Make It Stick

The Science of Successful Learning

finished (2020)Peter C. BrownHenry L. Roediger IIIMark A. McDaniel2014

1. Learning is Misunderstood

Most conventional wisdom about learning is wrong—for instance, the most common study hack (rereading / "review" of material) is among the least effective. Instead of retrieving information, one of the most effective ways to learn (think quizzes, flashcards, comprehension questions), students instead tend to rely heavily on repetition of information, which lends an illusion of comprehension through fluency with the text but ultimately proves ephemeral. In fact, reading something multiple times in a row has no statistically significant effect on learning over reading it once! Rereading is effective only when spaced out, giving the brain time to strengthen neural pathways. Highlighting? Reviewing? "Studying" (in the normal sense)? Nope! Not actually helpful.

What are then best ways to study and learn material, then? First, you must do it over a period of time (often interwoven with study of other subjects), so that unfocusing and refocusing build up the neural pathways associated with certain memories. Past that, seek recall-based methods, such as flashcards, rephrasing, answering questions / solving problems before you're told the answers, figuring out the core ideas behind topics and problems, and so on. Taking tests is actually one of the best ways to strengthen memory, incidentally; even if evaluative testing has flaws, it's perceptually helpful. Real-life scenarios? Even better!

Much research (yielding the above results) has been done in recent years in the burgeoning fields of cognitive psychology, developmental psychology, evolutionary psychology, and neuroscience. However, common wisdom (even in schools) has yet to catch up—most people's studying is unhelpful, even at top universities!

Most importantly, learning ability is not a fixed quantity—by practicing different techniques, almost anyone can learn how to learn well.

2. To Learn, Retrieve

Learning entails not forgetting—in fact this is the hardest part of long-term education by far. How do you get people to not forget things? By testing them, essentially.

The more retrieval-based evaluations—especially with feedback—that a class provides, the better students do, the longer they remember the material, and the higher they rate the class. Retrieval-based evaluations (even low-stakes quizzes) bolster learning by stopping the process of the brain's habitual forgetting. The sooner after learning this happens, the more pronounced the effect. It stacks, too—students taking three tests after learning something score as well after a week as students taking one test after learning do on that one test!

Generally, 70% of what we learn sticks around for a small time, probably in short-term memory. We learn—then we forget. The other 30% lasts longer, but slips away eventually.

Interestingly, testing is not the only way to interrupt our forgetting. Almost any kind of retrieval practice has this effect! The more effort it takes, the better it's learned; hence, a short delay (on the order of minutes) helps learning more than either none (too easy) or a long delay (too much is forgotten). Repeated retrieval (such as routine flashcards) is also more effective, similarly to repeated tests. Such frequent (self-)evaluation also gives a heightened self-awareness about both competence and weaknesses, enabling more efficient studying (hopefully retrieval-style!).

Finally, evaluations should give feedback (for use improving), but not instantaneously! Otherwise, it could develop into a crutch.

3. Mix Up Practice

Once again, common knowledge and habit proves wrong. We all practice things repeatedly, doggedly keeping at it until we improve. Once we're satisfied that we know or have mastered something, we move on to the next item. But that's not the right way to do it! This practice is called "massed learning" and is used to pick up things quickly—but, again, the hardest part of learning proves to be not forgetting what we know! Things learned through massed practice quickly slip away—we just don't notice that part, remembering (and coming to prefer) the swiftness of initial success through repetition.

A better—yet seemingly harder, in the moment—method is that called "interleaved practice", where different categories of practice are constantly mixed. It's harder, because you can't do things mindlessly, but retention is markedly better than massed learning! You also get practice distinguishing between different types of problems, something markedly absent in traditional practice.

This may explain the perennial ineffectiveness of textbooks: they're designed almost as altars to massed learning, topics being completely segregated along with their examples and practice problems.

Not only is the order of cramming wrong, but the timing too. Like when studying, things should be spaced out so as to allow the brain to transform memory traces into more permanent pathways. For full learning, time is necessary!

Lectures (especially at conferences), likewise, are bad—because as fast as information comes in, it proverbially "flows out the other ear." Information reception, as with practice, should be spaced out.

4. Embrace Difficulties

Your performance in the moment is not an indication of durable learning. On the other hand, when you let the memory recede a little, for example by spacing or interleaving the practice, retrieval is harder, your performance is less accomplished, and you feel let down, but your learning is deeper and you will retrieve it more easily in the future. (pp. 74–75)

"Learning" or knowing has two parts (besides "not forgetting," a plausible third): knowledge/memory o the material, and a reliable way to retrieve that memory. When we "forget" long-term memories (post-consolidation), we're most likely just losing the cue we used to bring them to mind—the memories are still there! Smells and tastes can often bring back torrents of recollections unexpectedly—this is the system at work.

The ease of retrieval and usefulness thereof to learning are inversely related: the harder it is, the better you'll remember it. That said, difficulty is only desirable when surmountable: if so hard as to be impossible, challenges are clearly not helpful. Push yourself to the uncomfortable edge of capability to practice recall—but no further! (Csikszentmihalyi, anyone?)

Generative learning is the process of trial-and-error. It's hard to come up with answers without background—but when you try, even mistakenly, it primes your brain to learn correctly. Those who guess and make mistakes learn the correct thing better than a control group.

5. Avoid Illusions of Knowing

Most of this chapter is an explanation of the Dunning-Kruger effect—the idea that those who are least competent don't know that they simply don't know—as a critique on unschooling and self-directing learning. If students don't have an accurate measurement of their performance, how can they be expected to know where they stand and what to do as a result? By the research of both Dunning and Kruger, it seems likely that as students enter and get an initial impression of a subject, they'll think their comprehension far outclasses its actual reality—and, as a result, make suboptimal progression and practice decisions. Self-directed learning should not exist in isolation, then: it should always be accompanied by evaluative recall practice and experts to gauge and aid progress.

Corrective feedback from peers and instructors is important. The better you know something, the less well you can teach it. Peers are better than instructors at helping new learners to work through roadblocks, because they have a similar perspective.

Memory is fluid and can easily be disturbed by incorrect recall, suggestion, leading questions, and others' conflicting accounts.

Our brain has various speeds of process: instinctive reflect acts first but uncritically, while more labor-intensive and accurate thinking is harder and slower. See "Thinking, Fast & Slow."

6. Get Beyond Learning Styles

The common notion of "learning styles" per person is not backed by evidence, but the idea that learning styles matter for different material (geography = visual, etc) is. Likewise, only three "dimensions of intelligence" have empirically been shown to be real: analytical, creative, and practical intelligence.

When learning, understanding the underlying material (and patterns between examples) is more important than just learning the examples. People are typically more "rule learners" or "example learners" (I didn't take very good notes on this part).

7. Increase Your Abilities

The authors call upon the work of Dweck and company to emphasize that human capacity or learning and brainpower is dynamic and increasable. Effortful learning, especially intentionally difficult and improvement-based practice, can actually expand and change mental ability! On top of that, self-discipline, persistence, and grid are keys to all kinds of success—and can clearly be learned and improved.

Memory is especially prone to improvement; mnemonic devices and the like can be used to learn massive string of arbitrary data quickly and precisely. One common device is to picture a familiar space, then associate imagined figures within said space with the information you need to remember. Holding this mental image in mind, you can then progress through it, from character or action to character, much more easily recalling information than if it were just strings of arbitrary text! Memory masters can use even more sophisticated strategies, such as assigning each number a sound, then stringing together sets of three sounds into 1000 number-matched words/pictures, which then be assembled into such scenes or storylines to remember large strings of numbers.

8. Make It Stick

No new learning content, just a compilation of tips and examples throughout the book (plus a few new ones). You could probably read this chapter and get 90% of the book's lessons. Below is new content

  • Salespeople use the FORE mnemonic for getting to know someone new: learn about their Family, Occupation, Recreation, and what they Enjoy doing.

  • It's better to give people algorithms than general targets, especially when learning new skills (in a business environment). The FORE method is much better than saying "go get to know someone" or even "learn these things about someone"!

  • Almost all studying—at least in medical, fact-based fields—is better when you're asking the text questions then looking for answers. If you're reading every word, you're using a textbook inefficiently.

  • Unsurprisingly, "learning" improves when classes have grades attached to regular retrieval practice. "Highly structured" classes with quizzes, tests, practice questions, reflections, mandatory notes, etc generate higher average scores than "traditional" classes with only high-stakes tests. However, it doesn't seem that they consider efficiency (how much time is spent per "learning"); this performance increase might be attributable to just getting people to spend more time on a class (rather than being a better per-time method).

if this is your website, log in to make changes