Seen Exams

This past semester I got to try out using a seen exam for the first time.

For those of you unfamiliar with it, you publish the exam paper some time ahead of the sitting date (a week, in this case), so students can prepare their responses, which they then write under controlled exam controls (without notes or materials to hand).

The logic of this is that it provides a more meaningful test of students’ abilities, since they since have to revise, plan and produce, but without the added peril of “I can’t find a question I can do” or “I answered the question wrong”.

Having inherited the format from a colleague, I was keen to try it out, especially since last year’s use of an open-book, online exam had worked very well. Indeed, this year’s module was with the same students.

The practicalities are very simple indeed: an email to the class and a posting on the VLE at the appropriate time, plus being available through the week to answer any queries or clarifications.

The day before the exam I emailed everyone again, just to run through any points that had come up and to remind them again that the format meant some things were different from a ‘normal’ exam.

Firstly, my expectations on factual accuracy would be higher, since they’d have had time to prepare.

Secondly, I’d like to see more references to the literature: not direct quotes, but certainly mention of relevant authors.

And most importantly, I’d expect clear organisation and argument in each of their answers.

So?

Having now finished my marking, I’m able to say a bit about how this all played out.

As with the other format, this approach seems to be good for pulling up the tail of students who might otherwise have found things difficult: even the worse-performing student still produced relevant answers with some detail.

Likewise, the almost total absence of factual errors and of very short answers was a pleasant development, suggesting everyone had actually done work for the exam.

So the knowledge front seems to be positive.

Having seen a few students straight after the exam, I’m not sure that they found it any less stressful though: yes, they knew what the questions would be, but they also noted that they were also conscious I would be marked in line with that, so maybe their extra work wouldn’t count for anything.

While we’ve yet to complete all the feedback cycle, I think that anxiety is understandable, but hasn’t played out. Instead, the performance of the class has been strengthened and their capacity in the subject will be that bit more for future modules they take.

In sum, this exam has further convinced me that closed-book, unseen exams aren’t that useful, either in measuring knowledge or managing student stress: unless I have to use them in future, I’m not going to be.