Make It Stick

Authors

Peter Brown, Mark McDaniel, Henry Roediger

First Thoughts

This could be the most important book I read this year, given the impact it could have upon my ability to learn and retrieve information.

Key Learning Concepts

1. Retrieval practice is important to cementing knowledge in your mind and preventing forgetfulness of a subject.
2. Interleaving learning increases retrieval practice implicitly, as you essentially have to do some form of retrieval to remember where you are in the context of reading / learning. (I.e. context switching can be good for learning, however this is probably not good for application of flow as there is significant time associated with context switching).
3. Repetition is not an effective way of learning, as it doesn’t test the mind in a way which encourages remembering. I wonder if this is a reason to try and describe a concept in different ways - so whilst it is repetition, it is forcing you to think about the problem in a format which is different to the original way it was described.
4. Massed learning (cramming) has the same downfalls as repetition in that it doesn’t stimulate persistent memory.
5. Learning in the form of a story builds a timeline which people can associate with memories, so commits these stories to memory. Is it possible that people can become more emotionally attached to a story or that they are more likely to reflect on a story later on, which will re-invigorate the memories.
6. Attempting to find a solution before knowing the answer improves learning, even if the answer you originally get is wrong. There is likely to be some essence of actions and learning requiring some previous knowledge, so it’s probably not something to really hold yourself back on. This is also going to help problem solving skills, whether this is in the pattern of understanding how you solved a problem where you previously did not know the solution, or if it mainly just covers that usually the problems you solve will include the solution you have already worked out, I don’t know.
7. If you can perceive the underlying principles of similar sets of problems, you are more likely to be able to recall how to solve a problem and be more likely to pick the correct principle to solve a problem when in unfamiliar problems.
8. Mechanical repetition has a limit, whilst elaboration, the act of taking the knowledge you have gained and expressing it in your own terms is boundless, such that there need be no end to your learning. So this is another good reason for this random blog in that if I can turn what I am reading into cognitive thoughts upon the subject, link it to existing knowledge I had, then the chances of me remembering are much higher. I guess the question on this will be is how much time can I afford myself to be able to effectively re-quiz the concepts in these books or test myself regularly as well as actually have a life 😅

The Importance of Retrieval

The recurring theme I believe will be brought to the fore throughout the book will be that retrieval is key. Early discussions in this chapter note how 53% of knowledge is retrieved in classroom environments when simple quizzes are interspersed throughout the studying period (one close to the initial learning, and then periodically inbetween). Where a set of students previously lost 13% of knowledge to recall, they stayed at a level with their initial recalling percentages after a week.

I wonder how this works in practice for reading a book - should I make notes on the book and then write myself key questions, to answer instantly and then at a later date?

Massed Practice = Massed Waste

This chapter carried the key theme of how repetition of a tasks with the aim of mastery is a strong method of forgetting said skill, and that varied, spaced, and interleaved practice are all more appropriate for longer term mastery of a skill, despite the perceived difficulties initially found in learning.

Perceived improvements in massed learning were often explicitly shown to students who would learn skills more consistently if they had spaced learning vs repeated massed learning - however, students still often felt they had learnt better with massed learning, which seems to be an apparent syndrome of humans struggle to contradict their own deep based beliefs. An area of perceived benefits is blocked massed practice which intimate varied practice, such as practicing 3 putts from 3 different distances in repeated cycles. Whilst this may appear to be varied, the repetition in cycles again prevents mastery of the skill in a way that a much more varied style would have, such as practicing 3 putts at short range at one hole, then 3 putts at a long range at the next, then perhaps 3 short putts again before moving to medium putts.

Real world integration of the skills is a solid way to practice and interleave “tests”, whilst having the key addition of implied space between testing forced upon you by you general work tasks. For example, doctors switching between patients in A&E have to context switch regularly, which creates the state of interleaving, and also creates a natural break in the repetition of treating a single patient / cause. This can be seen as similar to a programmer solving a problem in one domain and moving to a separate domain instead of staying in the same area. In the workplace or productive environment, it is likely that this will need to be thought about in the context of flow, in that people can be effective in a consistent mindset for an extended amount of time without having to context switch. The obvious answer here is what is already harped on about endlessly in this chapter, variation. 😅 Mixing your work between interleaved work where you are looking to do some work in a somewhat timely manner, but learning from it is almost as key to doing it, compared to work where you may already have solid foundations in the core concepts and you need to be efficient to meet a deadline.

A similar context of variation & interleaving is solving a climbing wall problem - if you split this into two pieces, the first and final third, if you learn to solve these problems you are more likely to understand how to solve the middle segment despite having not traversed it! You will have the context of the shape your body needs to be to lead into the final problem, and will understand the position you will be in leaving the first third.

Something which is mentioned regularly which I think has helped my learning is Reflection. Reading through this book reading a lot of places where I haven’t been effective with my learning, reflection is one skill I (ironically) do without thinking too much about. This is where you are able to consider the concepts being taken, look at them in different angles, consider them against things you already know, question them (& particularly question them against what you already know, given the importance of associating learnings with existing foundational knowledge).

Look into the Leitner flash card system - where you have four pots of flash cards with 4 levels of difficulty for you remembering. Practice the ones you find hardest most, but make sure you practice with the easier ones relatively regularly too. Practicing foundation knowledge on a less regular basis is key to retaining that knowledge on a level of mastery. If you happen to get one of the easier questions wrong, move it up to the more difficult pot. If you get one right, move it to the lower pot.

In difficulty the answer lies

Easy learning should be a thing of the past. To an extent 😺 Correlations have shown that when recalling information is more difficult, the actual learning on the back of it is more ingrained in your memory.

The three steps of learning appear to be:

• Encoding, where your initial learnings are adapted into your neural pathways.
• Consolidation, where the neural pathways are cemented through repeated learning.
• Retrieval, re-consolidating the learning you have already made. This helps you retrieve and associate knowledge in relation to what you’ve learnt.

This chapter re-affirmed that difficulty in learning leads to better results in a longer testing period, even when being tested on a separate technique (similar to the one which they have learnt) which a separate group have practiced regularly through massed learning.

This is regularly told in anecdotes of how people have learnt how to do a certain skill, despite having “zero prior knowledge” in it. It is more likely that the person has a lot of prior knowledge in similar skills, which they have used in varied and spaced situations, and then apply those similar skills to achieve the new skill. By then associating it with their prior knowledge, they are then able to more significantly store their new information.

It is important to show students that effortful learning and failing are parts of the learning process, as once they understand that failing is part of the process they are less encumbered by the anxiety and pressure of not getting questions incorrect, and are more likely to apply the whole of their cognitive capacity to solving the problem without trying to find cheats to the answer, or being worried about getting it incorrect on the first attempt. (This should probably exclude absolute guesswork, as that is not really thinking about the problem at hand).

Attempting to solve a problem before being shown the solution has a lot of benefits of priming the brain for learning, in such a way as you will begin to understand how the solution will be helped by understanding the solution, so you can then plug in the knowledge once you have gained it.

This is all part of Generative Learning, where people learn by doing, in that people build the knowledge needed by trial and error, and create stronger learning pathways because they have gone through the process of trying something, understanding why it has failed, moving onto the next step, and beginning to understand how this new way might work before failing again. Eventually something will stick, but it won’t be long before they fail again at something new where they are testing the boundaries.

These are all cases of Desirable Difficulties, where the difficulty is enough to encourage deeper thinking about the problem, reflection on why some paths haven’t worked out, before understanding any of the solutions. It is important to note that if no solution is found to difficulties, this does not help too much.

One final instance is to think about an instance of solving a problem on the climbing wall again. Often I’ve seen people try a problem again and again, failing at particular parts, only to go and spend 5-10 minutes on a different problem before coming back to flash the problem in moments. This is a case for interleaving and spaced learning, where the ability to take your mind off a problem helps it disconnect and start to apply other knowledge to the existing problem you had when you return to it. One professor once exclaimed something similar to me in Edinburgh, where he said the only time he really came up with novel solutions was when he truly understood the problem at hand, but the solution would generally come to him when he was doing something completely arbitrary at another time. The case of your brain cementing the knowledge with other prior knowledge, reflecting & consolidating and applying to new concepts. Eureka?

Know Not What You Know

The third chapter is on the illusion of knowing, discussing the fallibility of the human brain to illusions and misconceptions where there is either incomplete memory, or where the memory could have been compromised by imagining other cues or memories since the initial events or learnings.

The core concept is drives home the point of Daniel Kahnemmans book, Thinking, Fast and Slow, which describes the two states of the human mind: System 1, and System 2. The former is the intuitive & reactionary system which has ingrained knowledge over time & repeated learning. This is elsewhere referred to as the computer system. The latter is the minds ability of reasoning and decision making based on assessing the environment more carefully.

The brain is inclined to make quick decisions based on prior knowledge in system 1, however this can lead to poor decision making as it can ignore influences and factors which would lead you to take a different route if you had taken a more considered approach.

Accounts which sound true to the reader / listener, particularly if they have resonance with personal experience, are more likely to stick in the memory and embed beliefs even if they are not true. This is a simple example of why people believe cramming and massed learning are good methods of learning, as they have learnt through those methods and have it confirmed by other people telling them it is so. When compelling evidence to show them otherwise is passed to them, it can take a lot to override a belief in system 1.

Experts can be poor teachers based on their Hindsight Bias, or the Curse of Learning, where they have built strong mental models of the topic being taught and are unable to see where students won’t be able to join two concepts together, as they do not yet have the strong knowledge of a separate concept the teacher is expecting them to know. Students ended up being the best of teaching each other, as they are closer to understanding the best way to frame the teaching to the concepts they are currently able to grasp.

We all have differing opinions and views on the same topics. It is often taken that your view on the world is the same as everyone else’s. Confirmation Bias adds to this if you listen or read material which matches your own opinion.

People who overestimate their ability when they are unable to do tasks, are often condemned to not being able to improve. Their inability to see their deficits causes them to never leave their safety zone and learn further. It was also important to not overestimate how well you have learnt something, for example dropping a flashcard out of the Leitner system too early. This can prevent the spaced, interleaved recollection from fully taking affect.

Learning Styles are Learning Fails (Sort Of)

The core focus on this chapter is that learning styles are unsupported as concepts, and that whilst there may be some truth in their ideas these are generally empirically non-statistically relevant. However, there is one grouping which is empirically backed:

Creative Learning, Practical Learning, & Analytical Learning

These three groupings are shown to be of significance to a persons ability to learn and solve problems within that type; for example a person growing up in the countryside may have a stronger practical understanding of how plants are grown and crops harvested, however they may struggle at an analytical interpretation of their knowledge. This could be part of how Van Gogh saw country people, who he saw as simpler folk but who were much more in tune with their natural surroundings [in comparison to city dwellers]. The environmental aspects of country folk cause them to be much closer to the questions being asked of them by their surroundings, and as such learn from it successfully. City dwellers are not overly exposed to natural environments in such a way, and may struggle to learn as their background in analytical learning would not translate as well to practical learnings.

Finally, a key point was again that being able to comprehend and extract the underlying principles to a problem were key to being able to translate the problem to different areas & build concepts on top of the originals (i.e. understanding why the cog ratios matter in K-Nex meant you could expand outside the pre-built models that came with the instruction booklet). For example based learners it is useful to take two examples and compare them to start to see the underlying principles in practice.