Seven Steps, From Everything You Bring Before Class Through the Exam and Beyond
by Professor Stephen T. Abedon Steve Abedon, Ph.D., has taught well over 100 individual classes and ~10 individual courses (including introductory biology, microbiology, physiology, biochemistry, etc.) over his 30+ years as a college professor. He has over 150 publications, including multiple monographs, and is ranked in the top 1 or 2 percent among researchers of all disciplines worldwide. His research focuses on bacteriophage biology and phage therapy. (abedon.1@osu.edu)
phage.org | phage-therapy.org | biologyaspoetry.org | abedon.phage.org | google scholar | publications
Jump to: 📋 In Brief | 🩹 Diagnostic | 0️⃣ Step 0 | 1️⃣ Step 1 | 2️⃣ Step 2 | 3️⃣ Step 3 | 4️⃣ Step 4 | 5️⃣ Step 5 | 6️⃣ Step 6 | 📖 Glossary | 🏫 For Teachers | 🔖 Index | 🧮 More Calculators
biologyaspoetry.org/laboratories/studying/ · Abedon’s Books
How can I improve this page? contact: studying@biologyaspoetry.org
Almost everyone who gives study advice was good at school. That is worth knowing before you take any of it, including mine. I was a reasonably good student (99th percentile on the ACS general chemistry exam and 95th on the organic exam; CRC chemistry book award recipient; Cum Laude; Phi Beta Kappa; 3.85 GPA; at least interviewed by MIT for graduate school), but I was a reasonably good student because I worked hard at it — in biochemistry, a subject which does not forgive a method that mostly works. Nothing here is a shortcut around effort. All of it is about making effort pay instead of evaporating.
That list is also less impressive than it looks, and it is worth saying so early. It came out of a particular few years with a particular cause behind them, and the same intensity was gone by graduate school. Step 4 has the longer version.
Most students who are struggling are not lazy and are not incapable. They are running a strategy that fit every course they have ever taken, and have now met one it does not fit (mine was quantum mechanics in physical chemistry). That is not a character problem. It is a mismatch, and the useful response is to find out where the mismatch is.
Click any step to open it.
If you are here because an exam went badly, the Diagnostic is the faster route in: four questions about one question you got wrong, and it will tell you which step to read.
Rereading feels productive. A highlighted page feels known. Two-column notes in three colors feel like understanding. An answer that looks familiar feels correct. Every one of those is the same mistake: ease produced by the material being smooth, mistaken for ease produced by you having learned it (the fluency illusion). It is why the techniques that work feel worse while you do them, and why people abandon them.
The other half of the same fact, stated positively (the generation effect): you keep what you produce, not what you receive. Producing well is where the effort goes. That is the work.
Read that as a one-way test, because that is all it is. Ease is reliable evidence that something is wrong. Effort is not reliable evidence that something is right. Eight hours a day of rereading, highlighting, and recopying is genuinely exhausting and still gone in two days.
So the full form has two halves. If it doesn’t feel like work, you almost certainly aren’t learning. And feeling like work only counts when what you are producing is the thing you will later have to produce — from memory, in your own words, under time pressure.
Students run the sequence once and stop. That is the single most common structural mistake, and it is why studying can feel complete right up until the exam says otherwise. The traffic runs in both directions:
3 → 2 Organizing sends you back to your own raw notes.4 → 3 Studying turns up a hole; back to the organized set.4 → 2 The organized set doesn’t have it; the raw notes might.→ 0 None of them have it; back to the textbook.→ out Book and notes both run out; now go outside them —
a search, an assistant, a classmate, the instructor.
Where the loop runs out is where you find out what kind of problem you have. If the answer is nowhere in any of your own materials, the failure sits at Step 0 or Step 2. If it was there all along and you still could not produce it, nothing upstream is broken and the problem is Step 4. Those need opposite responses, and they feel identical from the inside.
The Diagnostic walks that chain for a question you actually got wrong. Step 6 is what you do with the answer.
Each of these works with no prior skill, costs very little, and tells you something within a week. Start here rather than with a system.
Not a system. One cheap notebook, no categories, no color scheme, no filing. Every due date, every announcement, every this will be on the test goes in it. This asks nothing of your understanding, which is exactly why it is the right place to start: in a week you will be able to look at it and see something it caught that you would otherwise have lost. That is a checkable promise, and once it pays off, the harder claim — that writing is where thinking happens — is much easier to believe.
From memory. No notes, no slides. Then compare. Most people find the gap shocking, and the shock is the point: you will have just watched the difference between having followed something and being able to produce it. That gap is the entire argument of this page, and it is cheaper to meet it now than on an exam.
Given both modes, students overwhelmingly pick multiple choice, and they are not
cheating — they pick the mode that feels like it is working, because it
returns a steady stream of correct answers. But recognizing a term among four options
and producing it from nothing are different skills, and only one of them is what an exam
asks for. Same material, same afternoon, opposite value.
termschallenge.biologyaspoetry.org →
Each exam question you trace on the Diagnostic tab is kept as you go: your four answers and the steps they pointed to. In most browsers they are still here if you leave and come back on the same device, though a private window, a different device, or cleared browser data will start you fresh. When you are done, download a summary to keep, so that you can look back at which steps keep coming up.
These are other people’s accounts, listed by the step they bear on. Some are research based and some are one person’s experience; both are worth having, and it is worth knowing which is which as you watch.
A warning about searching for more. The further you get from institutional sources, the more confident the neuroscience becomes and the more the advice converges on the same handful of real techniques wearing new names. Unsourced percentages, named brain regions, and “studies show” without a study are all worth noticing.
Most people arrive at a page like this after something went badly, so let us start there rather than with a method. Take one question you got wrong on an exam — a specific one, not a general feeling about the exam — and answer the four questions below about it.
What comes back is a place in the sequence where that question was lost, and a tab to read. That is the whole point of numbering the steps: the seven steps exist to be diagnosed against. “Why am I bad at this?” has no useful answer. “Which step is failing?” does, and the fixes for different steps are different enough that guessing wrong wastes a semester.
No exam to look at yet? Then start at Step 0 and read forward. Come back here after the first one.
The chain in question 1 is the useful part, because each answer localizes a single handoff:
| Where the answer was | What that tells you |
|---|---|
| Not findable anywhere | Either it came from somewhere you did not know was fair game, or it was not fair. Worth raising with the instructor, politely and specifically. |
| In the textbook or lecture, not in your notes | The material was available and never got captured. Step 0 or Step 2. |
| In your raw notes, not the organized set | You had it and dropped it while deciding what mattered. Step 3, and the most interesting failure on this list: your effort was fine, your judgment was off. |
| In your organized notes | Nothing upstream is broken. Step 4 — and almost certainly recognition standing in for production. |
That last one is the most common and the least expected, because everything up to it looks right. The notes are good. The work happened. And a wrong answer is not a verdict — it is a repetition, and the whole reason retrieval practice works is that failing and correcting is the mechanism. The only genuinely wasted wrong answer is one you never look at again.
Step 0 is not just a task. It is the material every other step operates on, and it is why two people can sit through the same lecture and not receive the same lecture. It has two halves, and both of them are mostly settled before the semester starts.
Previous courses. Whether you ever took chemistry, and how much of it stayed. What you happened to be curious about at fifteen. A student who looks at the formula for a carbohydrate and calls it an amino acid is not missing a fact from last week’s lecture — amino acids contain nitrogen, and that is not a detail attached to the definition, it is what the definition means. Which in this case involves also looking at a chemical’s actual structure, a supremely important aspect of biochemistry. That gap is Step 0 surfacing late.
Prior capacity matters as much and gets discussed less. Whether you have ever sat with one problem for an hour. Whether you can tolerate not understanding something for a while without concluding you are stupid. Whether boredom registers as information or as an emergency. None of that was built in a study session; it was built by whatever you have spent your time doing.
Which explains something otherwise puzzling. Retrieval practice is not harder for someone with a low tolerance for not-knowing — it is the same technique, with the same steps. But the discomfort it produces is unfamiliar, so it gets read as evidence that the method is failing, and then abandoned.
Capacity is not built in a study session, and it is not built by deciding to have more of it. It is built by what you habitually do, which is a slower and more interesting answer than most study advice offers.
One researcher’s way of putting it, about screen time but general enough to be worth borrowing: do anything every day for 133 minutes and it will change you. Not a claim about any particular activity — a claim about what a daily block of time spent in a given mode does to the person spending it. If a substantial part of every day goes to something that demands nothing, offers no stakes, and has no natural end, then that is what is being practiced. And attention is not the only thing being trained. Staying with something long enough to care about it is a habit too, and it can be practiced in either direction.
Which suggests something better than a study tip. Being attentive and involved works much better as a disposition than as a mode you switch on at a classroom door — and if it has to be switched on, you are already fighting. That is worth thinking about at the level of how you spend a week, which is Step 1’s territory but is built here.
The verdict students most often reach for is I’m not good at this. Dennis Shaffer suggests replacing it with something more accurate: I am weaker at this right now, and time spent on it will change that. The framing is Carol Dweck’s growth mindset, and what it changes is how a setback reads. Under the first sentence a setback is a judgment of what you are. Under the second it is information about what to work on. The second is also simply true: what you know is changed by what you do, which is the premise of everything else on this page.
Closely related, and with solid evidence behind it, is self-efficacy — Albert Bandura’s term for your belief that you can accomplish a specific task. Shaffer cites a meta-analysis of more than a hundred studies in which academic self-efficacy was the strongest psychosocial predictor of college achievement. A predictor is a correlation, and belief and past success certainly feed each other. The practical point survives either way: people who believe they can do something keep going after a setback, and keeping going is most of what a setback asks of you.
How much a short lesson in mindset changes anyone’s grades is genuinely debated, and some large trials have found little effect. The reframing is worth having for a simpler reason, which is that it is accurate.
The framing is Amy Wang’s. Adding a new habit means new work in a schedule that is already full, so it competes with everything else and usually loses. Removing something that is already costing you returns time and capacity without asking for either. If your real problem is that there is no room, subtraction is the only move that does not require finding room first.
The general version, which you can answer about your own week: before adding a habit, look for what is already costing you and returning nothing.
The distinction is Thomas Frank’s, from Crash Course Study Skills, and it is the most useful single idea I found in any of this. There is an irony worth noticing (glossary). Students who will not let themselves take two real hours off, because they have too much work, routinely lose three hours in five-minute pieces to things they do not even enjoy.
So the counterintuitive instruction is that you need the high-density kind, deliberately scheduled. Denying yourself real time off is what leaves the low-density kind as the only available relief — and the low-density kind is also the daily block that trains the wrong disposition.
Advice about textbook reading contradicts itself constantly, and the reason is that a textbook is doing three jobs and almost nobody separates them. Sorting out which job you need is most of knowing how to read it.
The goal is exposure, and the currency is familiarity. You are not trying to master anything. You are trying to make the material stop being new before you meet it again. That is cheap, and the cheapness is the whole point.
Here the job is retrieval rather than understanding, so marks earn their keep: flags, brackets, a dot in the margin, anything that gets you back to the right page in fifteen seconds. This is also the mode every open-book exam runs in.
In some courses the text is the material rather than an account of it — a novel in a literature course, a primary source in history, a paper in a seminar. Here “textbook” is being used broadly. In those courses the expensive kind of active reading is entirely justified, because there is nothing else to spend the effort on.
If your current preparation is zero, the useful instruction is not a system. It is: look at the chapter. Deeper is better, and almost anything beats nothing, because nearly all of the benefit of reading before class comes from familiarity, and familiarity is cheap. It does not require mastery.
Cheap moves, in rough order of return per minute:
The problem with marking a text is that deciding what matters is a different operation from reading, and doing both at once means the reading gets choppy and the decisions get made too early. There is a neat way around this, sometimes called the Morse code method:
Two marks, no slowing down, and when you come back the structure of the chapter is already annotated — which is to say you have left yourself the raw material for Step 3 without having done Step 3 while reading. If you own the book, pencil; if you rent it, flags or sticky tabs work identically.
Some of the most commonly wasted student time goes here, so it is worth being blunt. Skilled adult readers run roughly 200 to 400 words per minute on difficult material. Past about 400 you are skimming, and comprehension falls away quickly. The standard speed-reading techniques have been tested:
What actually makes you faster is unglamorous: read a lot, read things that are hard enough, and know what you are looking for before you start. The last of those is free and is what the cheap moves above are for.
Most reading advice treats the textbook and the lecture as two parallel sources feeding the same exam, and concludes that if they overlap you can skip one. That conclusion follows from the premise, and the premise is wrong for a science course. The redundancy is the point. Meeting the material twice is the mechanism, not the waste. Specifically:
Asynchronous courses reorder this. The instructor is not present, so reading first loses its best payoff — but you gain something better, which is that the textbook and the lecture can be run together, pausing one to consult the other. That makes the order conditional on how easy the book is to navigate:
In science courses the text and the lecture rarely diverge far, which is what makes this work. In a course that departs substantially from its text, it will not, and the text may be nearly unnecessary — but that is uncommon in the sciences, and worth figuring out rather than assuming.
Attendance is not presence. It is entirely possible to be in the room, awake, facing forward, and not there — with a phone in your lap, or eleven tabs open, or having decided somewhere before the hour started that this one does not matter much.
That distinction explains something instructors find genuinely puzzling. When I say this will be on the test, a good number of students write nothing down. That is not a note-taking failure. Note-taking is Step 2, and it never got a chance. It is a presence failure, and it usually has a perfectly rational history behind it.
I used to run triathlons, and the week before an event does not look like the months before it. You taper the training down so that you arrive fresh rather than fit — you cannot get fitter in the last week, and trying only costs you. You adjust what you eat. You check that the equipment works. You think through roughly what your strategy is going to be, while knowing you cannot plan for everything.
None of that makes you faster. What it buys is the ability to give the event your whole attention, with nothing avoidable pulling at you, so that you can stand at the start and think: I have done what I can do to be as effective as I can be. Bring it on.
On a much smaller scale, that is the state of mind to walk into a classroom with — and it is where Step 0 and Step 1 meet. Step 0 is the training. Step 1 is race day. Which raises the awkward question:
You are physically there. But preparation does not show up as knowledge so much as in the quality of your presence. Arriving unread means spending the hour finding out what the topic is rather than following where it goes — which is a way of having already given away the most valuable thing about being in the room. Ten minutes with the chapter beforehand is the taper.
The analogy has a limit worth stating. A taper works because the training happened. Arriving rested and empty is not the same as arriving prepared, and no amount of being well rested substitutes for having read the chapter.
There is a useful way of putting this from outside education entirely. We try to get more done by finding more hours, and hours are fixed — a hundred and sixty-eight a week, and never one more. Energy is the other resource, and unlike time it is internal, which means it can be spent well or badly and can be renewed.
Which lands exactly on the distinction this step is about. An instructor has your fifty minutes whether or not you slept, whether or not you argued with someone at breakfast, whether or not you are thinking about the other four things you owe today. Having your time is not the same as having your engagement, and only one of those does you any good.
“Paying attention” is only half of what being there consists of, which is why removing distractions never quite finishes the job. Phillip C. Schlechty (1937–2016), an American educator who spent his career on school reform and student engagement, set out a framework in 2002 that separates engagement into two independent things: attention — being focused on the task rather than elsewhere — and commitment — finding the task worth doing for its own sake. Cross them and you get five recognizable states:
| State | Attention / commitment | What it looks like from inside |
|---|---|---|
| Rebellion | diverted / none | Actively doing something else, and not quietly. Nothing is learned. |
| Retreatism | none / low | Checked out. Not disrupting anyone, just gone. Little or nothing is learned. |
| Ritual compliance | low / low | Doing the minimum to avoid trouble. This one is worth being able to name, because it feels like attending. |
| Strategic compliance | high / low | Working hard, performing well, aimed squarely at the grade. Looks like engagement from outside and often is not. |
| Engagement | high / high | In it. Stays with the thing as it gets harder, and would keep going even ungraded. |
The useful cell is strategic compliance, because it is where good students live and because it answers a question this page keeps circling: how can someone work genuinely hard and end up with nothing durable? Schlechty’s account is that such a student learns at a high level and then fails to retain it or carry it into a new context. Which is exactly what learning to tell definitions apart from one another is — high attention, aimed at the grade, and it evaporates the moment the question is worded differently.
Dennis Shaffer arrives at the same place from a different direction. His advice is to focus on learning goals — getting better, improving what you know and how you work — rather than performance goals, meaning how you do on this quiz or that exam, at least at first. Good performance follows from good process; good process does not follow from chasing performance. And performance goals change from task to task, from course to course, and from college to whatever comes after it, while learning goals carry across all of them (learning and performance goals).
It also gives you a better question to ask yourself than am I paying attention?, which has one embarrassing answer. Which of these five am I in, in this class, today? That is answerable, the states are recognizable from inside, and you can usually move up one.
Schlechty built this as a framework for teachers designing tasks, not as a measuring instrument, and it is worth taking as a way of organizing experience rather than as a set of findings.
Everything on this page is technique, and technique is downstream of caring. A method makes effort pay off; it does not manufacture the willingness to spend it. There is an old observation that good students do well more or less regardless of the instructor, and what it is really describing is the difference between being intrinsically and extrinsically motivated. To be fully there, you have to actually want to be. There is no cure for not caring.
Two things make that less bleak than it sounds, and both are worth checking before you conclude anything about yourself.
Those five states are levels in this class, today — not personality types. Which is why the same person is genuinely engaged in one course and ritually compliant in the next. “I don’t have it” is a verdict; “I don’t have it here, yet” is a question with somewhere to go.
A course that has been incomprehensible for three weeks produces not-caring as a result. So does a course that demands nothing. Boredom from having nothing asked of you and boredom from everything being out of reach feel identical from inside, and both feel like not giving a damn — but each has a fix, and neither is a dead end. If you genuinely do not care, no page will help. Check first which one you have.
The deeper version of all this is that attention and caring are better thought of as dispositions than as modes. If you have to switch them on at the classroom door, you are already fighting. If they are simply how you go about things, then the classroom is one more place where you do what you always do — and being present there stops being an act of will.
This is the same point as the taper. Nobody becomes a person who trains on race morning; race day is continuous with how they live. Someone who is curious about why their car is making that noise, or what a plate spinning in the air is actually doing, does not have to generate curiosity for a lecture. They arrive with it.
Which is worth turning into a question rather than a verdict. If being attentive and involved in class feels like something you must muster, that is information about more than the class. And it is genuinely the one thing on this page that cannot be fixed in a week — a disposition is built the way any disposition is built, by what you habitually do. It may also be the most valuable thing here.
Note the escape route, because this can otherwise read as a judgment on a person. Almost everyone who feels disengaged in class is deeply engaged with something else, which proves the capacity is entirely intact. The question is only what it is currently attached to.
There is a chain here worth following, because it explains something that otherwise gets blamed on the subject or on the lecturer: no preparation, no engagement; no engagement, no creativity; and without creativity it is boring.
By creativity I do not mean anything artistic. I mean having something to do with what is being said. Prepared, you can connect this to the thing from two weeks ago, object silently, predict where the argument is going, notice that it contradicts something you read last night, or wonder why it is not the other way around. Every one of those is a small act of making, and making is what turns an hour into something interesting rather than something to be endured.
Unprepared, none of that is available. You are receiving, and receiving is passive by construction — there is nothing to be right about, nothing to be wrong about, and nothing of yours in the room at all. The same thing happens later in the sequence, at Step 4: rereading is duller than testing yourself for exactly this reason, because one of them asks you to produce something and the other does not.
Which means boredom in a lecture has two distinct sources, and they want different responses. One is the difficulty being mismatched — too easy or too far out of reach. The other is that you have nothing to make of the material, which is usually a Step 0 problem rather than a Step 1 one, and which no amount of sitting up straighter will fix.
Unlike most of what this page covers, Step 1 does not have a starting problem. The lecture begins whether or not you feel like it, there is no blank page, and nothing has to be initiated. What has to be protected is attention, and there is a better test for that than “don’t multitask,” which nobody follows anyway:
The failure worth naming first is not obvious inattention. Andrew Deutscher, in a TEDx talk on engagement, calls it the gray zone — straddling two worlds and doing a disservice to both. His example is standing in a doorway after a week of work travel, pushing his five-year-old on a swing while answering his phone, until the boy asks: why do you go to work if you do work at home?
A student in a lecture with a messaging window open is in exactly that position, and the reason it is hard to correct is that the gray zone does not feel like not paying attention. It feels like handling both. That is the illusion to watch for, and it has a simple test:
If no, it is probably harmless and may even help. If yes, it is not a distraction — it is a second task, and one of the two is going to lose.
Instrumental music mostly does not compete: it occupies the channel that would otherwise be scanning the room. Lyrics do compete with reading, because both want language. Walking occupies your legs, which the thinking was not using at all, which is why people think well on a walk. But reading a text message while someone explains something to you is two streams of language fighting over one piece of machinery, and only one of them can win. A lecture wants exactly what a phone wants, which is why a phone is not a small problem.
Shaffer is stricter about music than I am: none while studying unless it is familiar and has no lyrics. Familiar is the useful addition. A song you have never heard asks to be listened to, and listening wants the same part of you that the work wants.
This also explains what switching actually costs, which is more than the seconds it takes. Part of your attention stays behind on the thing you left, and that residue is expensive precisely because the new thing wants the same equipment the old one was using. Then there is the ramp-up: coming back is not instantaneous, because you have to rebuild where you were before you can go anywhere from there.
And in a lecture there is a worse version. If what you missed was load-bearing, you do not merely have to get back up to speed — you may be unable to. The next ten minutes are built on the thing you were not there for, so the cost of the interruption keeps compounding long after the interruption is over.
Andrew Deutscher, in the talk mentioned above, puts it as the difference between a sprinter and a marathoner. A sprinter can empty the tank because the end is visible. A marathon with no finish line anywhere produces something else entirely — a pace that can be sustained indefinitely and that is never full effort. His point is that modern working life has lost its finish lines, and that a great deal of the day gets spent in that second mode: present, at partial throttle.
A class period is one of the few genuinely sprint-shaped things in a student’s week. It has a start, an end, and a length you already know. That is worth using deliberately: fifty minutes is a distance you can run kind of flat out in terms of your attention, and running it that way is easier than hovering through it. Then it is over, and being over is the point — a real stopping point is what makes full expenditure affordable in the first place.
The same principle turns up again in Step 4, where the finish lines have to be built rather than found, because studying has no bell.
The most consequential moment in any lecture is the one where it stops making sense. The pull right then is to check out — and checking out does not cost you the minute you did not follow, it costs you the rest of the hour, because everything after that is built on what you missed.
Staying present through not understanding is a skill. It is trainable, it is uncomfortable, and it is probably the single highest-value thing you can do in a classroom. Some ways to hold on:
Nobody sustains fifty minutes of full attention, and nobody ever has. Attention is not a switch that is on or off for the hour — it comes and goes, and it always has. What separates students is not how often they drift but how fast they notice and return. Noticing is itself the skill, and it improves with practice in exactly the way you would expect.
Which is worth knowing, because the alternative belief is corrosive: if you think attention is supposed to be continuous, then every lapse is evidence that you are bad at this, and evidence of that kind makes the next lapse likelier.
Presence exists in order to catch things, and a few of them are worth far more than the rest. An instructor typically covers less than the textbook and holds you responsible for less, so what gets said in the room is largely what defines the scope of the course — and that exists nowhere else.
Catching those is also where Step 1 hands off to Step 2: they are precisely what should end up on the page, and a transcript of the slides will not contain any of them.
Presence is participatory, and a question asked in the room costs about twenty seconds. The same question a month later, alone, costs an hour — if you ever get back to it, which mostly you will not.
Worth sitting with, because it settles what live attendance is actually for in a course with recordings. The content will still be there tonight. The slides will still be there. The transcript will still be there. What will not be there is the person who can answer you, at the moment your confusion is still small and specific. In a hybrid or recorded course, the whole marginal value of the live hour is the question — which means going and not asking anything is close to the worst available use of it.
Almost nobody asks enough, and the reasons are consistent. Each has an answer, and none of the answers is “be braver.”
| What stops you | What is actually true |
|---|---|
| I will look stupid. | Nobody has ever thought worse of a student for asking. Instructors remember the ones who ask, and remember them favorably, because a question is evidence of someone paying attention hard enough to notice a gap. |
| Everyone else seems to get it. | They are also sitting quietly hoping somebody else speaks. This is close to a rule: if a sentence lost you, it lost several other people in the room. The silence is not evidence of comprehension, it is evidence of the same reluctance you are feeling. |
| I do not know what to ask. | The real obstacle most of the time, and it has a method rather than a pep talk. See below. |
| I do not want to slow everyone down. | A question takes twenty seconds and frequently saves the next ten minutes for several people. And it tells the instructor something they otherwise cannot discover until they are grading the exam, which is much too late to act on. |
Shaffer points out that looking foolish, or feeling incompetent, is among the most common fears people have, and not only students. His own practice is the useful part. When he drafts a paper for a journal, he sends it to his coauthors and to someone who does not know the research, and asks them to criticize it hard. The journal’s reviewers then see only the polished version. Exposing something while it is rough, to people who can help, is how it gets good before it counts.
“I do not understand the lectures” is a real and common condition, and it is unanswerable as stated — not because it is a bad thing to say, but because there is nothing in it for anyone to take hold of. The work is turning it into something specific, and that work is most of the understanding:
Asking in front of everyone is only one of them, and for many people it is the least comfortable. These are roughly in order of cost:
In a course where the lectures are recorded and the live session exists mainly for questions, treat that session as what it is. Arrive with your questions already written down, because you will not generate them on the spot — the confusion happened last night, in front of the recording, and by the time you are sitting in the room it will have faded into a general sense that something was hard.
It is worth saying plainly, because students routinely treat office hours as a last resort for people in trouble. They are not. They are the purest available form of what this whole step is about: a live encounter with the person who knows the answer, with nothing competing for your attention and nobody else’s questions in the way.
Everything above applies there, more so. Being fully present matters, being willing to say what you do not understand matters, and arriving having done something first matters most of all. What separates a useful office-hours visit from a wasted one:
There is a further reason to go, which has nothing to do with the material. Instructors are more use to students they know — about what to prioritize, about whether to worry, about what is coming, about the course after this one. That relationship does not form during a lecture, and it cannot form over email. It forms in the ten minutes when you turn up.
Practically: office hours are also where a retake, a second attempt, or a walk-through of a returned exam usually happens, which makes them the natural home for Step 6 as well.
One last reason, which is about you rather than about the answer. Formulating a question requires deciding exactly what you did not follow, and that decision is most of the work of understanding it — which is why questions so often dissolve while you are writing them down. Shaffer counts asking as a form of the generation effect for the same reason: you produce the question yourself, and then fit the answer into how you already understand the material. It also converts a passive hour into a participatory one, and that is the difference between two of the five states described earlier on this page.
Capture — writing down what needs to be learned — runs on someone else’s clock. The lecturer keeps talking whether or not you have finished thinking, which means the one thing you cannot do here is stop and consider. That is not a flaw in your technique. It is the situation.
Which is why the standard advice — don’t just transcribe, engage with the material as you go — asks for something close to impossible. It is asking you to do Step 2 and Step 3 simultaneously, and since only one of them has a deadline attached, thinking is what gets dropped. A student who transcribes is not failing at note-taking. They are doing Step 2 correctly but may never do Step 3.
It is worth asking this plainly, because the answer is not obvious to everyone and the alternatives are worse than they look. If you do not take notes, then when the exam is two weeks out, what exactly are you sitting down with?
None of those is nothing. All of them are considerably less than a page you made yourself, and the reason is not mainly that your notes are more accurate. It is that making them required decisions, and the decisions are the part that stays with you.
Plenty of instructors hand out worksheets that mirror the slides with gaps to complete, and there is a real argument for them — they guarantee that everyone leaves with the skeleton, and they are a kindness to anyone who genuinely cannot keep up.
The cost is worth knowing anyway. A gap to fill tells you exactly what to write and exactly where, which means the two decisions that make note-taking valuable — what matters, and how to put it — have both been made for you. Filling in a blank is nearly free, and things that are nearly free tend to produce nearly nothing. If you are given one, the useful move is to treat it as the floor rather than the ceiling: fill it in, and then write your own material in the margins alongside it.
Dennis Shaffer, who hands out skeleton outlines of his own lectures, describes the failure one level further down: students who look at the outline and never fill in the blanks at all. Then even the floor is missing. Generating the notes yourself, in your own words, is the part that helps memory, and an outline only gives you a place to do it.
Working memory holds something like four items at a time. That is not a figure of speech, it is a limit, and it is why a page is necessary rather than optional. You take notes so that you can follow something too big to follow in your head.
Everything above might read as an argument for writing down only what the lecturer says. It is not. Your own thoughts belong on the page too, and they are frequently the most valuable marks on it — the objection that occurred to you, the connection to something from three weeks ago, the question you would ask if the room were empty, the idea that has nothing to do with this course at all.
So the rule is: catch it, do not chase it. The thinking happens in Step 3, which is exactly what Step 3 is for.
Here is where Step 0 pays off in a way that is easy to miss. If everything being said is new to you, every second of the hour is spoken for, and there is no spare capacity for an aside even if you have one. Having read the chapter beforehand changes the arithmetic: you are not writing down what is already in the book, you recognize where you are, and the attention that would have gone to getting content onto the page is free for something better.
The more preparation, the more freedom to do something other than transcribe. That is the practical return on ten minutes with a textbook, and it compounds — the less of the hour spent copying, the more of it spent thinking, and thinking is what you were there for.
There is a real finding here and it is easy to misread. Asked to take notes on the same fifteen-minute lecture, laptop users wrote about 310 words on average and longhand writers about 173 — and the laptop users recalled less when tested later.
The mechanism is the useful part. Typing is fast enough to let you capture nearly verbatim, and your working memory can only handle so much at once. Spend it on the words and there is none left for the meaning. Handwriting is slower, so it forces you to select — and selecting is thinking.
So the conclusion is not that paper wins. Speed is the hazard, not the keyboard. A typist who deliberately types less gets the same benefit. Paper simply has the advantage of enforcing it without any self-control, and of not having a browser attached.
Worth knowing honestly: this study has had mixed replication. The direction holds up better than the size of the effect. The mechanism — that verbatim capture crowds out meaning — is the durable part.
Typing notes directly into an AI assistant during a lecture works, and the reason is not the one people assume. There is no conversation happening — there is no time for one. You are typing as fast as you can into a box, and the same speed hazard applies as with any other keyboard.
What it changes is the cost of Step 3. The notes come out legible, searchable, cumulative across a whole semester, correctable, and sitting in the one place that can immediately quiz you on them. The usual reason notes die is friction, and this removes most of it.
And the same holds for notes typed anywhere else — a document, a notes app, anything on a computer can be dropped into an assistant afterward. There is genuine value in that interaction, provided it runs the right way round: ask it to find what is thin, to quiz you, to point out where your version of an idea does not hold together. Ask it to summarize your notes for you and you have handed away the one part that was doing you any good. The rule, in full, is in Step 4.
Notice, though, that the freedom to do any of this comes from the same place as the freedom to write an aside. If the hour was spent racing to get content down, there is nothing on the page worth interrogating.
Not everyone knows this is available, so it is worth stating plainly: if a lecture is recorded — by the instructor, or by you with permission — that recording can be turned into a text transcript in minutes, and a transcript can be handed to an AI assistant. Many courses now post recordings as a matter of course, and phones transcribe well enough to be useful.
Used properly, though, this is one of the genuinely new things available to students now, and the best use of it is not summarizing. It is comparison: what did the lecture contain that your notes do not? That is a question your notes cannot answer by themselves, for the same reason a survey of returning aircraft cannot tell you where the fatal hits land — and it is worked out in Step 3, where it belongs.
The heading is the mathematician Eric Temple Bell’s line, which Crash Course Study Skills applies to note-taking. A map is useful because it leaves things out. So are notes. You cannot record everything, and a set of notes is a reviewing tool only if what matters is in it and what does not is not. Some guidance on what to catch:
“I am not following the lectures” is the symptom. Note-taking is one of the places the cause can be, and there are five states you might be in. They need different responses, and telling them apart is worth more than any general advice about taking better notes.
| State | What it means and what to do |
|---|---|
| No notes | Nothing left the room with you, so there is nothing for Step 3 to work on and nothing for Step 4 to study from. Usually a Step 1 problem rather than a skill problem. |
| Too few | You have a page, and three weeks later it does not reconstruct — you can read your own writing and not know what was happening. Take denser notes: more of the reasoning, more of the examples, more of what was said rather than what was on the slide. |
| Too many | You have a transcript. The problem is not the volume in itself but where the volume came from: it was paid for out of the same attention budget that was supposed to be following the argument. This is the state most likely to produce I took notes and still do not understand it. |
| About right | Dense enough to reconstruct, sparse enough that you were listening while writing. You will know because the notes make sense later and the hour did not feel like transcription. |
| Enough notes, no understanding | The notes are fine and the trouble is elsewhere — most often Step 0, where the material was new when it arrived, or Step 4, where having good notes was mistaken for having learned from them. This one gets misdiagnosed most, because a good set of notes feels as though it ought to be sufficient. |
Notice the problem with all of this: the decision has to be made while the lecture is running, sentence by sentence, and you cannot take five seconds to weigh it. So what actually determines whether you get the amount right is not a technique. It is whether you already know, roughly, what is in the book.
Three words and an arrow, because you know where the rest lives and what it says. If you have no idea what is in the book, you cannot do that — every sentence is equally likely to be the important one, so the only safe strategy available to you is to write down as much as you can. Which is the too many row above, arrived at rationally.
That is the causal chain worth seeing whole, because each link explains the next. No Step 0, so nothing can be discriminated; nothing can be discriminated, so everything gets captured; everything gets captured, so the attention goes to the words rather than the meaning; and the hour ends with a full page and no understanding. The fifth state in the table is usually the first state in disguise.
Which makes preparation the real instrument for calibrating note density, and it works in proportion: having skimmed the headings lets you skip the definitions, having read properly lets you skip most of what is on the slides and spend the hour on the connections, the emphases, and what the instructor said that the book did not.
Two different activities get called “organization,” and blending them is why advice about staying organized grows into a nine-component system nobody adopts.
Not a system. One cheap notebook — and small is better than large, because a small
one goes everywhere and a large one gets left behind. No categories, no color scheme, no
filing. Every due date from every syllabus goes in it at the start of the semester, and
everything announced in class goes in it as it happens. A week later it will have caught
something you would otherwise have lost, and that is a promise you can check rather than
take on faith.
Once one sheet of paper has proven useful, then the harder claim — that writing is
where thinking happens — is a much easier thing to hear.
For what it is worth, I have dispensed with the paper version entirely and use a notes app on my phone plus a calendar, also on my phone. What matters is not the object but that there is exactly one place things go and that it is always with you. A phone qualifies on both counts. It is also, as Step 1 notes, the single most effective distraction ever built, so the tradeoff is real.
The Cornell method is worth knowing, and it is worth knowing properly, because the popular version has had its engine removed. The layout is just two lines on a page:
During the lecture you write in the right column, and put questions in the left one as they occur to you. Everything else about Cornell happens after — which is the part that gets dropped, and which is covered in Step 3. If the left column gets filled in live with headings and vocabulary, and the bottom strip never gets filled at all, what is left is a transcript with a margin. The geometry survives; the mechanism does not.
About color coding: if you enjoy it, by all means. It is worth knowing only that color which encodes a decision you have already made is a retrieval aid, while color applied during capture mostly makes a page look processed — and it costs attention at the moment attention is scarcest. So it earns its keep better in Step 3 than here. Not a prohibition, just an observation about where the return is.
This is the step that connects taking in information to learning information. It suffers though from a name problem: called “organizing your notes,” it sounds like filing. It is not filing. It is deciding what mattered, working out what connects to what, and finding out what you did not follow — and all three of those are thinking.
What you wrote during a lecture is a pointer to something that was in your head at the time. Come back two hours later and some of those pointers no longer resolve — you can read your own handwriting and not know what it was about. That portion of your notes is now lost. Step 3 cashes the pointers in before they expire.
Sit down with what you wrote and flesh it out. Finish the half-sentences. Write out the thing the arrow was pointing at. Put a question next to what you did not follow. The difference this makes is between notes you can use in three weeks and notes you cannot, and it is largest when done soon — while the pointers still resolve.
The version of this I first heard is Marty Lobdell’s, a psychology professor who taught a college study-skills course for years: take the first moment you get after a class, five minutes, and expand what you jotted down — because wait until you get home and you will already have forgotten what some of your own notes meant, and a note you cannot reconstruct is a wasted note.
The principle is right. The five minutes is worth some qualification, though, because it is not quite the panacea a short video makes it sound.
There is a tradeoff hiding in the number, and it runs in an awkward direction. If your notes came out reasonably clean — because you had read ahead, or because the lecture was well organized — then five minutes is genuinely enough, and it also moves the bar the least, since there was not much to repair. If your notes are a mess, five minutes will not come close, and that is precisely the case where doing the work moves the bar furthest.
So the honest version is that the messier the notes, the more this is worth and the longer it takes. Budget by how the hour actually went rather than by a fixed number.
Some practical realities the advice tends to skip:
One reason to do it sooner that has nothing to do with memory: Step 3 is the first point in the sequence where nothing external makes you begin. The usual reason it does not happen is not that students disagree with it — it is that anything requiring initiation loses to anything that does not. The closer you do it to the class that just ended, the less initiating it requires.
And if a note will not reconstruct, you are standing in a room full of people who were just in the same lecture. Ask whoever is still packing up: what was that bit about? That is a thirty-second fix now and an unanswerable question next month.
Expand the shorthand while you still know what it meant. This is the mechanical half and it is the half that actually decays.
In the left column of a Cornell page, or the margin of any page, write the questions the notes answer. Not topic labels — questions. Why does the curve flatten? rather than Growth curves. The difference matters because a question can be answered from memory later and a label cannot. The left column is not decoration; it is a built-in self-test, and it is the entire reason the layout exists. Cover the right side, read the question, answer it out loud, then uncover and check.
What were the big ideas of that hour? Compressing a lecture into a sentence forces a decision about what it was for, and the compression is where the understanding shows up or fails to. This is the step the popular version of Cornell leaves out, and it is the most valuable part.
Arrows between things that turned out to be the same idea. A note that this contradicts what was said two weeks ago. A line to the figure in the textbook. Connections are not decoration either — a fact with no attachments is a fact you will lose, and connecting is the work that makes material stick.
If a process was described, draw it. Drawing beats writing for retention, and the reason is that you cannot draw a thing without committing to what it looks like, which is a decision you can dodge indefinitely in prose.
If a linear page is fighting you, a mind map is worth trying: the concept in the middle, branches outward to the things that hang off it, sub-branches from those. It looks less like a page of text and more like a tree, and it makes relationships visible in a way that an indented list does not.
The dashed line matters as much as the solid ones. A mind map has room for the connection you noticed between two branches that were taught weeks apart, and noticing those is a large part of what Step 3 is for.
And notice why this cannot be done during a lecture. You can only place something on a map once you know what it relates to — which you do not, while it is still arriving. Drawing one is therefore a Step 3 activity almost by definition, and the drawing is the thinking rather than a record of it.
Everything above is about the content of your notes — what is missing, what connects, what you did not follow. There is a second job happening in this step, and it is easy to dismiss as cosmetic when it is not.
My own reason for rewriting notes was unglamorous: my handwriting is not especially pleasing to look at and not especially easy to read, and it gets considerably worse when I am writing fast to get everything down. Which meant that what came out of a lecture was not something I could comfortably study from — not because the content was wrong, but because sitting with it for two hours was unpleasant enough that I would find something else to do.
Generalized: Step 3 is also where you turn your notes into whatever you personally need them to be in order to spend hours with them. Legible. Laid out so you can find things. Broken into pieces you can work through. Converted into a form that suits how you actually study — a term list, a set of cards, a diagram, a sheet of practice problems, a deck in whatever tool you use. None of that is the thinking, and all of it determines whether the thinking ever happens.
Here is the practical payoff, and it is one of the more useful scheduling facts on this page. Rewriting, reformatting, building a term list, making cards — this is real work that genuinely needs doing, and it does not demand the deep, sustained concentration that Step 4 does. So it is available on the evenings when Step 4 is not.
That matters, because the realistic alternative when you are depleted is not deep work. It is nothing, or something worthless. Matching the task to the capacity you actually have beats both forcing the hard thing and abandoning the day, and low-intensity productive work is a legitimate move rather than a way of avoiding the real one.
This page is fairly hard on rereading, and it is careful not to call Step 3 studying. Both of those are right, and both can be taken too far. So, in fairness to the step:
Interacting with your notes is time spent with the material that would not otherwise have been spent. Rewriting, reorganizing, turning a lecture into a term list or a diagram or a set of cards — none of it is retrieval, but all of it involves reading, handling, and making decisions about content you are going to need. An hour of that is not worth an hour of self-testing. It is worth considerably more than an hour of not thinking about the course at all, which is the realistic alternative.
Here is the part that is easy to miss. The steps are separated in time by construction:
That is four separate encounters with the same material on four different occasions. Which is to say the sequence produces spacing on its own, without anyone planning a schedule — and spacing is among the most reliably useful things you can do with a fixed number of hours. It is discussed in Step 4 because that is where it is usually deliberate, but it is running through the whole process whether or not you notice.
The practical consequence is worth stating plainly. Even in a stretch where you are not studying in any Step 4 sense — a week where the exam is far off and nothing is due — doing Steps 0, 2, and 3 means you are still meeting the material repeatedly, at intervals, rather than attending a lecture and then leaving it alone for three weeks. That is most of the difference between material that is merely familiar by exam week and material you have to meet almost from scratch.
The caution stands, though, and it is the one from the previous section. Exposure is worth something; it is not worth what production is worth. A month of recopying with no retrieval in it will leave you with beautiful notes and a poor exam, and the fact that recopying beat doing nothing is not much consolation at that point.
If a recording or transcript of the lecture exists (see Step 2 for the conditions), there is one thing to do with it that is worth more than all the others, and it follows directly from the problem above.
Your notes are the survivors. Whatever you missed left no trace in them, because missing it is what kept it off the page — which means no amount of rereading your own notes will reveal it. A transcript is the one artifact that contains the material your notes do not, so comparing the two is the only cheap way to see the shots that did not come home.
Practically, this is a job an assistant does well, because it is a mechanical comparison over more text than you want to read: hand it the transcript and your notes, and ask what the lecture covered that your notes do not, or where your version of something differs from what was actually said. Ask for a list of differences rather than a summary of the lecture. The summary you do not need; the differences are the entire point.
Then do the work yourself. Each difference is a decision: did I miss this, or did I leave it out deliberately because it did not matter? Both answers are useful and only you can make the call — the assistant does not know what your course is responsible for, and this is exactly the judgment that Step 3 is.
The same material feeds Step 4: once you have a transcript and a corrected set of notes, asking for practice questions drawn from what was actually said is a good deal closer to the exam than questions drawn from the textbook.
The sequence is not a line, and Step 3 is where most of the traffic turns around. Students run through once and stop, which is why studying can feel finished long before it is.
3 → 2 Something in the organized set does not make sense; go back to
the raw notes.3 → 0 The raw notes do not have it either; go back to the textbook.4 → 3 Studying turns up a hole; the organized set needs fixing.4 → 2 The organized set does not contain it at all; the raw notes
might.
Note what this means about Step 4. Finding a hole while studying is not a sign that studying failed. It is what studying is for, and the correct response is to go back a step rather than to reread harder.
Eventually the loop exhausts what you have. The notes do not have it, the book does not have it or does not make it clear, and now you go to something outside them. Before you do, it is worth knowing which of four situations you are actually in, because they cost different amounts and they teach different amounts.
| Where you are | What to do about it |
|---|---|
| In your head you believe you can do it unaided |
Verify that. Running through something mentally and producing it on paper are different acts, and the gap between them opens as soon as the problem is one step longer than the one you rehearsed. If you have not written it out, assume you are on the row below. |
| On paper you can do it with a pencil and some time |
Good. This is the honest version of knowing something, and it is the standard an exam will hold you to. |
| Workable you can get there yourself, with help |
Do that. The textbook counts here, and so does a search, and so does an AI assistant — provided you ask for a missing piece rather than for the finished result. More on that distinction below. |
| Needs a person you cannot get there alone |
A classmate first, because it is fast and cheap. Then the instructor. Either way, bring a specific question rather than a general difficulty. |
These four are about where the answer has to come from, and deliberately have no numbers attached — the numbered steps on this page are a different scheme entirely, and two sets of numbers meaning two different things would be one set too many.
Looking something up sits in the workable row, not the needs-a-person row, and that is not a technicality. When you look something up, you decide what is missing, where to find it, and whether what you found answers the question. That is the whole skill. My own calculus instructor’s version, forty years ago, was “if you can’t figure the problem out, go to the library” — which sounds quaint and is not, because the library was only where the corpus lived. The internet is the same corpus with the walk removed.
What decides which row you are in is not the source but the request. The clearest case is a problem set:
Dale Corson, a physicist who was dean of engineering and later president of Cornell, gave students advice that sounds obvious and is not. Before going to the professor, ask yourself: what is it, exactly, that I do not understand? Most students arrive with a sweep of the hand and “I don’t get this,” which is unanswerable. Instead:
A surprising fraction of the time the problem dissolves during those fifteen minutes, because locating the confusion is most of resolving it. Programmers have a version of this: explain the broken code aloud to a rubber duck on the desk, and the explanation frequently reveals the bug. When it does not dissolve, you arrive with a question that takes two minutes to answer instead of twenty, and instructors are genuinely glad to get those.
Richard Feynman kept a notebook titled “Notebook of Things I Don’t Know About.” Not a record of failure — a map. Every confusion, every half-formed question, every concept that did not click, written down as a question rather than an answer, and returned to.
The same shape turns up wherever anything gets made: decide, learn, do, and then revise — because a first draft is not a final draft, and revising is where you find out what is actually worth keeping. Step 2 produces the draft. This step is the revision, and in both domains it is the half that gets skipped.
Most people keep the opposite notebook: neat summaries of what they already know, which is satisfying, looks productive, and teaches nothing. Knowledge you already own does not need protecting. The half-grasped, slightly embarrassing material is where the building happens.
The practical version costs nothing. Keep a running list of things you did not follow. Once a week, take one, and work on it until you could explain it to someone who does not know the subject. Over a semester those scribbled confusions become the most solid things you know, because they are the only ones you had to fight for.
There is a well-known piece of wartime statistics that makes the underlying problem vivid. In 1943 the Statistical Research Group at Columbia was asked where to add armor to bombers. The available data was a record of where returning aircraft had been hit, and the obvious reading was to armor the areas covered in holes.
The mathematician Abraham Wald, working for the group, argued the reverse. The data came only from planes that made it home. A bomber that returned with its wings full of holes was evidence that wings could be hit and the plane would still fly. The regions with no holes in the sample were not the safe ones — they were the ones where a hit meant the plane never came back to be counted. Armor the emptiness.
Worth being accurate about the story, since it is usually told as a single clever remark. Wald’s work was a series of eight memoranda amounting to a long and fairly dense piece of mathematics for estimating how vulnerable each part of an aircraft was, given only the damage recorded on the survivors; the application to armor placement is his own suggestion, made in his comments on the memos. It stayed classified until the Center for Naval Analyses reprinted it in 1980. The familiar diagram of a bomber covered in dots is a much later illustration rather than anything Wald drew, and the phenomenon now carries a name he did not use for it, survivorship bias.
Your material is in order — you have been to the lectures, your notes exist, and Step 3 has turned them into something you can work with. Now it is time to actually study.
There are only a few things here that are strongly supported, and they are all versions of one instruction: produce, then check. Everything that feels like studying and is not — rereading, highlighting, recopying, watching the lecture again — shares the property that too little is demanded of you.
Most students prefer the first, and it is worth being clear that this is not stupidity. The first one returns a steady stream of positive signals and the second one returns failures. Every available feeling says the wrong thing.
† The weights image is not mine. It comes from a short video by the Explanatory channel on active recall, which puts it better than anything else I have seen: rereading is watching someone lift weights, and active recall is lifting them yourself. One makes you feel productive; the other makes you stronger.
Retrieval practice. Flashcards, self-quizzing, closing the book and writing out what you remember, answering the questions at the end of the chapter without looking. Retrieval changes the memory rather than merely visiting it. It also, unlike rereading, tells you the truth about what you actually know.
And when you cannot remember the answer — that attempt still helped. Racking your brain activates related material, so when the correct answer does appear, it lands somewhere prepared. This matters practically, because the usual reason people give up on self-testing is that getting things wrong feels like evidence they are not ready, which can be genuinely unpleasant. It is, however, the mechanism working. Better to find out what you do not know before the test, while you can still do something about it, than during — which is also an argument against leaving all of your studying until the last minute, since by then finding out is all you can do.
One refinement, from Dennis Shaffer. Testing yourself is also how you find out what you know, and as a measurement it is only accurate if you wait. Quiz yourself the moment you close the book and you are checking short-term memory: the material is still warm, and it will come back whether or not it has been learned. The exam will be testing something else.
His prescription is to let at least an hour pass without thinking about the material, with no notes or book in front of you, and then test yourself, find the weak links, and spend your time on those. Judgments of how well you know something are considerably more accurate when they are delayed (delayed judgment of learning). It is the fluency illusion again, in a form that is easy to miss because self-testing is supposed to be the cure for it.
It also fills a gap in the schedule. The last night before an exam is for proving to yourself that you have the material, as described further down this tab. Delayed checks are what belong between studying and that night: each one finds weak links while there is still time to repair them, so that by the last night confirming is all that is left to do.
This is about trusting the result, not about whether to recall at all. Recalling right away is still practice, which is why review before you stand up appears further down this tab. It just should not be what tells you that you are done.
Spacing. The same total hours, spread across days, beat the same hours in one sitting — by a lot, and reliably. Twenty minutes today, twenty tomorrow, twenty three days from now. Sleep in between appears to matter, which is one of the few solid arguments for not staying up. Note also that the sequence hands you some of this for free: reading, lecture, and the pass through your notes are already three encounters on three different days, as Step 3 discusses.
Expect it not to feel that way. Shaffer makes the point that even in studies where spaced practice clearly wins, students still feel that they did better when they massed their practice — crammed. Massing produces immediate, visible gains and hides the rapid forgetting that follows; spacing feels slower while it is working. The feeling and the result point in opposite directions, which is the theme of this whole page.
Forgetting, usefully, is not a loss. Think of it as a library where the books never leave the shelves but the catalog gets disorganized: what fades is your path to the memory, not the memory. That way of putting it comes from Robert Bjork’s work on storage versus retrieval strength, by way of Benedict Carey’s How We Learn. Every retrieval reorganizes the catalog. And the harder the retrieval was — the closer you came to not finding it — the more the attempt was worth. There is a catch, which is that if you wait too long you get nothing back at all. The target is roughly the point just before you would have forgotten.
Everything above is about memory, and there is a second argument for spacing that has nothing to do with memory and is felt far more often. Call it restart cost.
Come back to something after three days and you open it and start working. Come back after three weeks and the material may still be largely there, but the context is not: where you were, what you had already decided, what the next move was supposed to be, what state your notes are in. Rebuilding that can take an hour before any actual work begins, and it is genuinely unpleasant — not difficult so much as disorienting, and disorientation is what makes a task feel like something to avoid.
Which sets up a loop worth recognizing, because from inside it feels like evidence about you rather than about the gap. The longer you leave something, the more expensive restarting becomes; the more expensive restarting becomes, the easier it is to put off; and putting it off lengthens the gap. Nobody decides to abandon a project. They decline to pay the restart cost, repeatedly, until it has been abandoned.
This is the practical case for spacing that a student can actually feel. Retention is invisible until an exam tells you about it. Restart cost is charged every single time you sit down. Familiarity is the reward for having stayed in touch, and it is paid out in not having to spend an hour finding your place.
Two notes from having done this badly. The reward that actually works here is often not a treat but the removal of an anticipated cost — let me get some of this done today so that I do not have to worry about it tomorrow is a stronger motivator than anything you could give yourself afterward. And reconstruction is real work even though it produces nothing: reaching the point where you are ready to begin is a legitimate place to stop for a while.
Interleaving. Studying one topic until it is solid and then moving to the next feels orderly and performs worse than mixing topics within a session. Cycling forces your brain to let go and retrieve again, which is the same mechanism as testing, and it also surfaces the differences between things that looked similar when studied separately.
Shaffer puts the case for it most usefully: practicing one kind of problem in a block teaches you how to use a method, but not when. His example is twenty practice problems that all need the Pythagorean theorem. You get good at the theorem, and you never once have to decide that the theorem is what you need. Exams ask for both. The biology version is twenty Punnett squares in a row, which teach you to draw a Punnett square and do nothing for recognizing which question calls for one.
He adds a detail that is easy to skip and should not be. If you are practicing spotting errors — what is wrong with this figure, this argument, this sentence — include some examples with nothing wrong in them. Otherwise you are also practicing the assumption that there is always something to find, and an exam will not grant you that.
Unlike the other two, though, this one need not feel worse while you are doing it. What gets judged as more effective afterward is usually the blocked version, one topic until it is finished — but moment to moment, cycling is frequently the more pleasant way to spend an hour, because monotony is its own kind of difficulty.
Think of it less as making things harder and more as alternate-tasking. Rather than stopping for five minutes, you switch to something different enough that the switch itself is the rest. It is the same logic as working with music on, or as spending a morning on a paper and an afternoon on problem sets: the things you alternate between should draw on different enough resources that the change refreshes rather than collides. Genetics for half an hour and then metabolism is a change. Genetics and a group chat is not.
Marty Lobdell’s distinction, and the most useful thing on this page for a biology course. Ask it before choosing a method, because it decides which method is right and biology is full of both.
Memorize the name of a bone without understanding what it does and you have done a lot of work for very little. Concepts, once grasped, stay; and in most college courses the concept is what the exam is actually reaching for. Terms Challenge and any flashcard system handle the fact half well and cannot do the concept half at all. That is not a criticism of flashcards; it is a statement of what they are for.
Biology courses hand you long vocabulary lists, and there is a specific wrong way to learn them that works well enough to be dangerous. Most students take the stack of cards, look at the word, say the definition from memory, turn the card over to confirm that it matches, and put it on the done pile.
That is a real skill and it is worth having. The trouble is only that it is the second rung of four, and the card going onto the done pile is what stops you climbing.
On this point I give memorizing more credit than some. Shaffer advises against spending much time memorizing definitions at all unless a course requires it, because a definition learned by itself is an isolated fact, connected to nothing. We agree about the failure, which is stopping there. Where we differ is that I would not skip the rung: having the words exactly is what the understanding then gets built on.
The specific trouble with stopping at the second rung is that it trains you to tell the definitions apart from one another, which is a different skill from knowing what they mean. It holds up right until the exam words things differently — and instructors reword things deliberately, because that is how you find out whether someone understood.
Shaffer makes the same point about question banks. Drilling the questions in a set such as Quizlet makes you familiar with how those questions are worded, without adding much understanding of the material, and you are led astray when the same content arrives in different words. Flashcards used only to identify a term, he notes, are worse than they look, because they build familiarity without learning, and you come away believing you know it.
The four steps, which work on any definition:
The instrument that enforces this — and the whole four-step method above — comes from the Long Beach City College study skills workshop cited earlier. It is a single word: and? Recite the definition, and someone says and? If the answer is “and what?”, you are not ready. If you can say it differently, and then give an instance, you are as ready as you are going to get, and it will not matter how the question is worded.
Psychologists have a name for the difference running through this tab. Shallow processing deals with how information looks or sounds — repeating it, recopying it, highlighting it, recognizing it on a card. Deep processing deals with what it means: where it fits, what it connects to, what it is an example of. The framework goes back to Fergus Craik and Robert Lockhart’s levels of processing model of 1972, and the account here follows Shaffer’s talk, which is built on it (shallow and deep processing).
Shaffer’s list of shallow processing reads like an inventory of what most students actually do: repetition, rewriting notes without connecting anything, highlighting, skimming, flashcards used only to identify terms. He also notes that textbooks invite it, by putting a term in bold or in color without placing it in anything larger. Deep processing, in his account, takes a few forms besides the producing that the next section covers.
Elaborating means relating what you are studying to other things you have learned. Think of applications and examples. Ask where it fits, and how you would categorize it within the rest of the course. The last of the four definition steps above does this on a small scale, and a mind map in Step 3 does it on a large one.
The self-reference effect: people remember information better when they relate it to themselves. Connect what you are studying to something in your own life, a friend’s or a family member’s, something you have seen, or something from another course. In biology this is unusually easy, because the subject is in large part about you — the metabolism is yours, the immune system is yours, and the infection you had last winter was a real instance of the chapter you are reading.
Shaffer treats interest as something you do rather than something the material either has or lacks. Set the material up as a challenge you intend to beat. Give it a story. Look for the deeper connection, which is usually where the interest was hiding. The same idea appears in Step 1 as Feynman’s habit of carrying real questions around: interest is mostly a matter of having somewhere for new material to attach.
From Crash Course Study Skills. Take a single index card — or a page of the small notebook you are already
carrying — and put the most important material for the exam on it, in handwriting
as small as you like. Most instructors will not let you bring it. Make it
anyway. The card is small, so you are forced to choose — and choosing requires
deciding what matters, which is exactly the judgment the exam will test. An hour spent
selecting and writing is an hour of concentrated, active engagement with the material,
disguised as a craft project.
The card is also portable, which is most of its second life. It fits in a pocket, so the
ten minutes before a bus, between classes, or standing in a queue become usable in a way
they otherwise are not — and those minutes are naturally spaced from one another,
which is the schedule you would have designed anyway.
Toward the end of his talk, Shaffer gives a list of warning signs that the way you are studying, rather than the amount, is the problem. They are worth having in one place, because each is something you can notice about yourself without waiting for a grade.
One rule covers every case: use it to be questioned about what you have produced, not to produce things for you. An answer that helps you understand something is fine. An answer that is the deliverable is not — and the reason is worth spelling out. A set of notes has no value sitting on a desk. Its entire value was in the making of it, which is what changed you. Get a machine to make it and the object exists, you are exactly as you were, and you now own a study guide that studies nobody.
The test to apply to anything you are about to ask for: will I end up having said the thing, or having been handed it?
A practical note on generating practice questions: unless you specify otherwise, you will get back a stack of definition prompts, which reinforces exactly the recognition strategy you are trying to escape. Ask for questions that require applying the concept.
A lecture starts whether or not you feel like it. Studying does not, and that difference matters more than it looks. Everything here has to be initiated by you, unsupervised, against no external clock — which is the same situation a writer faces in front of a blank page, and it responds to the same things.
Which is why framing this as motivation makes it useless. Nobody can decide to want something more. It is a starting problem, and starting problems have a known fix that is not willpower: make the first mark cost almost nothing. Notice that every technique below is a subtraction or a cost reduction, and none of them adds discipline.
None of the four below is mine; they circulate widely and are collected here because they work. Sources given where I can name them.
Procrastination is not a time-management problem. It is emotion regulation. You avoid the task because contemplating it feels bad, and the avoidance works — the relief is immediate and real. Which is exactly why the habit sticks, and why knowing better does not dissolve it.
And one correction to a common assumption: procrastinators are frequently not the people who care least. High fear of failure is one of the strongest predictors. People put things off because they are afraid the work will not be good enough, which means the behavior that looks like indifference is often the opposite.
Boredom while studying usually gets read as a fact about the subject, or about yourself. Neither is likely. It is nearly always a mismatch between the difficulty of what you are doing and what you can currently do — too easy and attention wanders, too hard and you quit.
Which explains something about method. Passive rereading is unbearable because nothing is being demanded of you: nothing to succeed at, nothing to fail at, no feedback of any kind. Self-testing is many things but it is rarely boring, precisely because you can miss. Every technique on this page that works happens to produce the conditions that hold attention — a clear goal, a bounded unit, and immediate information about whether you got it.
The same logic runs the other way. If the demand is far too high — a chapter resting on a prerequisite you never had — more effort will not fix that either, and it is a Step 0 problem wearing the costume of a motivation problem.
Setting an hour after which you do not study sounds like a concession and is closer to the opposite. It does two things at once. It creates a real end, which open-ended work never has — and open-ended work absorbs everything while returning progressively less, which is why eleven o’clock work is rarely any good. And it runs backwards through the day: knowing you stop at six is what makes twenty more minutes at half past five worth doing. A day with no deadline until the real deadline has no urgency at all until everything is urgent at once.
I have a private name for the other case: the earning tenure moment. It is the point in an evening where you can stop, or you can push a little further and either finish the thing or set the next session up. And it genuinely mattered in my career, because fifteen more minutes at the right moment frequently bought an entire extra day of research — not by adding fifteen minutes of output, but by moving past a threshold, so that tomorrow began somewhere useful instead of beginning with yesterday’s unfinished business.
That plainly contradicts the paragraph above, and I do not think the contradiction resolves into a rule. What it resolves into is a distinction: the fixed time is about when you stop, and the tenure moment is about where. Those are different questions, and the second one is the one almost nobody asks.
Stopping mid-thought, mid-problem, mid-paragraph is expensive, because tomorrow begins with reconstructing what you were doing rather than doing it. Stopping at a clean boundary — the chapter finished, the problem set through number six, the argument written down even if badly — is cheap. Same clock time, very different restart cost.
One or two sentences. Where you got to, what you were about to do next, and anything you
had in your head that is not on the page yet. This is the cheapest thing on this page and
close to the highest-returning, because it converts an expensive restart into a cheap one
without requiring you to have stopped at a good place.
It is also the same principle as everything else here. You are not going to remember
where you were. Write it down.
There is a harder version of this too, and it applies to anything without a deadline attached. Sometimes the question is not whether tomorrow will be expensive but whether there will be a tomorrow for this particular piece of work — because if you stop now, you may have no idea when, or whether, you will want to come back to it at all. Stopping can be permanent without anyone deciding that it is. When that is the situation, pushing on is not stubbornness; it is recognizing that the alternative is probably not “later.”
The thing that makes people push on in that situation is usually not discipline either. It is caring how the thing turns out. Wanting it to be good is what gets you through the part where you are tired of it, which is worth knowing because it means the useful question when you stall is often do I actually care about this? rather than why am I so lazy?
The honest caution is that the tenure moment can be wrong. Pushing on past the point of usefulness burns the next day as surely as a bad stopping point does, and then the extra day you bought gets spent recovering from having bought it. There is no formula. But if you are going to overrun, overrun to a boundary rather than simply until you cannot continue — and if you are going to stop on time, spend the last two minutes on the note.
You do not know what a course demands until it has tested you once. Syllabi do not say, other students guess, and the same instructor can set very different exams in different courses. Which means the first exam of a semester is being sat blind, and the two ways of being wrong about it are not symmetrical.
Overpreparing costs you hours. Underpreparing costs you a grade you may not be able to recover — and, worse, it teaches you almost nothing, because you learn that what you did was not enough without learning what would have been. So study too hard for the first exam, find out what the course actually wants, and taper from there on evidence. Going the other way round, starting light and correcting upward, means paying for the calibration with the one result that is hardest to fix.
Not learning anything new, and not rereading. What I used the night before an exam for was proving to myself that I had it — going through everything once, out loud or on paper, checking that I could produce each piece from memory and produce it quickly. Speed matters here, because an answer you can reach in three seconds and an answer you can reach in forty are not the same answer once a clock is running.
Then, having been all the way through, spend whatever remains on the items that did not come back cleanly. If you have been running boxes, this is the same instruction in different words: sweep the whole set once, then work box one.
Which is worth stating plainly: the night before is not the time to find out that you do not know the material. If that is where you find it out, the most you can do is cram. Finding out belongs earlier, in the delayed self-checks spread across the days and weeks before, so that the last night holds no surprises.
Note what this presupposes. A last night spent confirming that you know the material only works if the learning already happened. The same evening, spent trying to learn it for the first time, is the version everyone means by cramming, and it is a different and much worse activity that happens to look similar from the outside.
The list of credentials at the top of this page could be read as evidence that some people are simply built for this. I would rather be honest about where mine came from, because the honest version is more useful.
I took a year off between high school and college. I skied, I hiked part of the Appalachian Trail, and by the end of the summer I was bored out of my mind. I arrived at university more mature than I had been a year earlier and genuinely hungry for something to do, and that combination — not discipline, and not talent — is what powered close to four years of high-intensity studying.
It did not last. By the time I started graduate school the same enthusiasm was not there, and it showed in how I performed on exams. Some of that was the material getting harder and less familiar — theoretical population genetics, I am looking at you — but a good deal of it was simply that the fuel had run out. Which is the part worth saying out loud: the intensity was a phase, it had a cause, and the cause was not virtue.
It is a marathon rather than a sprint, but marathons end, and this one ends sooner than it feels like it will from inside the second year. What you build during it — how you read, how you take something apart, what you do when you are stuck — stays with you for the rest of your life. The part where studying is the main activity of your days is a brief stage of it.
Two things follow. The first is permission: the goal is not to sustain maximum effort indefinitely, because nobody does. The goal is to arrange the rest of your life so that it supports putting your best in during the hours that are actually for studying — which is what sleep, real time off, and stopping at a fixed hour are all in service of.
The second is a reason to take it seriously while it lasts. A finite thing is worth spending properly. It is also worth not wrecking yourself over, since running out of fuel in year one is a poor trade for a degree that takes four.
An exam is one pass, on the clock, with no editing available. This step is about the present tense only — what to do in the room, while the clock is running. What to do afterward, with a paper full of red marks, is Step 6.
An exam is unforgiving in a specific way. A paper can be drafted badly and fixed over a week; there is no equivalent here, which is why almost everything that makes an exam survivable happened weeks earlier. What is left on the day is real but limited, and it is worth knowing exactly what it consists of.
Most test-taking advice goes straight to question-level tactics and skips the budget, which is backwards. The single most expensive mistake available is grinding on one hard question while answerable ones sit unread later in the paper — and a hard question is usually worth the same as an easy one.
Leaving feels like being finished, and sitting with a half-blank question feels like
failing. But leaving spends the one resource still generating answers for free. Sitting
quietly and not forcing it is doing something, even though it looks like doing nothing
— which is the exact inverse of the problem everywhere else on this page, where the
thing that looks productive is not. Doodle. Sing something in your head. Let your
attention drift off the paper for a minute and come back.
This is also why getting some sense of a hard question matters before you move
on: you cannot have a sudden insight about a question you never engaged with.
More points are lost here than to not knowing the material, and every fix is a few seconds long.
That last move is the same one the Deconstruction laboratory teaches for word problems, which is not a coincidence: an exam question is a word problem with distractors attached.
“Check your work” mostly fails, because rereading your own reasoning confirms itself — you cannot catch a wrong interpretation by inspecting the steps that followed from it. A second pass only counts if it comes in by a different route:
Coming back to a question ten minutes later supplies the change of perspective for free, which is one more argument for the two-pass structure above.
Partial knowledge is the normal condition. Nobody walks in knowing everything, and the useful question is what partial knowledge is worth.
Elimination arithmetic. On a four-option question, guessing blind gives you 25 percent. Eliminate one option and you are at 33. Eliminate two and you are at 50. That is worth internalizing, because it reframes the task: you are not trying to know the answer, you are trying to reduce the field, and reduction is done with understanding.
Well-written wrong answers are wrong in ways you can detect if you know a little. That is what they are for. The amino acid case is exactly this: nitrogen eliminates an option without your having to recall the whole definition.
If the options are numerical and you cannot do the calculation, a rough estimate will often throw out the wildly high and wildly low ones. Knowing roughly what a quantity should be is itself knowledge, and this is where it pays.
A good question to ask yourself on every item you are unsure about. It sounds extreme and it works, because it converts a vague preference into a commitment — and the gap between this one looks right and I would stake something on this is precisely the gap between recognition and knowing.
Staring longer does not make a blind guess more accurate; it makes second-guessing more likely and spends time that answerable questions need.
Test formats are not arranged from easy to hard. They are arranged by how much the format gives you — and that is what determines how you should have studied. It is the same ladder as recognize, remember, understand, apply, seen from the examiner’s side.
| Format | What it asks of you |
|---|---|
| True / false | Recognition, with one bit of information. Beware of absolutes: always, never, all. |
| Matching | Recognition with the whole answer set visible. Do the ones you are sure of first; each one you remove makes the rest easier. |
| Multiple choice | Recognition, with plausible wrong answers built to catch partial understanding. Try to answer before reading the options. |
| Fill in the blank | Recall. Nothing is handed to you, which is the important part. Telling definitions apart is still useful here — ruling candidates out is often how you land on the right term — but it only works once you have candidates, and here you have to generate them yourself. A strategy built on the differences between definitions, rather than on the definitions, runs out at exactly this point. |
| Short answer | Recall plus production. Your own words, which means the definition drill in Step 4 is what prepares you. |
| Essay / problem | Production and structure. Show the work; write the steps even when you cannot finish, because a graded step is worth something and a blank is worth nothing. |
| Open book | Off to the side of the ladder: hands you everything and demands the most. See below. |
Writing down everything you know is a reasonable tactic where partial credit exists. It becomes actively dangerous when the grader cannot tell which part is your answer. Speaking as one: if the correct answer is not indicated as the answer, and both correct and incorrect statements are sitting on the page, I generally count the incorrect ones against the correct ones. Overdelivering with no answer marked is not a safe bet. And none of this applies to multiple choice, where there is no partial credit and no pity.
These are frequently the hardest kind of exam, and they are the clearest demonstration in existence that having the material is not the same as knowing it. Everything is right there, and it does not help.
The reason is design. Instructors write these questions knowing you have the material, so they ask you to apply rather than to locate, and there is not enough time to read anything you should already have read.
“Open book” gets used loosely for three quite different arrangements, and they reward different preparation. Ask, because guessing wrong is expensive.
| Kind | What it actually rewards |
|---|---|
| Open text the textbook, and perhaps only the textbook |
Knowing the book — its structure, where things live, what is in which chapter. Marking and tabbing pay off directly here, and so does having read it, since a book you have never opened is a very large haystack. Your own notes may not be allowed at all. |
| Open notes your own materials, sometimes without the text |
This one is graded on Step 3, whether or not anybody says so. The resource you are permitted is the artifact you built, so its quality is your exam-day capability. Contents pages, numbered pages, one-line chapter summaries on the tabs — the work that felt like tidying turns out to have been the preparation. |
| One sheet a single page or card you prepare |
The cheat card exercise, made official. The selection is the studying, and it is nearly impossible to make a good one without having understood the material. Which is presumably why instructors set it. |
The advice below applies across all three, and comes from people on both sides of the desk:
The goal, well put by an accounting academic who writes these exams: with the material understood and properly referenced, the exam becomes a matter of going to the right place to confirm what you already think. Not to find out.
Some anxiety before something that matters is normal and is not a defect — it is a signal that the thing matters, and the alternative is apathy. Henry Fonda was still throwing up before stage performances at seventy-five, and still walked out and performed. The goal can be less to eliminate it — though that certainly helps where it is possible — than to be able to function despite it.
Too much of it does real damage, and the mechanism is specific: worry consumes working memory — the same limited resource the question needs. Which is why it does the most harm on exactly the hardest items, and why a student can feel fine on recall questions, fall apart on the application ones, and conclude they did not know the material. They may have known it perfectly well.
Reported by Gerardo Ramirez and Sian Beilock, and discussed in Crash Course Study Skills. Students given about ten minutes to write out their fears and worries immediately before an exam scored meaningfully better than controls — roughly on the order of a grade point. The mechanism is the same one behind every other recommendation here: the page holds what your head cannot, and the freed capacity goes to the test. “Just don’t worry about it” does not work, because your brain thinks the worries are important. Putting them somewhere it trusts does.
Some other things that help, briefly. Preparation that resembles the actual test is the single best anxiety treatment there is, because by the time you sit down it is not really the first time. Reframing the exam as a game with just enough challenge is a real move, not wishful thinking — you are re-reading the same sensation, not pretending it is absent. And bringing something to fall back on is sensible even when you know the structure by heart: knowing something and being able to produce it under pressure are different, and a blank mind is a thing that happens.
If anxiety is affecting your life beyond exams, that is worth taking to someone qualified, and there is nothing on this page that substitutes for it.
The other six steps are about getting ready and performing. This one is about the only thing that makes any of them improve: looking back afterward and changing something.
It is worth separating from the Diagnostic, which asks a different question. The Diagnostic asks where was this particular question lost? — one item, about content, pointing back at a step. Step 6 asks what should I do differently? — about your methods taken as a whole, pointing forward at the next round. The first is the instrument. The second is what you do with what it tells you.
Steps 2 and 3 already send you back and forth, and Step 4 keeps returning you to Step 3. Step 6 closes the outermost loop: what you learn here goes into Step 0 for the next unit, the next exam, and the next course. Without it the whole thing runs once and you repeat whatever you did, having learned nothing about how you learn.
Almost nobody looks at a returned exam properly, and it is the highest-value material anyone has ever handed you: a precise list of things you believed and were wrong about. The comparison is Crash Course Study Skills’s: chess players at the top of the game spend enormous amounts of time reviewing their own lost games, and for exactly this reason.
That is the content half, and the Diagnostic is a structured way of doing it. Then comes the harder question, which is about you rather than about the material: why did the mistake happen? Not enough time? Ignored the study guide? Prepared, but by a method that does not transfer? Rushed and made careless errors? Ran out of time and never reached the last page? Those have completely different fixes, and “I’ll do better next time” is not one of them.
One reaction deserves special attention. If your first thought on seeing the grade is but I thought I knew this, Dennis Shaffer treats that sentence as a red flag in its own right: it usually means the studying was shallow, or that your sense of what you knew was off, or both. See the red flags in Step 4.
Here is the trap in looking back, and it is the same one that Step 3 describes for knowledge. When you review how you prepared, you are surveying the methods you actually used. Those are the survivors. And the thing that selected them was not whether they worked.
The selection pressure on study methods is comfort, not effectiveness. Which means your repertoire has been optimized for the wrong thing, by a process you never consciously ran. And when you survey it after a bad exam, you find no obvious fault in the methods you can see — because there is nothing visibly wrong with rereading — so the natural conclusion is to do more of them. Study harder. Reread more.
That is armoring the bullet holes. The damage you can see is the damage that was survivable. What is missing from the picture is the thing that would have made the difference, and it is missing precisely because it was the uncomfortable one.
Inspecting your methods will not reveal the gap, because the gap is an absence. So ask the question directly. Did I ever, at any point, close everything and try to produce this material from memory? Did I ever attempt a problem before seeing it solved? Did I ever say any of it out loud to another person? If the honest answer is no, you have found something more useful than any critique of what you did do.
This is the same reasoning as Abraham Wald’s, described in Step 3, applied to your own habits rather than to your knowledge. Both are cases of survivorship bias, and in both the fix is the same: reason deliberately about what is absent, because absence does not announce itself.
A post-mortem that produces a list of twelve resolutions produces nothing. Pick one change, make it specific enough that you will know whether you did it, and check at the next exam whether it helped.
Some changes that are small enough to actually happen:
After a midterm is the good case, and it is not too late in the way it feels. There is another exam, frequently a cumulative one, and the material you just got wrong is the material most likely to return. Doing this within a few days of getting the paper back is worth more than any equivalent hour spent rereading.
After a final it is still worth doing, for a different reason. The finding is no longer for this course; it is for the next one. And the strategy you are evaluating is the strategy that will otherwise run again, unchanged, in January, against material that may be harder.
And it is worth doing after an exam that went well, which almost nobody does. A good result tells you that whatever you did was sufficient for that exam. It does not tell you which parts of it were doing the work, and it does not tell you whether the same approach survives a harder course. The most useful version of this question, asked after a success, is: how much of that was the method, and how much was that the material happened to suit me?
The named ideas used on this page, in one place. Each entry says which step it belongs to, so this doubles as a way of finding the part you want.
Close everything and write out a topic from memory — the logic, the diagram, the equations — noting exactly where you hit a wall. Open the source only long enough to get that one missing piece, then close it and start again. The failures are the output: they are a precise list of what to study. Write them down.
A page divided into a narrow left cue column, a wide right notes column, and a summary strip along the bottom. Notes go in the right column during class; the cue column and the summary are filled in afterward, and the cue column is then used to quiz yourself with the notes covered. The popular version omits the afterward part, which is the part that does the work.
Deciding how well you know something after a gap — an hour or more, with the material out of sight — rather than immediately after studying it. Immediate judgments mostly measure short-term memory and run optimistic; delayed ones are considerably more accurate. The practical version on this page is Dennis Shaffer’s.
The principle that the harder a retrieval was — the closer you came to not finding it — the more the attempt was worth. Which is why effective study feels worse than ineffective study, and why people abandon it.
There is an upper bound, though. Leave something so long that the attempt fails completely — nothing comes back at all — and you are no longer retrieving anything. You are learning it again from the beginning, and the extra difficulty bought you nothing.
Relating what you are studying to what you already know: examples, applications, where it fits, what category it belongs to. The most general form of deep processing, and what turns a definition into something with somewhere to be recalled from. An established term; its use on this page follows Dennis Shaffer’s talk.
On a four-option question, a blind guess is 25 percent; eliminating one option makes it 33; eliminating two makes it 50. The point is not the numbers but the reframing: partial knowledge is worth something, and the job is to reduce the field if you do not know the answer outright.
Ease produced by the material being smooth, mistaken for ease produced by you having learned it. A reread page, a highlighted passage, tidy notes, and a familiar-looking answer option all generate a fluency illusion. It is the single mechanism behind most of what goes wrong in studying, and it is why the methods that work feel worse.
Generation here means producing something yourself — saying it, writing it, drawing it, working the problem — rather than reading or watching someone else do it. Nothing to do with generations of people, or with what a computer generates. You keep what you generate far better than what you receive.
Which is the same fact as the fluency illusion, turned the other way round: the effort you feel while producing something is not an obstacle to the learning, it is the learning.
Andrew Deutscher’s term for straddling two things at once and doing a disservice to both. The reason it is hard to correct is that it does not feel like inattention; it feels like handling both.
Carol Dweck’s term for treating ability as something that changes with effort rather than a fixed quantity. The practical form, as Dennis Shaffer puts it: replace I’m not good at this with I am weaker at this right now. How much short interventions change grades is debated; the reframing itself is accurate.
High-density: something you actually want to do, with a beginning and an end, which returns something. Low-density: more attractive than working, easy to tell yourself will only take five minutes, engineered to have no stopping point, and not actually enjoyable. Refusing yourself the high-density kind is what leaves the low-density kind as the only available relief — and relief is important.
Mixing topics within a study session rather than finishing one before starting the next. It feels disorganized but performs better, partly because it forces repeated retrieval and partly because it surfaces differences between things that looked similar when studied apart.
A learning goal is about getting better — what you know and how you work. A performance goal is about a particular result on a particular assessment. Dennis Shaffer’s advice is to focus on the first, since good performance follows from good process but not the reverse. Effort aimed only at performance is what Schlechty’s strategic compliance looks like from inside.
Five boxes at increasing review intervals. Every card starts in box one; get it right and it moves forward, get it wrong and it returns to box one from wherever it was. Spaced repetition that adjusts itself to what you personally find hard, runnable with index cards.
A page with the concept in the middle and branches radiating outward to what hangs off it, with room for cross-links between branches. You can only place something once you know what it relates to, which is why drawing one is a Step 3 activity rather than something to attempt during a lecture.
Cal Newport’s marking scheme for reading: a dot in the margin beside a big idea, a dash beside an example or explanation supporting it, and nothing else. It leaves the material for Step 3 without interrupting the reading to do Step 3.
Attempting questions before you have studied the material. The attempts mostly fail, and they still improve how well the material is learned afterward — and they tell you what level of processing the exam is going to ask for.
Four rungs, each resting on the one below. Recognize: you know it when you see it. Remember: you can produce it from nothing. Understand: you can say what it means in your own words. Apply: you can use it on something new. Memorizing is not the wrong thing — it is the second rung, and understanding what a definition means usually makes memorizing it easier. What goes wrong is stopping, and people stop at remember because it is the first rung that feels like success.
The three ways a memory gets accessed, and the three things test formats are built on. Recognition needs a cue and is what rereading and multiple choice train. Recall supplies its own cue and is what most exams actually demand. Confusing the two is the most common way to study hard and still fail.
What it takes to rebuild your context after time away — where you were, what you had decided, what comes next. It grows with the gap, it is charged every time you sit down, and it is the reason abandoned projects get abandoned without anyone deciding to abandon them.
Trying to produce the answer from memory rather than reviewing it. The single best-supported study technique there is. Failed attempts still help, which matters because failure is what often makes people stop doing it, to their detriment.
Two of Phillip Schlechty’s five levels of engagement. Ritual compliance is doing the minimum to avoid trouble, and it feels like attending. Strategic compliance is high attention with low commitment — working hard, aimed at the grade — which looks like engagement and predicts poor retention and little transfer.
The tendency to remember information better when you relate it to yourself — your own life, people you know, things you have seen. A form of elaboration, and unusually easy to use in biology, which is largely about you. An established finding; its use on this page follows Dennis Shaffer’s talk.
Handling information by how it looks or sounds — repeating, recopying, highlighting, recognizing a term on a card — versus by what it means: where it fits, what it connects to, what it is an example of. Deep processing produces far better retention. The framework is Fergus Craik and Robert Lockhart’s levels of processing model (1972); its use on this page follows Dennis Shaffer’s talk.
Reasoning from a sample that contains only the cases that made it through. Abraham Wald’s bombers are the standard example: the returning planes showed damage where damage was survivable, so the undamaged regions were the lethal ones. Your own knowledge is such a sample — what you missed left no trace, because missing it is what kept it off the page — which is why gaps cannot be found by introspection. So are your study methods: the ones you still use survived because they were comfortable, not because they worked.
The same total hours spread across days beat the same hours in one sitting, reliably and by a wide margin. Sleep between sessions appears to matter. Reading before a lecture produces spacing automatically, without anyone planning it.
A way of describing forgetting as a filing problem rather than a loss: the books stay on the shelves while the catalog degrades. What fades is your path to the memory, and every retrieval reorganizes the catalog.
The point at which your own materials run out — notes, then textbook — and you go to something outside them: a search, an assistant, a classmate, the instructor. Not a step of its own, just the place in the loop where the answer is no longer anywhere you already have.
States describing where an answer has to come from, named rather than numbered so as not to collide with the steps. In your head — which is worth verifying on paper, because most people who believe they are here are not. On paper. Workable, meaning you can get there with a book, a search, or an assistant. Needs a person. What separates the last two is not the source but whether you are asking for a piece you will then use or for the finished answer.
A short tab, addressed to colleagues, on the parts of this that sit on our side of the desk.
The most useful thing to convey may simply be that studying has steps, and that a student who is struggling can ask which step rather than concluding something about themselves. Two specific points land particularly well:
Dennis Shaffer ends one recording of his talk on study habits with an assignment he calls a mission for empowerment, and it is worth borrowing whole. Each student writes a plan for improving how they study in the course, built from at least one general principle and five specific strategies from the talk. For each strategy they say why they expect it to help them in particular, and whether it is similar to or different from what they have been doing. Then they turn the plan into a day-by-day weekly schedule — including days on which they will deliberately do nothing, with a reason for each.
Three things make it better than simply telling students the same content. Writing the plan is itself generation, so the strategies are produced rather than received. Asking what is different from current practice is exactly the question that Step 6 argues students never ask on their own, because the methods they abandoned leave no trace. And requiring rest days, with reasons, builds in the stopping that Step 4 treats as part of studying rather than a lapse from it.
One difference in emphasis from the rest of this page. Step 6 recommends changing one thing at a time after an exam, because a list of twelve resolutions produces nothing. Five strategies at the start of a course is a different situation — a plan made before there is any evidence, rather than a correction made after it — and the weekly schedule is what keeps five from becoming a list of intentions.
Phillip C. Schlechty’s levels — rebellion, retreatism, ritual compliance, strategic compliance, engagement, set out in his Working on the Work framework — were built for us rather than for students, and the design implication is the uncomfortable one. A quiet, orderly, hard-working room can be almost entirely strategic compliance, and strategic compliance produces exactly the outcome we then blame on students: high performance during the course, poor retention afterward, and little transfer to anything new.
The half we have direct leverage on is commitment, and we get at it by making the material worth wanting — why this question was ever asked, what it is good for, where it stops working. That is also the only part of the whole business that a student cannot simply be told to do.
There is an entire genre of test-taking advice built on exploiting how people write multiple-choice exams. Most of it is defeated by ordinary care:
A story against my own interest. In the physics course I took as an undergraduate, the last section of the semester was electronics — circuits, and everything that goes with them. I learned it, and I learned it reasonably well at the time. Then the professor announced that anyone above a certain grade need not sit the final, and I was above that grade, so I did not sit it.
Forty years later there is an emptiness in my head where that material should be. I can tell you that I once knew it. I cannot tell you any of it. And the gap is specific to that unit, which is not a coincidence.
Would knowing this have changed what I did at nineteen? Almost certainly not, and that is the honest part. The offer is genuinely attractive and declining it would have felt absurd. Which is why the observation is aimed at us rather than at them: a student cannot be expected to price a benefit that arrives decades later. We are the ones in a position to know what the policy costs.
None of which is an argument against the incentive, which does real work — it rewards sustained performance and it reduces end-of-term load for people who have earned it. It is an argument for noticing the tradeoff and, where it is possible, arranging things so the last unit gets retrieved somewhere regardless. Exempting from the grade consequence rather than from sitting the exam does it. So does putting the final unit somewhere else — a required low-stakes quiz, a problem set, a question on a later assessment — so that nobody leaves the course having met that material exactly once.
Offered as a worked example rather than a recommendation, since much of it depends on circumstances:
The pieces above combine into something I think is worth more than any of them separately. Because assessment runs in Canvas, in the classroom, on a lockdown browser, and because I can generate fresh questions from a bank — with help from recorded lectures and from AI — it costs me very little to let a student sit an assessment again.
So I do. Students may retake twice, with the grade on the second retake capped at 90 percent. Retakes happen during my office hours, which I deliberately hold in an empty classroom rather than in my actual office, which is cramped. Canvas grades most of it automatically, so I am not returning to the same pile of papers three times over.
I can do this because of circumstances, not because of virtue, and it would be misleading to present it otherwise:
Remove any one of those and the arithmetic changes. But some of it scales down further than it looks — a single low-stakes retake on the unit that went worst, offered once a semester, captures a good deal of the benefit at a fraction of the cost.
A question resting on prior knowledge that the current course has never addressed is not a hard question. It is an unfair one. You should have known this already is, in my opinion, a mark of poor instruction rather than of insufficient studenting.
This is worth stating because the failure is invisible from the front of the room. Material that is obvious to us because we have carried it for twenty years reads to a student as something they are apparently supposed to have and do not. They will conclude they are not smart enough. They are frequently just missing one specific thing, and the fastest fix in existence is naming it out loud.
Nobody should have to read all of this. Every section of every tab is listed below, in order, with its subsections indented beneath it. Click anything to go straight there.
There is also a search box under the tab bar, which looks through the whole page and will usually find what you want faster than scanning this will.
Building the index…