TouchPaper Problem #7

Last week I was fortunate to attend the first TouchPaper Party. I was in group 7 and we had the following question to unpick:

TouchPaper Problem #7: What is the optimal number of times for a student to (a) read, (b) hear, or (c) say information aloud if they are to retain for 1, 3, & 6 month intervals?

After much discussion, we broke the problem in two. One group looked at what we need to do in the first place to ensure students learn new information and the other group – the one I was in- looked at how we can ensure it stays remembered.

A century of memory research tells us that students forget. Even when the teaching is good and we can say that students have learnt the material, unless we design opportunities for that information to be recalled, it is going to slowly disappear. For information to be retained at the specific intervals identified in the problem (and beyond) we need to revisit the content.

Our discussion led us to 3 main ideas: Spacing, Interleaving and Retrieval and I’ll look at each of them in turn.


Carpenter et al (2012) explain that “… performance on final tests of learning is improved if multiple study sessions are separated—i.e., ‘spaced’ apart— in time rather than massed in immediate succession” and “Studying information across two or more sessions that are separated (i.e., spaced apart or distributed) in time often produces better learning than spending the same amount of time studying the material in a single session.”

This is all very well, but determining the optimal spacing period is difficult. In one study of 1350 individuals and their recall of facts, Cepeda et al (2008) came up with an optimal gap of 1 day for a test in 7 days, 11 days for a test in 35 days, 21 for 70 and 21 for 350 days. The optimal gap being longer depending on the test delay. The results were slightly different for tests of recognition e.g. supplying multiple choice answers. Also, the results would clearly be different if we revisited the content multiple times.

This does seem to go a little way to answering one element of our Touchpaper question and is a helpful reference point even though the optimal gap for a given set of information in a given time frame is very difficult to be precise about. There is simply no formula suggesting that x sessions at gaps of y% of the test interval is the best way to remember.


Suppose you have X, Y and Z to study. Typically, the study of these would be X, X, X, Y, Y, Y, Z, Z, Z. This is ‘blocking’ and is how our curriculum is traditionally laid out. Each topic is mastered before moving on. Interleaving the topics would tend to look like: X,Y, Z, X, Y, Z, X, Y, Z.

Richland, Bjork, Finley & Linn (2005) show us that interleaving is more effective in the long term than studying something in a block. Bjork calls this a ‘desirable difficulty’ because it would seem counter intuitive and is harder to get right. Indeed, students may perform less well in the short term but retain the information over the long term. Even after students have benefitted, they still often feel as if they have not.

Rohrer (2012) suggests that the greatest benefit is in terms of discrimination. For example, a maths teacher who interleaves methods will find that students are more able to choose the correct method for completing a given problem. There is some of the effect here attributed to the effects of ‘spacing’ material but the effect exists still after taking this into account.

Our TouchPaper problem is about learning one set of information but teaching involves students remembering many pieces of information. Spacing is an important concept but our curriculum isn’t often designed in such a way as to make spacing manageable. Interleaving helps us to space content (thus the effects of spacing) plus we also get these additional benefits of interleaving. We could expect to see the benefits of interleaving at the ‘6 months’ interval of the TouchPaper question and further into the future.


Revisiting content isn’t just as straightforward as rereading or restudy. One important finding from our research was that testing is a way of improving learning, not just measuring it. So when we are talking about spacing intervals, one thing we should be doing at these intervals is testing!

This seems to fly in the face of common sense. Ask students and teachers and they would likely agree that studying is better than testing. Students certainly have the perception that testing isn’t as good as study. Simply rereading the material can create a sense of familiarity which increases the perception that this is better and that material is learnt. For teachers, we talk about how weighing the pig does not make it fatter. It is thought of as the ending point, not the means.

However, there are benefits associated with testing, most notably in retrieval. Roediger & Karpicke (2006) write: “Although restudying the passages exposed students to the entire set of information, testing permitted practice of the skill required on future tests and hence enhanced performance after a delay. If students retain information in their long term memory, they need to be able to access it, and testing may help to develop the cues and ‘retrieval routes’ to stored information.”

They conclude: “Frequent testing leads students to space their study efforts, permits them and their instructors to assess their knowledge on an ongoing basis, and—most important for present purposes—serves as a powerful mnemonic aid for future retention.”

The original TouchPaper question addresses reading, seeing and saying but perhaps we could add in ‘retrieving’ to that list?

Having laid out these ideas, and having barely scratched the surface of the original question, I will propose some points which now need addressing:

    • Teachers are not always curriculum designers and designing a course to cater for optimal learning gaps is difficult.
    • Much of the research base referred to above is on small-scale lists of facts and does this necessarily apply to larger bodies of knowledge and the schemata that underpin our subjects?
    • Does interleaving work better for some subjects than others?
    • What form should an interim test take to maximise retrieval?
    • How do we decide on the optimal spacing gap?
    • A classroom context is different to a research context. Can these research findings be transferred into classroom practice?

On the day, we concluded with the following simple recommendations:

Distribute study of content across multiple sessions rather than a massed session.

Make spacing manageable by weaving numerous topics together throughout the year rather than blocking topics discretely.

Quiz pupils regularly on previous content as the most effective way of retaining it.

A final point. The conversations, discussions and even disagreements during the course of the TouchPaper party helped shape ideas better than any CPD I’ve been involved with. Presentations from other groups contained many insights that have inspired a lot of thinking this week. It felt great to be a part of this occasion and I look forward to reading more blogs and seeing how things develop.

Some useful reading:

Using Spacing to Enhance Diverse Forms of Learning

Spacing Effects in Learning

Why Interleaving Enhances Inductive Learning

Interleaving Helps Students Distinguish Among Similar Concepts

Active Retrieval Promotes Meaningful Learning

Taking Memory Tests Improves Long Term Retention


Removing the cues

Stabilisers helped me to learn how to ride a bike. But until I took the stabilisers off, I couldn’t say that I was able to ride a bike. I have been considering the role of classroom ‘stabilisers’ and how in some cases we might be keeping them on for too long. These stabilisers- or cues- take different forms, from displays to thesauruses and the way that we inflect our voice when we say something like, ‘Is that a simile or a metaphor?’ With too much of these, students become very good at recognising, but not so good at recalling.

We can think of students’ knowledge as appearing on a continuum as follows:

subconscious understanding > familiarity > recognition > recall with a cue > recall.

(paraphrased from Memory: A Very Short Introduction)

Our job is to shift students’ knowledge along that continuum. If they cannot get to the ‘recall’ part, then do they know it at all? Much of the time we keep students at the ‘cues’ stage, where they can perform pretty well in a lesson but without actually being able to do the thing that we want them to do or remember the thing we want them to remember in the long term.

With too many cues, students are just using ‘maintenance rehearsal’ i.e. just learning something enough to get through a lesson or a task. With this, there is a danger that nothing stays in long term memory as the deep thinking required for this to happen is not present. Our ultimate job as teachers is to encode the knowledge, ensure it is stored, then put students in a position where they can retrieve the knowledge without needing our cues.

Interestingly, the act of retrieval itself helps to make things stick and allows students to create their own cues: “Each act of retrieval alters the diagnostic value of retrieval cues and improves one’s ability to retrieve knowledge again in the future.” (Karpicke, 2012)

Bearing all this in mind, I have been considering some of my current practice and implications for change:


feb2013 065This is one of my displays, showing a range of connectives. It is helpful to refer to but now I have a concern that the display doesn’t help students in the way I want it to. Students are never asked to recall the connectives if they are in plain sight. Recently, I have been removing connectives and asking students to recall the missing one. I must confess that I cannot make up my mind whether this display is a help or a hindrance. Advice is welcome.


When we provide feedback we often provide prompts or directions for next steps. We could design even better feedback which requires a follow up activity some time afterwards where students are asked to recall. For example, we might give students an exercise on homophones which explains the rules and asks students to identify (i.e. recognise) which homophones have been used incorrectly in a range of sentences. One week later, we might ask them to write their own sentences or explain the rules of homophones themselves- asking them to recall. This is not something I currently do. I ask students to reflect on whether they meet their target but not by asking them to recall. I’m going to experiment with this form of follow up retrieval homework over the next term.


For students to be able to use a range of vocabulary expertly, we need to shift it from that zone of recognition- they understand it when they read it- to it being part of their regularly used spoken and written vocabulary. Spending too much time searching for the right word in a thesaurus will not have that effect. Displays with key words will be helpful to a point but with no guarantee students will recall them in the future. We know that students need multiple exposures to words in different contexts before they will be readily recalled and used. See this post for some further thoughts on vocabulary instruction.


I think literacy mats or similar can play a part in the classroom. I can see that they might be beneficial for students writing at length in other subjects but I have mixed feelings about them in English. Much like the connectives on the display above, by giving students a mat with, for example, all the punctuation they need to use for a certain level, are we in fact hindering them over the long term? The mats need to be taken away at some point and I would argue that they should be taken away completely. The cognitive load produced by a learning mat can be overwhelming too. Students spend so much time looking at a number of things that they have to include that they lose sight of what they have to do.

I will be using ‘mentor mats’ this half term. These are still a form of cue but consist of mentor texts for study with some elements highlighted for closer inspection. The cognitive load is reduced. There are still cues but these are the isolated areas for study in the unit, rather than a checklist of things to include. Because these cues are in the form of model sentences, they require more thinking than ‘You should include x’. Here is the text from a mentor mat to get the idea.

Curriculum Design

We should design curricula with the opportunities for students to encounter the material they need over and over again, removing the cues and increasing the amount to be recalled. This means that we might teach a concept for the first time and provide all of the supports and then revisit it later without them. For example, in a year 7 scheme of work I have just designed, students will write about The Kraken. Tennyson uses personification in the poem and I want students to be able to comment on it. Therefore, several weeks before, at the start of the scheme, students are taught about personification, encounter the technique later in a short story, revisit it in another poem, The Second Coming, and then read The Kraken. At this point, we hope students can recall the technique unprompted.

This is not a post arguing that cues and scaffolds and supports are not necessary- they play an important role in the learning process. Sometimes it is only repetition of certain cues that force things into long term memory. For example, the C, B, A, mnemonic was important for me to recall the pedals when learning to drive. This cue was never ‘removed’, it just became unnecessary after a point.

If we wish to isolate a skill for practice, we may also provide support on other aspects. For example, when I am looking at sentence structure, I may well offer model sentences and example sentence starts. When I want students who struggle with extended writing to comment on a text, I will often provide scaffolds for support. However, this is always with an end in mind, leading students towards the place where they will be able to do this without the prompts.

Further reading:

David Didau: Deliberately difficult – why it’s better to make learning harder – essential reading on ‘desirable difficulties’.

David Fawcett: Can I be that little better at…using cognitive science/psychology/neurology to plan learning – a must read exploration of key principles of cognitive science.

Stephen Cavadino: Why calculators should be banned (and for balance- a counter argument)