfit for exams by your students to the extent that students have a good reason to assume that it is not relevant for the exam. If you (do not) ask a question about this topic in the first exam, students taking the second exam have a better estimate of the probability that you ask questions about this subject and thus have an advantage if you (do not) ask a question about it in the second exam – even if it is entirely different. So, if there is such a subject, you either have to announce beforehand that there will be a question about it in each (none) of the exams or you have to ask about it only in one of the exams. The latter case is however problematic considering the difficulty of the exams, unless you have two such topics at hand.

Some further aspects to consider:

  • Another not so obvious information contained in the first exam is the general way you ask questions and what general aspects you consider important. Giving a sample test is a good way to provide this information beforehand.
  • Neither somebody who focusses on the questions of the first exam nor somebody who focusses on everything that was not asked in the first exam, should have an advantage. Thus some deviation from the rule of not repeating anything without announcing it is ok in my opinion.
  • No matter what you do, somebody will complain about it.
  • In some cases, you can compile a huge catalogue of potential exam questions beforehand (part of this can be the exercises to your course) and give them to your students sufficiently long before both exams. Announce that you will pick questions randomly from this catalogue for each exam (under the constraint of equally difficult exams) and do so. This is very fair, but a lot of work and if your catalogue is too small (which is unavoidable for some subjects), it encourages learning without understanding.
share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.