A week working at UCLA in the Bjork Learning and Forgetting Lab


Dear readers,

Over the summer I was privileged to visit UCLA in Los Angeles to work with a number of educational researchers looking into how cognitive science can be used to improve teaching. In a highly stimulating week I learned about both the historical and contemporary research findings relating to our thrusts and was also able to share some of the work we have done at Wyvern College with the academics.

I thought that many of you may be interested in hearing about what I learned during the visit and so have summarised it in this blog post in order that it might provide some ideas and inspiration for your own professional learning this coming year.

I am not an expert in the minutiae and detail of all the following content. At UCLA there are fulltime researchers spending years of their career looking into specific, narrow phenomena. In my visit I tried to capture the main concepts and ideas of many people’s research across the breadth of the department and will summarise them forthwith. To provide you with some direction for further reading if you are interested, I have hyperlinked papers and articles to this post.

I spent a week working in a research group of the UCLA Psychology Department called the Bjork Learning and Forgetting Lab. This has been led by heroes of mine, Professors Robert (Bob) and Elizabeth Bjork since they both got tenure 37 at UCLA 37 years ago. Bob is a Distinguished Research Professor and is considered one of the top two experts in the world on memory and cognition. He has been President of the American Psychological Society and is also a previous winner of that institution’s Distinguished Service to Psychological Science Award and also their Lifetime Achievement Award. Elizabeth is a Fellow of the Association for Psychological Science and is a previous winner of UCLA’s Distinguished Teaching Award.

Performance versus learning

The Bjorks’ research is focussed on understanding the way in which human memory works and the implications for learning and teachers’ classroom practice. I’ll summarise the main points, pulling out their key ideas and laying down some context before then sharing their most important findings in terms of implications for classroom practice. It is important that the learning-enhancing strategies later discussed are seen through the prism of the context that underlies them in order to prevent subsequent misinterpretation or people inferring they are ‘universal silver bullets’; they most certainly are not.

The Bjorks say that any definition of learning needs to include both an element of knowledge needing to be retained and accessible over time and also that the knowledge needs to be transferable to different contexts to that in which it was acquired. They talk of learning as ‘retention and transfer’.

In any observation you make of a classroom they argue that what the students are demonstrating at a particular point in time is performance, not necessarily learning. Their mantra is that performance is often a very poor proxy for learning. If a student gives you a correct answer during a lesson, can you infer from this whether they will be able to give a correct answer on this topic in six months’ time and when the question is worded in a different context? That correct answer today may in fact just be the result of the students’ mimicry of the teacher, the recency with which the student was given the information or other factors that can ‘prop-up’ performance. It is refreshing to know that Ofsted also now understand this point having seen the Bjorks’ and others’ research in this area, facilitated by people such as David Didau, which was a big influence in them deciding no longer grading individual lessons and becoming much more focussed on exam performance data.

However, rather than interpreting this notion as entirely futile, it does have some strong positive implications for the way in which we conduct summative assessment. Alan, my Headteacher speaks regularly about the need for college assessment to be accurate and reliable. If mimicry and recency can desync performance from learning, by ensuring summative assessments of learning are time-delayed and contextually-varied (to the way in which the knowledge was taught) we can be more confident in what students have retained and transferred (learned) from what we taught them. If students can apply the knowledge from today’s lesson on an assessment independently in a few weeks from now in a context that is different to the lesson then we can be reasonably confident we are measuring learning and not performance.

We experimented with this approach to assessment in maths last year by doing time-delayed end-of-unit assessments with Year 9 students. It was sobering to see the significant difference between what students could do in their exercise books during their lessons versus that they then couldn’t do on a delayed and contextually varied end-of-unit assessment. Formative assessment is an excellent in-lesson tool to manage the pace of learning, but is often near-useless for measuring retention and transfer– learning. Time-delayed, contextually-varied summative assessment certainly has its place if you want to accurately track learning…

In this video Bob discusses these ideas in more detail: dissociating learning from performance (GoCognitive, 2008a).

Memory– if you measure it you change it

A common misconception people have about memory is that it ‘operates like a video recorder’, i.e. information is somehow ‘recorded’ in its entirety and in the format in which it was input. The Bjorks’ work has shown this is a very poor analogy.

Memories typically become less accessible over time and the drop-off is significant. If you are to retain access to new knowledge over time periods of months and years you need to retrieve it a number of times after it was first learned. The gaps between retrievals need to be of the order of days and weeks to be effective. If this is done, by retrieving previously learned information, what the Bjorks call a retrieval event, you are in fact boosting its accessibility in future retrieval events. Perhaps a metaphor of a seed ben of plants and only regularly watering only the ones you want to grow would be a better analogy than that of a video recorder?

For more details, see this video: using our memory shapes our memory (GoCognitive, 2008c).

The New Theory of Disuse

The Bjorks are perhaps most well-known for their theory of memory called The New Theory of Disuse. Since its publication it has been rigorously tested with empirical studies and is widely accepted amongst cognitive scientists as capturing important behaviours. I will summarise The New Theory of Disuse as concisely as I can for your interest below. If you are interested, you can read the original article here (Bjork and Bjork, 1992) or a more detailed summary I have written previously here (Emeny, 2015). Bob also explains it in this video: the theory of disuse and the role of forgetting in human memory (GoCognitive, 2008b).

The theory says that everything committed to memory, each memory representation, has two strengths, a retrieval strength and a storage strength. The retrieval strength corresponds to how easy it is to access the memory representation; how easy it is for you to retrieve it into your working memory. Memories with high retrieval strength are highly accessible. Storage strength relates to how deeply you have learned the memory representation, in particular how linked and connected it is to other memory representations. Memories with high storage strength are deeply learned and conceptually well understood.

The theory specifically uses the word disuse in its title as the Bjorks’ research has shown that memories are never completely forgotten, you just loose access to them over time. In the context of their theory, they have shown that gains in storage strength are entirely cumulative; you never lose storage strength. So when you learn something, over time you will lose access to it (a loss of retrieval strength) but you won’t lose the depth of understanding for which you learned it.

Consequently, you can end up with memory representations that have high storage strength and low retrieval strength. For example, if I asked you to name everyone in your primary school class you would struggle. If instead I showed you a list of names and asked you to pick out the ones that were in your primary school class, you could probably get them all. This high storage strength memory is still retained by you, but its accessibility (retrieval strength) was low through disuse. The cued recall would be sufficient however to increase the retrieval strength.

There is a relatively more challenging part of the theory to understand which relates to the gains in retrieval and storage strength each time you retrieve something from your memory. Please do not worry if you do not fully follow this; the main messages to take from it are:

  1. You need to retrieve learning a number of times with a spacing gap to build storage strength
  2. Gains in storage strength (depth of understanding) are much greater if done so after you have had time to forget your previous learning (i.e. done from low retrieval strength). Conversely, if you try to get gains in conceptual understanding by retrieving something from the point when it is already ‘automatic’, the gains are minimal; you do it ‘without thinking’ and because you don’t think, you don’t link it to other things in your memory. As Bjork puts it, ‘teachers need to realise forgetting is the friend of learning’.
  3. The implication of these ideas (which have been verified experimentally) is that if you want ‘deep learning’ (high storage strength), you need to build it over time through revisiting topics a number of times when you find them difficult, but not impossible, to retrieve from your memory. Put bluntly, sustained learning is not a rapid gain, but instead requires a number of revisits over time. Mid-term planning to revisit concepts a number of times is essential if you want your students’ learning to be sustained. Telling them once and expecting them to remember it is futile.

For completeness, I’ll describe how these observations relate to the theory. Please feel free to skim over the next two paragraphs if you are happy just to know the findings and are less concerned about their theoretical foundations.

The Bjorks showed that each time you retrieve a memory representation into your working memory you it gets a gain in both retrieval and storage strength. The growth in each of the strengths depends upon the current relative strengths. They found that gains in retrieval strength are positive correlated to the current storage strength. Put simply, the deeper you have learned something, the faster you gain access to it when you have lost access to it. ‘You never forget to ride a bike’, is the cliché. You lose access to it over time, but it comes back quickly if you spent a lot of your childhood riding your bike.

They then showed that gains in storage strength are negatively correlated with current retrieval strength. In simple terms, you get a greater gain in the depth of learning if you retrieve something after you have lost some access to it. Effortful retrievals are great for new learning and making connections to previous learning. Retrieving things that are already automatic see minimal gains in learning (storage strength).

The Bjorks realised that to get the greatest gains in learning (storage strength), teachers need to introduce techniques in their instruction that limit gains in retrieval strength. By keeping this low, the gains in storage strength, making connections and links, the gains in learning are maximised. They called these teaching techniques desirable difficulties. There has been considerable research effort put into understanding desirable difficulties in the ensuing years and they are still very much the focus of contemporary research in the Bjork Learning and Forgetting Lab and in other research institutions around the world.

Before I discuss desirable difficulties and share with you both the historical and contemporary findings, it is important to emphasise the desirable part of desirable difficulties. These techniques make learning seem more effortful and challenging, certainly at the superficial level. The gains are long-term and in fact lead to lower performance during the early stages of learning. Their message is certainly not ‘make learning more difficult’, the increased challenge and difficulty must be desirable ones– specifically ones that potentiate learning. The Bjorks have shown that students often prefer to take the easier path than engage in desirable techniques that make their learning seem more challenging; their intuition for what maximises their learning is often very wrong. Desirable difficulties seem counter-intuitive for many people, teachers and students alike.


Bob and Elizabeth Bjork in their lab at UCLA that has been conducting cognitive science research for 37 years.

Desirable difficulties

The following are the main desirable difficulties that have a strong research base and are currently or previously the focus of study in the Bjork Learning and Forgetting Lab.


Building storage strength (learning) is a cumulative process that takes time. Each time your retrieve something you get a gain in its retrieval strength. However, the more times you retrieve it, the rate at which the retrieval strength falls (the speed at which you lose access to it) decreases. The more times you retrieve previous learning, the longer it will be retained for the future. The result of this is that spacing students’ practise activities leads to greater retention of their learning over time. If they are going to complete 20 questions on a topic their retention in the long run will be better if they do 10 today and 10 in a fortnight than if they do all 20 today.

The spacing effect has been experimentally verified many times and with a broad research base stretching back over the last century. Doug Rohrer, a Professor at the University of South Florida has published many papers on the spacing effect. He has summarised much of the main ideas and research base in a literature review, Student Instruction Should Be Distributed Over Long Time Periods, (Rohrer 2015).

In an impressive study, Spacing effects in learning: A temporal ridgeline of optimum retention, Cepeda et al (2008) explored the effect of varying both the time intervals between each time some learning was retrieved (the spacing interval) and between the last time some learning was retrieved and the test (the testing interval). They found there is an optimum spacing interval which, if you are aiming for long-term retention (a long test interval) is of the order of 3 weeks. In this study which involved over 1350 subjects (!) they were still only able to look at retrieving learning on two occasions. We know in practice that maximising storage strength is likely to take more than two retrievals and so current research in the lab is looking into whether fixed 3-week spacing intervals is optimum if more retrievals are done. Recent findings suggest that expanding spacing intervals are even more effective, particularly if informed by formative assessment (based on students’ response times) (Mettler, Massey and Kellman, 2016).

Whilst the retention benefits of spacing have been shown many times in the literature, these are mostly laboratory studies typically performed on undergraduate students tested on semantic, list-type information. Currently, there is little more than testimonial evidence for impact of spaced learning in secondary school classrooms. A number of colleagues in the maths department and myself have been experimenting with spaced learning the last couple of years and we believe we it can certainly be applied and benefit learning in secondary maths. I’m personally in the process of working with Bob to publish the impact study for Numeracy Ninjas which incorporated spacing principles and had a strong impact on students’ numeracy. In the coming year I’m going to do some collaboration studies with Bob and his team hopefully to demonstrate strategies for successfully applying spacing in secondary maths classrooms.

A number of maths colleagues have tried various approaches to spaced learning. I will include a variety here in case they are of interest to you and something you might like to adapt and make the subject of your own professional development work in the future:

  1. Spaced starters. Beginning lessons with a small number of question on previously taught topics. Repeat the topics of the starter questions until students are getting them right over a period of days, then move onto different topics
  2. Lagged homeworks. Make homework topics based on what was taught a few weeks ago
  3. Staggered revision schedule. Giving students a revision schedule where each topic gets revisited at least 3 times by staggering the tasks. See an extract below of the staggered revision schedule many of our class of 2016 cohort followed:


In addition to teachers incorporating spaced learning ideas into their practice, Bob is also keen for students to understand these principles and to realise that by incorporating them into their own independent study that they can get significant learning gains without an increase in study time. At present, I am not aware of any good case studies of schools who have got this working in practice. This may be something you would like to look into?


Interleaving is the second main desirable difficulty for which there is a significant research base to support. In principle, interleaving is simply giving students questions on a variety of topics, rather than ‘blocked practise’ on single topics. Put simplistically:

Blocked practise Interleaved practise
Study A

Practice AAAAAA

Study B

Practice BBBBBB

Study C

Practice CCCCCC

Study A

Study B

Study C


There are different competing theories about why interleaving leads to greater retention and context-transfer benefits over time. Some researchers believe it is because students have to learn to select the strategy in addition to executing the strategy. In blocked practise they effectively have already been given the strategy. Others believe interleaving somehow captures ‘higher-level processing and linking’ than blocked practise.

An important thing to understand about using interleaving in practice, which is typical of many of the desirable difficulties, is that students will find doing interleaved practise much more challenging and difficult than blocked practise. Their performance, quantity of work produced and accuracy in lessons will be lower. However, the gains in the longer term can be highly significant. In a 2007 paper, The shuffling of mathematics problems improves learning, Rohrer and Taylor (2007) showed that students who undertake interleaved practise in maths lessons perform lower in end-of-lesson tests, but that they outperform students who did blocked practise on a delayed test 1 week later. The effect was highly significant; the interleaved group performed with 3 times greater accuracy on the delayed test!


The mixers (interleaved group) performed lower in end-of-lesson assessments (practice performance), but out-performed the blockers group by 3:1 on a 1-week delayed test (test performance)

These findings, which have been replicated, raise profound questions about pitch and challenge in lessons, whether work sampling, lesson observation etc can accurately judge and quantify students’ learning or whether we should focus ourselves more on outcomes and progress over time, as Ofsted now do. Interleaving and many of the other desirable difficulties that lead to better outcomes in the long run will see lower ‘in-lesson performance’ in the short term as students work on activities which are ‘desirably more difficult’ than they would traditionally have worked on.

We incorporated spacing and interleaved principles into our Numeracy Ninjas programme and ran a year-long study of its impact with 470+ Year 7 and 8 students in 2015-16. We found that it increased the average students’ retention of mental numeracy strategies and important KS3 topics by at least an additional 70% when compared with a control cohort who did not have Numeracy Ninjas the previous year.

The testing effect

Testing can be used for many reasons. Traditionally we think of it in terms of assessing students’ level of attainment and tracking their progress. Around the turn of the millennium, Dylan William (1998), strongly advocated using assessment formatively, not just to assess learning that had happened, but to then use it to guide and design future instruction through Assessment For Learning strategies. The community of researchers associated with desirable difficulties have highlighted a different way in which we can use testing, one that is a learning event in itself.

Before we go any further, I should clarify that when I say ‘testing’, I really mean a ‘retrieval event’. In what follows think of testing within a ‘low-stakes’ format, perhaps as a ‘quiz’ rather than a ‘test’. A test is one way of getting students to retrieve previously learned information from their brain (rather than giving it to them); there are others.

The desirable difficulty community of researchers want teachers and students to understand that testing, i.e. low-stake quizzing, cannot only be used for assessment purposes (both summative and formative), but as a learning event itself. The process of students having to retrieve the information during the test is a highly-effective memory-modifier and enhances their learning.

In a study, Test-enhanced learning: Taking memory tests improves long-term retention, Roediger and Karpicke (2006) gave students prose passages from “The Sun” and “Sea Otters” to learn. However, they instructed students to follow different study schedules:


True to the typical desirable difficulties format, the students that undertook the STTT schedule performed lower in a test 5 mins after studying, but higher in a delayed 1-week test.


One of the most remarkable findings in this study was that the STTT schedule was even more effective than SSSS even though on no feedback was given on the tests during the learning period. In the 3 ‘tests’ (T) that the STTT group took during the learning period students answers were not marked.

Bob and Elizabeth talks about this study showing the significant benefits of students undertaking learning activities where they have to repeatedly retrieve previously learned information. To clarify, this isn’t talking about the misconception many people have that ‘if they discover it for themselves they’ll learn it better’; there is no research foundation for this. The point is that to build retention of previously learning, students will get better results if they undertake learning activities, such as low-stakes tests, which require them to retrieve the information themselves rather than simply re-reading it or passively receiving it from the teacher.

The message for teachers, as the Bjorks put it, ‘input less and get students outputting more’. Using frequent low-stakes quizzing of previously taught topics is very effective at building retention, certainly in comparison with techniques such as re-reading and copying notes. The old cliché of getting students to teach their parents/peers is another effective strategy that captures this message, but only if they generate the lesson themselves by retrieving previously learned information. In the ‘Brain, Book, Buddy, Boss’ strategy that we give students in maths to follow when they are stuck, it is very important they have engaged with the ‘Brain’ strategy first before turning to the others!

The Bjorks are also keen on students understanding the impact the testing effect should have on their own study habits. They believe most students engage in revision techniques that are not retrieval-based, such as highlighting, re-reading and re-copying notes. One effective strategy for students to know from the testing effect is that a very effective way to study is to generate their own quizzes based on previous lessons and to test themselves on these at regular intervals until their performance on them is high after a long delay. They should also ensure they can answer questions when they jump between quizzes (interleaving).

Alan talks often about getting students following ‘active learning strategies’ for their revision. Within this context, the desirable difficulties research suggests we these should include strategies incorporating high-levels of retrieval.

The pre-testing effect

Giving students a test or challenging activity on the topic of a lesson they have not yet been taught can enhance their learning of the content in that subsequent lesson. Even though students will perform very poorly on the pre-test, their learning in the lesson that follows can be greater than it would otherwise have been.

In one maths-based study, Designing for productive failure, Kapur and Bielaczyc (2012) showed that pre-testing students, a process they called ‘productive failure’, led to greater learning in the subsequent direct instruction lessons as evidenced by delayed test performance. Direct instruction was more effective if they were given a pre-test first. They demonstrated the effect across question types that ranged from recall to abstract problem solving and transfer to different contexts. Their findings suggest that fears that some teachers (myself included!) have about students embedding errors if you give them free-reign on a topic before some instruction may be counter-productive to students’ learning. It may be that the links and intuitions generated during the ‘productive failure’ pre-testing then make the subsequent learning from the subsequent direct instruction more effective.


I did not get a chance to speak with the Bjork researchers in depth about this desirable difficulty and so am not clear myself if there are particular caveats and limitations to this strategy or how best to implement it in a secondary school environment. Perhaps this might be something that you would like to look into in your own projects? Would giving students pre-reading to do before lessons enhance their learning?

For further reading on pre-testing studies see: Productive failure in learning the concept of variance (Kapur, 2012), The benefits of generating errors during learning (Potts and Shanks, 2014), Why does guessing incorrectly enhance, rather than impair, retention? (Yan et al, 2014) and The Pretesting Effect: Do Unsuccessful Retrieval Attempts Enhance Learning? (Richland, Kornell and KAO, 2009).


Checking out Santa Monica Pier (end of Route 66) with Bjork Lab researchers Dr Veronica Yan and Dr Courtney Clark. Both have done research work on the benefits of pre-testing and other desirable difficulties

Disfluency- fonts etc

During my visit I got the opportunity to meet with Danny Oppenheimer, Professor of Marking and Psychology, for a brief 45 discussion. Although his research interests are mainly outside the domain of desirable difficulties, he has contributed to this field via an interesting result around the concept of ‘disfluency’.

In a study, Fortune favors the bold (and italicized): Effects of disfluency on educational outcomes, Diemand-Yauman et al (2011) found that giving students learning resources which had fonts that were more difficult to read led to better performance on delayed tests on that information. The disfluency caused by an increased subjective experience of difficultly led to deeper processing of the information. The study demonstrated the effect first under laboratory conditions, but then also replicated it in a secondary school environment.

Before you start changing your worksheets to Curlz MT size 8 font, there have been a number of issues in replicating the original study findings and it would be fair to say how context-dependent the learning conditions need to be for this the effect to enhance learning. More research is needed to understand specific disfluency strategies that teachers can use to get students to process information presented to them more deeply.

For further reading on the concept of using disfluency to increase depth of processing see, Thinking, Fast and Slow (Kahnemann, 2011).

Other research

There was other research projects going on in the lab that show promise in enhancing learning, but at this stage are relatively impractical to put into practice in schools without developing software platforms and particular ICT infrastructure to help their delivery. For this reason, I will not go into depth about them here but am happy to discuss them with you in person if you have a particular interest. In particular, these areas relate to confidence-weighted multiple choice testing and perceptual learning. The paper, Perceptual Learning Modules in Mathematics: Enhancing Students’ Pattern Recognition, Structure Extraction, and Fluency, (Kellman, Massey and Son, 2010) gives you a flavour of the ideas involved in perceptual learning. The impact is quite impressive!

Identity-based motivation theory

One morning during my visit I got to meet Daphna Oyserman, Dean’s Professor of Education and Communication at the University of California. During a walk along the beautiful Manhattan Beach, Daphna explained to me her work into mindsets and how changes in context can change students’ mindsets and subsequent behaviours and outcomes.


Manhattan Beach, LA

Importantly, Daphna explained in detail why Dweck’s growth mindset work is an incomplete theory. Essentially, it’s a good idea, but flawed because it doesn’t incorporate human motivation factors. It is more of an observational theory than a theory of action. Yes, people with growth mindsets get better outcomes, but the theory doesn’t explain the motivations, strategies and so on that tell us why some people have growth mindsets and others have fixed mindsets. It is an observable symptom rather than a root cause and as such, teachers are not informed about what to do with it. Some colleagues and myself were frustrated with exact this point when we tried to do some work on growth mindsets last academic year.

Ultimately, Daphna argues that a mindset theory needs to start by considering human motivation and decision making that then moves onto giving tangible actions and strategies which influence mindset. Without starting at this point, any theory that focusses only on the observable outcomes of mindset, rather than the driving input factors isn’t likely to yield interventions and strategies that can be acted upon. You need to look under the bonnet to see what is wrong with a car…

Daphna has spent a considerable number of years deriving and then rigorously testing her own theory of mindset that starts from a motivational standpoint. Her Identity-Based Motivation theory, (Scott and Kosslyn, 2015) argues that people’s perception of their ‘self’, their identity strongly influences the way in which they react to both easy and difficult tasks and situations. Both their perception of their current identity and their future identity are important drivers in people’s mindsets and consequent actions. What kind of person they think they are now and what kind of person they think they will be in the future strongly governs their mindset. Their mindset is centred around their current and future perceived identities.

I have been thinking about this a lot anecdotally and the more I consider it, the more I think the concept of identity seems important in so many instances surrounding motivational decisions. The concept of branding is about buying a new identity for yourself. In high-performing and motivated sports and other professional teams people often talk about the ‘team spirit’ and people buying into something (an identity) bigger than themselves. The antithesis of this are the ‘lost souls’ as I sometimes think of them, the students without a plan, without a future identity to work towards. Even Dan Pink’s (2010) big three drivers of motivation, autonomy, mastery and purpose seem to be coherent and relevant when you view them through the lens of being components of a strong perceived identity. It seems to encompass my own intuitions and observations about real intrinsic motivation.

Daphna has shown that students’ identities are dynamically constructed, meaning they are highly dependent upon the environmental conditions and the ‘possible future identities’ students have in any given moment. Within the growth mindsets theory this manifested itself in the observation that students could be ‘growthy’ about some parts of their lives and ‘fixy’ about others. She has shown that by changing the perceived possible future identities the current ones that come to mind can be changed.

Next, her work has shown that in order to be motivated to take actions students need to see the relevance between taking the current action and it helping them work towards their future identity. The current task has to feel related to the long-term goal. There are a number of factors that Daphna has shown are important to achieve this:

  1. The action needs to be one that is consistent with important social identities. For example, if their future identity is based on a career as a doctor, the action needs to be consistent with the ethos and values of being a doctor. The flip side of this observation is for students who have a low-aspiration future identity, their current actions will reflect this as they try to fit in with the social identity they are working towards.
  2. Students need to feel that the actions are relevant behavioural strategies that are achievable. If they perceive the action you are asking them to take is not achievable they are unlikely to be motivated to take it.
  3. Students need to feel that the actions will help them achieve their goals. They need to seem relevant. Daphna has even done some experimental manipulations (Lewis and Oyserman, 2015) to make students feel like their future goals are closer than they really are and she saw resulting increased student motivation.

If the above action readiness conditions are met, as Daphna calls them, she argues that students’ interpretation of perceived difficulty will be different. Students who believe the task is relevant and consistent with their future identity will perceive a difficult task as being important and engage with it. Conversely, students who believe the task is not relevant with their future identity will perceive a difficult task as being impossible, ‘not for people like me’ and not engage with it. The defining factor in whether they are motivated to take actions to engage in difficult tasks, which under Dweck’s theory we would call ‘growthy’, is down to whether they believe the actions are relevant in helping them work towards their perceived future identity.

Fortunately, Daphna is as much a pragmatist as much as a theorist. She has developed a complete programme called Pathways to Success (Oyserman, 2015) that can be run over 12 sessions in schools to put her Identity-Based Motivation theory into action to boost students’ academic outcomes. She is currently developing a website that includes all the resources and even videos of her delivering the sessions in US high schools herself. Although not formally published yet, she has very promising impact data from schools right across the socio-economic spectrum.

Potential implications for maths teaching

One way to view desirable difficulties are that they are specific high-impact strategies for incorporating challenge into lessons. This is productive challenge that leads to better retention and transfer of learning over time. These are challenge strategies that are in addition to pitching high from a conceptual standpoint which are effective at consolidating learning for the long term.

Whilst the benefits of desirable difficulties have been replicated many times in laboratory studies, knowledge of how to implement them and maximise their benefits in real classrooms is still relatively in its infancy. In your own research projects and professional development, you may like to explore some of the desirable difficulties ideas.

Desirable difficulties enhance learning in the long term by making it more challenging in the short term. In order to get students to engage with desirable difficulty strategies and embrace the increased challenge you might like to explore how this fits within the Identity-Based Motivation theory ideas. Do we need to explain desirable difficulties to students to get their ‘buy-in’? Should we talk about them wanting to become someone who ‘studies short and smart’, not ‘studies long and poorly’? Perhaps you would like to get involved in delivering the Pathways to Success programme?

I hope this has provided you with some ideas for your own professional development and that it has been of interest. On a personal note, I feel as though I learned so much from the Bjorks and their colleagues who were so welcoming and generous with their time. They are very interested in what teachers do to put their research into action. I plan to collaborate with them myself in the future once I’ve had time to think in more detail about what all this means for maths teaching.


Bjork, R. and Bjork, E. (1992) ‘A New Theory of Disuse and an Old Theory of Stimulus Fluctuation’, in From Learning Processes to Cognitive Processes: Essays in Honor of William K. Estes. Hillsdale, JK: Erlbaum, pp. Vol. 2, pp 35–67.

Cepeda, N.J., Vul, E., Rohrer, D., Wixted, J.T. and Pashler, H. (2008) ‘Spacing effects in learning: A temporal Ridgeline of optimal retention’, Psychological Science, 19(11), pp. 1095–1102. doi: 10.1111/j.1467-9280.2008.02209.x.

Diemand-Yauman, C., Oppenheimer, D.M. and Vaughan, E.B. (2011) ‘Fortune favors the bold (and italicized): Effects of disfluency on educational outcomes’, Cognition, 118(1), pp. 111–115. doi: 10.1016/j.cognition.2010.09.012.

Emeny, W. (2015) Forgetting is necessary for learning, desirable difficulties and the need to dissociate learning and performance. Available at: http://www.greatmathsteachingideas.com/2015/01/20/forgetting-is-necessary-for-learning-desirable-difficulties-and-the-need-to-dissociate-learning-and-performance/ (Accessed: 17 August 2016).

GoCognitive (2008a) Dissociating learning from performance. Available at: http://www.gocognitive.net/interviews/dissociating-learning-performance (Accessed: 16 August 2016).

GoCognitive (2008b) The theory of disuse and the role of forgetting in human memory. Available at: http://gocognitive.net/interviews/theory-disuse-and-role-forgetting-human-memory (Accessed: 17 August 2016).

GoCognitive (2008c) Using our memory shapes our memory. Available at: http://www.gocognitive.net/interviews/using-our-memory-shapes-our-memory (Accessed: 16 August 2016).

Kahneman, D. (2011) Thinking, fast and slow. New York: Farrar, Straus and Giroux.

Kapur, M. (2012) ‘Productive failure in learning the concept of variance’, Instructional Science, 40(4), pp. 651–672. doi: 10.1007/s11251-012-9209-6.

Kapur, M. and Bielaczyc, K. (2012) ‘Designing for productive failure’, Journal of the Learning Sciences, 21(1), pp. 45–83. doi: 10.1080/10508406.2011.591717.

Kellman, P.J., Massey, C.M. and Son, J.Y. (2010) ‘Perceptual learning modules in mathematics: Enhancing students’ pattern recognition, structure extraction, and fluency’, Topics in Cognitive Science, 2(2), pp. 285–305. doi: 10.1111/j.1756-8765.2009.01053.x.

Lewis, N.A. and Oyserman, D. (2015) ‘When does the future begin? Time metrics matter, connecting present and future selves’, Psychological Science, 26(6), pp. 816–825. doi: 10.1177/0956797615572231.

Mettler, E., Massey, C.M. and Kellman, P.J. (2016) ‘A comparison of adaptive and fixed schedules of practice’, Journal of Experimental Psychology: General, 145(7), pp. 897–917. doi: 10.1037/xge0000170.

Oyserman, D. (2015) Pathways to success through identity-based motivation. United States: Oxford University Press.

Pink, D.H. (2010) Drive: The surprising truth about what motivates us. Edinburgh: Canongate Books.

Potts, R. and Shanks, D.R. (2014) ‘The benefit of generating errors during learning’, Journal of Experimental Psychology: General, 143(2), pp. 644–667. doi: 10.1037/a0033194.

Richland, L.E., Kornell, N. and Kao, L.S. (2009) ‘The pretesting effect: Do unsuccessful retrieval attempts enhance learning?’, Journal of Experimental Psychology: Applied, 15(3), pp. 243–257. doi: 10.1037/a0016496.

Roediger, H.L. and Karpicke, J.D. (2006) ‘Test-enhanced learning: Taking memory tests improves long-term retention’, Psychological Science, 17(3), pp. 249–255. doi: 10.1111/j.1467-9280.2006.01693.x.

Rohrer, D. (2015) ‘Student instruction should be distributed over Long Time periods’, Educational Psychology Review, 27(4), pp. 635–643. doi: 10.1007/s10648-015-9332-4.

Rohrer, D. and Taylor, K. (2007) ‘The shuffling of mathematics problems improves learning’, Instructional Science, 35(6), pp. 481–498. doi: 10.1007/s11251-007-9015-8.

Scott, R.A. and Kosslyn, S.M. (2015) Emerging trends in the social and behavioral sciences: An interdisciplinary, searchable, and linkable resource. .

Wiliam, D. and Black, P. (1998) Inside the black box: Raising standards through classroom assessment. London: School of Education, King’s College London.

Yan, V.X., Yu, Y., Garcia, M.A. and Bjork, R.A. (2014) ‘Why does guessing incorrectly enhance, rather than impair, retention?’, Memory & Cognition, 42(8), pp. 1373–1383. doi: 10.3758/s13421-014-0454-6.

first seen http://www.greatmathsteachingideas.com

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s