Why do we put so much weight on final exams?

Question

As we know from cognitive psychology that humans are most creative and learn things better when they are relaxed. I can also relate to this. I do much better in complex problem solving at home rather than in a test. Therefore, I have a hard time understanding why semester final exams are always given such importance and a student's knowledge is judged on the basis of these test results.

I think that what I have learned in 3-4 months can not be properly tested in only 2.5 to 3 hours. Why are exams in general are given this much importance when we know that test results do not fully reflect a person's potential?

Answer 1

What's the alternative?

Exams have been criticized since before I can remember, but if someone has come up with something better I'm not aware of it, e.g.

• Oral exams would "work", except those would also be high-pressure examinations, and they are also very time-consuming. Furthermore, unless one records the oral exam, it would also be difficult to resolve grading disputes.
• Homework-based assessment is vulnerable to cheating, since the student can get 3rd-party help with the assignments and it would not be easy to catch. Students can also work in groups, with associated problems (it's quite common for a few people in a group to do most/all the work and the rest just copy them).
• Subjective assessment where the teacher simply assigns a grade to each student is vulnerable to biases (which is part of the reason why grading schemes and/or blinded assessment are common).
• Finally, there's a school of thought that if you are unable to reproduce the learned material, under pressure, you haven't actually learned it.

Answer 2

What matters is (or at least should be) what you know at the end of the course, not in the middle of it. Thus it is legitimate to have some form of end-of-term assessment, and expected that is should have a disproportional weight, although it need not be in the form of a single exam; one might include student projects or term papers in addition to (or in place of) a final exam.

The difficulty is in ascertaining that the work submitted reflects the knowledge of the individual student (or group of students if there is a group evaluation). For term papers there are plagiarism tools; it is possible to test the knowledge acquired during a project by having specialized written or oral questions. However, it is not always possible or desirable to proceed as so.

The easiest way to assess students in a practical manner very often remains a final exam taken under similar conditions by all the candidates. (Of course this depends very much the discipline.)

In my experience, constructing an exam where students can, over the course of 2.5 or 3hrs, answers questions in a manner that fairly reflects their knowledge of the material taught over an entire term is not easy, and not always successful, but a well-designed exam should allow the more knowledgeable students to outperform the less knowledgeable ones, at least on average.

Of course, some students get lucky and some have a bad day, but in a large group the bulk of the students should be coarsely ranked by their level of knowledge of the material covered in class, else the exam is a failure of the instructor and remedial actions might be considered.

Answer 3

Assessment theory is a complex and large subject.

We must ask why we are assessing students. Off the top of my head, I can think of 3 possibilities:

• To give students a target to aim for. An understanding of what it means to have learnt the subject.
• To allow both student and teacher to assess progress and best plan where a students future time would be best focused.
• To certify that a student is safe or capable of performing a particular task (like a driving test, or a professional competence test).
• In order to rank students so that an employer can choose between them.

I think final exams probably do a bad job of the first two. The can do a good job of the third if the test is well constructed to actually be examining what the student will be doing in their professional role.

But I expect that we focus a lot on the final reason (to the great determent of the education and society in general). Exams are a really efficient and clean cut way of ranking people even if not the basis of genuinely useful criteria. The question is, what are they ranking them on the basis of? Almost certainly not the exact knowledge that an employer wants. Personally I think that if an employer wants to rank people then they should set an exam that actaully tests the thing want. But that's not the world we live in. Since an employer almost certainly doesn't genuinely care about your knowledge of medieval France, but just want some way to distinguish people, exams are an easy way to do that.

Also exams are probably the least work for professors of most assessment methods.

Answer 4

There are really two questions here: "what are pros/cons of final exams" and "why are we using them", which have very little to do with each other.

The answer to the second question is it's been traditionally done that way. Another post here rather peculiarly says that oral exams 'would work', when actually in many countries, oral exams are the main or a very common form of assessment. It is unlikely that they work better in those countries, rather, the traditions have been different. The main source of pedagogical expertise for a majority of teachers is their experience as students. There's very little over the course of their subsequent careers that has any chance to change that, and little incentive.

To make a point, consider an easier question: why do we still have 45min+45min chalkboard lectures on standard courses, like mathematics for economists or electomagnetism? They are really hard to defend. A lecture is a rigid format that goes too fast for some students and too slow for others. It is way too long; studies suggest that after 15-20 minutes, most people lose focus. It's not really interactive; 90% of students are too shy to ask questions in a large audience. It's imperfect with lecturer introducing confusing mistakes, talking to the blackboard, writing in too small letters not visible from last row. Why not replace lectures with high-quality videos, of just right length, with superb visualization, sound and text, pedagogically tested (see 3Blue1Brown series as an example)? As these are courses given every year at thousands of universities, such videos can be produced at a fraction of the cost of the current working time wasted on the lectures. Contact teaching resources can then be used in more productive ways.

Yet every September thousands of lecturers walk into classrooms, take the chalk and say "A matrix is..." Why? Because that's the way it's always been.

Coming back to final exams, they may have real merits, or they can be amended with post-hoc rationalizations, but that's beside the point: even if they were rationally undefendable, they would still be widely used.