This is the 9th in a series of blogs on assessment, which forms part of a larger series of blogs on the importance of starting strong in higher education and how academics can facilitate it.
There are numerous tropes that aptly apply to the current context in higher education: necessity is the mother of invention, through adversity comes innovation, it’s the survival of the fittest, and all that. Our current adversity renders traditional invigilated exams impossible, and certainly requires us to be innovative to solve the dilemma, but instead of simply looking for technology to innovatively recreate what we have always done, maybe it’s time to think differently about how we design examination in the first place.
Exams are summative assessments. They attempt to test a domain of knowledge and be the most equitable means of delivering an inference to stakeholders of what a student understands about that domain. They are certainly not the perfect assessment measure, as Koretz asserts here (conveyed in a blog by Daisy Christodoulou), but because they are standardised, and invigilated, they can and do serve a useful purpose.
Cheating is obviously easier in an online context and potentially renders the results of an exam invalid. Online proctoring companies, currently vigorously rubbing their hands together to the background sounds of ka-ching ka-ching, certainly mitigate some of these possibilities, with levels of virtual invigilation varying between locking screens, to some using webcams to monitor movements whilst being assessed. Timed released exams also help to reduce plagiarism because students have a limited amount of time to source other resources to complete the test, which inevitable self-penalizes them. I discuss this here. But the reality is, despite such measures, there is no way you can completely eliminate willful deceit in online exams.
So, do we cut our losses and become resigned to the fact that cheating is inevitable and that despite employing online proctoring that some will still manage to do the wrong thing? I’m not sure that’s acceptable, so I think it’s worth considering that if we design summative assessment differently, the need for online proctoring may be redundant.
WHAT DO YOU WANT TO TEST IN AN EXAM?
Do you want to see how much a student can recall of the domain, or do you want to test how they can apply this knowledge? If you want to test recall, then proctoring is a necessity, as answers will be mostly identical in all correct student responses. But should that be what an exam tests?
Few would argue that the aspiration of education is to set the students up in the course to be able to now apply their knowledge to new contexts. By designing a sequence of learning that incrementally delivers key content to students through the use of examples that help shape mental models of ‘how to do things’, and by continuously facilitating the retrieval of that knowledge to strengthen the capacity of students’ memory throughout the course (after all, understanding is memory in disguise – Willingham), we would have supported the development of their schema. This development enables students to use what’s contained in the schema to transfer knowledge and solve new problems, potentially in creative ways.
So exams needn’t be of the recall variety. They can test the application of knowledge.
Whilst we can’t expect the application of that knowledge to be too far removed from its present context (see discussion below), a well designed exam, and particularly those that require written expression, would generate answers that would be idiosyncratic, and then could be cross checked with TurniItIn to determine integrity.
In this way, timed exams in certain courses* could effectively be open book, eliminating a large component of the invigilator’s role. This may seem counter-intuitive, but the reality is that even if a student can simply access facts they haven’t committed to memory, they will still unlikely be able to produce a strong answer to a new problem. Their understanding of the content is limited simply because they haven’t spent enough time connecting it to previous knowledge which generates eventual understanding. The students will spend most of their working memory’s capacity trying to solve the problem, and invariably, in a timed exam, self-penalize in the process. It’s like being given all the words of a new language and being asked to speak it in the next minute. It’s impossible.
In order to successfully use the internet – or any other reference tool – you have to know enough about the topics you’re researching to make sense of the information you find.David Didau
4 REQUISITES OF A WELL DESIGNED OPEN EXAM
- Students have relevant schema
- Students have practised applying it to new near contexts
- Exam questions seek near transfer of knowledge
- Exam is timed and made available at a specific time interval – see here
I have just discussed the importance of schema, but if we want students to be able to apply that knowledge to new contexts we have to model and train them in doing so. This may seem obvious, but curricula are usually so crammed that educators often don’t have time to teach the application of knowledge. Or, as an ostensible antidote to such a context, some educators have fallen for the lure of problem based or inquiry learning, where students are thrown into the deep end and expected, without a sufficient schema, to solve complex problems. Such an approach doesn’t result in efficient learning, and often favours those with stronger cultural literacy, thus exacerbating the Matthew Effect. The ideal situation then is to support the development of a substantial schema and then allow space in the curriculum to help students learn how to apply that knowledge… and then test it in an open book exam.
The third requisite is the design of the exam questions. A strong design would have to ensure that the expected transfer of knowledge is not too ‘far’, and in fact is closer to ‘near’ transfer. We often exult in education’s aspiration of being able to transfer knowledge into new contexts, but the actual reality of this may render us less optimistic. The Wason experiments illustrate this well, suggesting that our knowledge is really quite specific, and that what we know about problem solving in one topic is not necessarily transferable to others. If you don’t believe me, try this experiment below, and click on the link above to see the answers.
Lots and lots of very smart people get this task wrong. What the experiment shows us is that it’s not how smart we are in being able to solve problems, but how much practice we’ve had related to the problem. So designing appropriate questions in an exam is crucial if we want the results to provide strong inferences about our students’ learning.
CRITICISMS OF OPEN BOOK EXAMS
A criticism of open book exams is that students are lulled into a false sense of security and fail to study enough for the test, believing the answers will be easily accessible from their notes – the fallacy that you can just look it up in Google, as discussed above. However, because we know that we need most aspects of the domain to be memorised to support the automaticity of its retrieval when engaging in new learning, (cognitive load theory), and have thus incorporated retrieval practice into our teaching, the need for a student to actually have to look up information will be quite low.
EXPOSURE TO OPEN BOOK ASSESSMENT IS CRITICAL
Like any learnt skill, you have to build the knowledge associated with it, and then practice until made perfect. Never expect that knowing how to function in an open book exam is a given skill. It is important to train the students in how to prepare for such an exam, by helping them learn to summarise their notes to reflect key concepts, to organise their notes so they can be easily used in the exam, and how to plan answers before committing them to writing.
A PEDAGOGICAL UPSHOT
As mentioned previously, the need for students to memorise key facts is an essential aspect of the learning journey, but sometimes summative exams tend to focus on this type of knowledge too much, or worse, expect transfer of that knowledge without providing necessary practice in doing so. The upshot of open book exams is that it not only requires students have sufficient knowledge, but also sufficient practice in applying it, and so the open book exam becomes a paragon of good teaching.
*online open book exams may not be so easy in courses like mathematics and equation based courses that require identical solutions.
I’m Paul Moss. I’m a learning designer. Follow me @edmerger