If a child has got a question wrong, we want them to improve their knowledge until they will get that sort of question right. Great. We can do reteaching and intervention questions in purple. But what if they get all the questions right? You can’t be any righter than right, if the answer’s 12V then you can’t make it more 12Voltish. Extension work, of course, is the answer we are given. But I’d like to question the value of extension work too.I’d like to make it clear here that I oppose differentiation as a forward-looking element of planning. I never ever want to plan an activity for Little Sammy to do because he’s bright but Little Katy won’t get on to that because she’s slow. But we do need something valuable for pupils to do when they’ve got everything right – when they’ve “mastered” the area, while others in the group are still practicing to master that area. We might hope that “extension work” can meet this need.

But let’s look at what extension work can mean in science. It might mean harder questions on the same topic. But this isn’t as straightforward as it sounds. Science does follow the difficulty model but that doesn’t mean there is an infinite hierarchy of difficulty in the questions we can set for each topic.

What determines the difficulty of questions is actually really interesting. Some questions are more difficult than others because they deal with more abstract concepts. For example, internal energy requires thinking about several things we can’t see, such as particles, work, kinetic energy, and potential. This is much more abstract than How geothermal energy works, and therefore harder.

Another reason a question can be difficult is if it is on a topic with high levels of element interactivity. This is when there are lots of factors that affect the solution and must be considered together. An example is momentum, where the mass and velocity of two objects must be considered along with an equation and a law. Contrast this with efficiency, where we just need to consider one short equation.

But if we are looking for extension work on a particular topic, questions from a different, harder topic just won’t do.


So we can try and make a harder question on the same topic. Some topics in science lend themselves very well to this – some examples are circuits and balancing equations.


We can make questions on these topics more difficult by increasing the element interactivity or the number of stages needed to reach the solution.


But other topics do not lend themselves to many levels of difficulty. We might be able get two or three levels of difficulty in questions on topics like the wave equation or uses of the electromagnetic spectrum, but after that there are no more levels of difficulty available without introducing new content. For these topics I’m not prepared to reserve the hardest questions for extension because they’re not that hard and all pupils need to be able to do them. We can’t just conjure up extension questions all the time because of the nature of scientific knowledge:


Not all knowledge in science has the same potential for increasing difficulty of questions. When we put knowledge at the centre of our curriculum we can see that extension work actually can not meaningfully exist in many cases.


We’ve seen that higher level difficulty tasks are not appropriate to all topics. What about extension work that introduces more knowledge, stuff you don’t need til A-level perhaps? This sounds lovely until we remember that forgetting curve! Little Sammy’s performance might be at 100% for this lesson, but he is unlikely to achieve 100% in his exam because he will have forgotten things.


Wouldn’t it be much better for a pupil who has got all the questions right to spend “purple pen time”, redoing questions from previous lessons for retrieval practice? This will often do more than anything else to improve the pupil’s final grade. Having all the topics, questions and answers together in one book like this makes this practical in the classroom.


I’m not saying we should not teach beyond the test, there are many instances where I think we should. But these instances should be driven by the knowledge, by the fact that a pupil’s schema will be more complete, more satisfying and more memorable because it is augmented by this extra knowledge. Here’s an example: Pupils are not required to explain the cause of the normal force or friction. But pupils often harbor a misconception that static and inert objects do not exert forces. If we teach that electrostatic repulsion between atoms is responsible for these forces, that misconception is easier to overcome. So I teach this extra material, because of the nature of the knowledge itself. We should not introduce extra knowledge because it fits our model of feedback, lesson structure or differentiation. There are better uses of our pupils’ time.