Why using exam questions in lessons to improve exam technique is probably a bad idea.

I think I’ve taught my students well but they really need to improve their exam technique, so I’m going to solve this by regularly giving them exam questions at the end of every lesson. 

I’m pretty sure we have all done this at some point in our teaching careers. Today’s blog is all about unpicking this logic, challenging it a little and offering a better solution. I’m going to focus on things from a science perspective but some points will be generalisable. 

Let’s start off with the fabled exam technique, what exactly is it? To me, it seems to be a mix of a few things: 

1. The existence of knowledge the students can recall
2. The awareness of when to apply that knowledge (see ‘Applicability conditions’ Reif 2008)
3. The ability to discern the surface features from the deep structures of a question. 
4. A structural understanding of what a good answer in this type of question must contain

Points 1,2 & 3 can be grouped together under the term flexible knowledge. If a student has flexible knowledge in a subject they can do all those things.

How do students gain these skills? Well we teach it to them, either formally or informally. Every comment you make, example you give and task you choose helps transition students from inflexible towards flexible knowledge. This is why great teaching involves clearly planned examples, non-examples and practice questions.

It is a common practice in lessons to improve exam technique by getting them to complete exam questions at the end of a lesson. I think this is often a flawed strategy because of the mismatch between the aims we want to achieve as classroom teachers and the objectives that examiners have when writing questions. (Note: For now we will ignore the implication that the lesson should be a single learning unit and the problems associated with this from a content planning perspective. See chapter 1 here or a blog here, here or here)

As teachers we want to know what particular declarative or procedural knowledge a student has developed within the lesson. We want to garner granular data of students strengths and weaknesses to inform future planning on specifically the things they have been introduced to today or in prior lessons. We also want to increase the students’ knowledge flexibility in the process by giving them appropriate practice.

An exam question is not always suitable for this. In general, science exam questions are designed to broadly assess either a single topic or a variety of connected topics within the specification. An exam question aims to have sections which attribute various marks to the different assessment objectives, according to the specification weightings. On top of all this the maths and ‘how science works’ (HSW) requirements must also feature in a set proportion. This can mean that a question on osmosis might look very different. From year to year it could have a theoretical slant, or be highly focussed on the required practical content. It might assess graphing skills or a percentage increase calculation. 

The imperfect match between exam questions and the way teachers want to use them creates two problems:

1. We do not generate useful data to inform planning. If the question has 8 marks and 3 are not actually content the students have previously been taught, then how much use is this activity? Those three marks could be some maths content, based on a more advanced part of the spec or HSW etc.. I call them wasted marks.
2. We end up teaching this exam technique via the completion of the question in an ad-hoc discovery-approach instead of specifically planning it in. This leads to demotivation as the students feel confusion and a sudden increase in challenge. They wonder “should I know this? Why haven’t I been taught this?”. Less resilient learners can become locked out in this situation.

While at GCSE this problem is maybe a minor irritation, at Alevel it is a huge issue. I vividly remember teaching Biological molecules many years ago frantically looking for an exam question to use as a plenary, only to realise there are literally no exam questions that do not contain a large proportion of synoptic content. Mindlessly I printed a load and just told the students to skip the bits they couldn’t do. What a numpty I was!

So what should I do instead?

If we want to gain informal data about a student’s current performance and improve the flexibility of their knowledge, then we do need questions. We need specifically planned and sequenced questions that are able to demonstrate to the students the applicability conditions and some of the common structures needed in answering certain types of questions. We need to map out the most appropriate times to teach vital HSW skills, teach them explicitly and then find examples for students to specifically practice them from that point onwards. We may use exam questions as a starting point for our own questions, but we should take care to carefully check the students can access the entire question and edit as appropriate. We might even be able to use a single scenario many times with minor adjustments.

This means curriculum resources need to be carefully thought out and pre-planned. The format doesn’t matter but my personal preference is booklets. To me they allow for the crafting of explanations and questions to cover the topic easily. However powerpoint and worksheets can also do this in the right hands.

If we want to teach students the structure required to write successful exam answers then we need to resource careful examples and model the answers, providing necessary scaffolding by example fading. The students then need practice questions.

So you’re telling me you never use exam questions?

Sometimes we might want students to spend time answering exam questions at the end of a topic. We could choose to make this a formal assessment or we could use it just as a way of building confidence and knowledge flexibility, it depends on our context.

We do know that the testing effect is well proven and that giving students exam questions will help that to a certain degree. My point is it is much less effective than alternative strategies that may seem less authentic than using ‘real questions’ are in fact more effective.

So if you currently use exam questions at the end of your lessons please consider the following questions:

1. How carefully have you chosen your questions? 
2. Do they have ‘wasted marks’ in them?
3. Have you taught all the prerequisite skills needed or have they been covered in earlier years?
4. Does student performance in this question inform your planning for the future?
5. Would the students get more out of spending the same time answering shorter, but more granular questions?