Running yourself: attention, procrastination, sleep, and your system

Everything you've learned so far has one awkward property: none of it runs itself. You know that fluency lies (lesson 1), that retrieval beats rereading (lesson 3), when to come back (lesson 4), when to mix (lesson 5), how to get through hard material (lesson 6), and what practice with feedback looks like (lesson 7). Every one of those techniques costs effort now for a payoff later, and the two most common study habits, multitasking and putting it off, are ways of dodging exactly that trade.

So this lesson is about the delivery mechanism. Attention, procrastination, and sleep, with the evidence at its actual size rather than its popular size. The myths the course has been stepping around, with the studies that test them. And a template that builds two real weekly plans, one for five hours a week and one for a phone and a commute. The plan you write at the end is the start of the course project.

The core idea

The whole idea fits in one sentence: every technique in this course feels worse in the moment than its useless rival, so any decision you make in the moment will go the wrong way, and the only fix is to make the decision before the moment arrives.

You've seen the evidence for the first half all through the course. Rereading feels smoother than recall, blocked practice smoother than mixed, a phone smoother than a problem set. Bjork, Dunlosky and Kornell's 2013 review gives the cleanest single figure.

That's why self-management isn't a separate subject from learning. It's the condition under which the other techniques happen or don't. Retrieval only strengthens a memory if you open the blank page. Spacing only works if the second session exists. Interleaving only forces discrimination if you sit through the mixed set instead of switching to the phone. Everything below is a way of taking the decision away from the moment.

Attention: single-task, in short blocks

Start with the habit almost everyone has: studying with something else open. May and Elder (2018) reviewed the research on media multitasking, which means using a second stream of media (messages, video, social feeds) while doing academic work.² Their finding is consistent across the studies they surveyed. Media multitasking goes with lower grade point averages, worse test performance, worse recall, and worse reading comprehension, and the effects show up both in class and while students are studying on their own.

Two limits on that finding. It's a narrative review, so it reports direction and consistency, not a pooled effect size; you're acting on the agreement of many studies, not on one number. And much of the survey evidence is associational, so some of the relationship could run the other way, with weaker students multitasking more.

The review also found that students misjudge the cost.² Picture a learner with a chat open who judges the lecture "fine". They're judging moment-to-moment comprehension, and lesson 2 showed why that's the wrong measure: working memory holds about four chunks, and a second stream takes some of them.³ What you feel is that you followed along. What you lose is the residue.

The practical rule is short blocks of single tasking. The block length is a convention. No study fixes 25 minutes, the figure in Francesco Cirillo's Pomodoro technique, which Oakley (2014) recommends to learners.⁵ What the evidence does fix is what has to be true of the block. One stream, because of May and Elder. A phone you can't reach rather than one you've resolved not to check, because changing the situation beats resisting it (Duckworth's point, below).⁴ And an end that is a retrieval act rather than a clock: close the notes and write what you have. A block with those three properties is doing the work; its length is whatever you can start. If 25 minutes is too long for where you are now, use 15. A short block is a startable block, which brings us to the next problem.

Procrastination: tactics from the self-regulation literature

Oakley's chapters are where most learners meet procrastination, and her vocabulary is a useful bridge: "focus on process, not product", meaning commit to a block of time rather than to finishing something, and her point, drawn from research on maths anxiety, that the discomfort peaks in anticipation and fades once you start.⁵ Her chapters are written for a general reader, so they carry the ideas on stories and metaphors and cite only a few studies. The three tactics below come from the self-regulation literature itself.

Implementation intentions. Compare two plans. "I will study statistics this week." And: "When I put my plate in the sink on Tuesday, I open the statistics notes." Gollwitzer (1999) called the first a goal intention and the second an implementation intention, which always has the form "when situation X arises, I will do Y".⁶ A goal intention leaves the decision of when and how to a future self who will be tired and holding a phone. An implementation intention hands the cue to the situation itself, so the action fires without a fresh decision.

Gollwitzer and Sheeran's 2006 meta-analysis of 94 tests found that adding an if-then plan to a goal produced a medium-to-large improvement in goal attainment over the goal alone, d = 0.65, across a wide range of behaviours.⁷ What makes it work is specificity: "after dinner" is weaker than "when I put my plate in the sink on Tuesday".

Shrink the first step. Steel's 2007 meta-analysis of 691 correlations found the strongest, most consistent predictors of procrastination were how aversive a task is, how far away its payoff is, low confidence in doing it, and impulsiveness.⁸ You can't move the exam, but you can make the first step nearly costless. "Study statistics" is aversive. "Open the notes to question 4 and read it" isn't. Once you've read question 4 you're in, and Oakley's point about the discomfort fading applies. Write the first step into the if-then plan so that the cue triggers something almost too small to refuse.

Design the environment. Duckworth, Gendler and Gross (2016) argue, from a process model of self-control, that because an impulse strengthens over time, strategies applied before temptation appears (choosing or changing the situation) should beat strategies applied during it (resisting), and that people underuse them because they feel effortless.⁴ It's an argument, not a trial result, but it's the reason the concrete moves work: the phone in another room or on airplane mode, the notes already open, the problem set printed the night before, the chat apps in a folder off the home screen. Each removes a decision from the moment when your judgement is worst.

None of these is a personality change. They're engineering around a known weakness in how people decide about future effort, which is why they work for people who have "tried to be more disciplined" many times.

Sleep, at its actual size

You've probably heard that sleep is where learning happens and that a night's sleep after study is worth more than the study. The evidence supports a weaker and more interesting claim.

The mechanism is well developed. Rasch and Born's 2013 review in Physiological Reviews describes active systems consolidation: during slow-wave sleep, memories encoded during the day are reactivated and gradually integrated into longer-term stores.⁹ That consolidation happens during sleep is established; how much of it runs through slow-wave sleep specifically is now in question.

The behavioural size of the effect is where the popular version overreaches. Cordi and Rasch (2021), with Rasch a co-author of that review, asked how reliable sleep-mediated memory benefits are and concluded they are "smaller, more task-dependent, less SWS-related, less robust and less long-lasting than previously assumed".¹⁰ In a separate 2021 study with 159 participants they found no within-person correlation between the amount of slow-wave sleep and how much was consolidated.¹¹ That's a careful group reporting that its own headline is smaller than it looked.

The classroom-level number is smaller still. Dewald et al. (2010) meta-analysed sleep and school performance across 15,199 children and adolescents.

Read that number for what it is. It's a between-person correlation across children who mostly sleep roughly enough, so it says the differences in their hours don't track grades much. It's the wrong number for "does a lost night hurt" and the right number for "is sleep the lever". Numbers this small are hard to feel, so here they are drawn on the full scale a correlation can take:

A lost night does hurt, and it hurts attention first. Lim and Dinges' meta-analysis of 70 studies of short-term sleep deprivation found lapses in simple attention among the largest effects (g of about −0.78), with reasoning accuracy barely moved.¹³ So a tired study session is a worse one, mostly because the single-tasking you just read about gets harder.

Classify this the way lesson 7 classified practice: established that sleep matters for consolidation, contested how much and by what mechanism. Sleep enough, because consolidation is real and because a tired session is a worse one. Don't skip a review on the theory that sleep will do the work. Matthew Walker's Why We Sleep is the book most learners will have met; this course doesn't cite it, because its claims run larger than the reviews above support and a 2019 critique documented factual errors in it.¹⁴

The myths, named, and one contested question

The course has already refuted "if it feels easy I've learned it" and "rereading is studying" (lesson 1), "practise one thing until it's perfect" (lesson 5), and "10,000 hours" (lesson 7). Four more belong here, plus one that is contested rather than false.

Learning styles. The claim: people have a visual, auditory, or kinaesthetic style and learn better when teaching matches it. Pashler, McDaniel, Rohrer and Bjork (2008) set out what would support it, a crossover interaction where the best method for one style is not the best for another, and found virtually no evidence of that pattern despite a large literature.¹⁵ Preferences exist; the benefit of matching doesn't show up.

Yet Dekker et al. (2012) found over 90 percent of surveyed UK and Dutch teachers endorsed the idea,¹⁶ and Nancekivell, Shah and Gelman (2020) found that many believers treat styles as innate, brain-based essences, which helps explain why the belief resists evidence.¹⁷ This isn't contested among researchers, so it doesn't get both sides.

Hard fonts help. The idea that a difficult-to-read typeface makes you process text more deeply is the canonical undesirable difficulty. Lessons 1, 5 and 7 gave Bjork and Bjork's condition: a difficulty helps only if you can meet it. This lesson adds the other half: it also has to engage the processing that builds the memory, as retrieval and mixed sets do. A hard font does neither. Rummer, Schweppe and Schwede (2016) found no benefit across three experiments.¹⁸ A meta-analysis of 25 studies by Xie, Zhou and Liu (2018) put the effect near zero (d = −0.01 for recall, 0.03 for transfer),¹⁹ and a 2021 reanalysis that found errors in it still left the transfer estimate at zero.²⁰ Of the studies of the purpose-built font Sans Forgetica published between 2020 and 2022, three found no memory benefit at all,²¹²²²³ and the rest found one only under conditions that don't resemble studying, such as when no test was expected or only for strong spellers.²⁴²⁵²⁶ Use a readable font.

Handwriting beats typing. Mueller and Oppenheimer (2014) reported that longhand note-takers beat laptop note-takers on conceptual questions, and the result travelled everywhere.²⁷

Lesson 2 still holds: notes that reconstruct beat notes that copy, and copying is easier on a keyboard. But the medium isn't the mechanism.

Ten percent of the brain; left-brain versus right-brain. Both sit in the Deans for Impact misconceptions table, produced with Willingham.³⁰ Imaging shows activity throughout the brain and damage almost anywhere costs something, so there is no idle 90 percent. And Nielsen et al. (2013) looked for "left-brained" and "right-brained" people in resting-state scans of 1,011 individuals and found lateralisation in particular connections, not in whole people.³¹

Growth mindset: contested, not a myth. The claim is that teaching students that ability can grow improves their achievement. Two camps read the same literature differently; here is each in its own terms.

Yeager and Dweck's position, set out in Yeager et al. (2019) in Nature and Yeager and Dweck (2020):³²³³ a preregistered national randomised trial of over 12,000 ninth-graders in 65 schools found that an intervention lasting under an hour raised grades among lower-achieving students and increased enrolment in advanced mathematics, with effects appearing where peer norms supported the message. On this reading the effects are real, small, and heterogeneous, and the heterogeneity was predicted in advance: the trial was designed to find where the intervention works, and averaging it with contexts where no effect was expected is, they argue, the wrong test. They also argue that the meta-analyses pool interventions of very uneven fidelity, and that a gain that is small per student, from under an hour at almost no cost, is meaningful across a population and compares well with far more expensive school programmes.

Sisk et al. (2018) and Macnamara and Burgoyne (2023) read it differently:³⁴³⁵ across 43 intervention studies Sisk found an average effect of d = 0.08; across 63 studies with 97,672 participants Macnamara and Burgoyne found d = 0.05, falling to 0.02 in the six studies they rated best designed. On this reading the effect shrinks as design quality rises, which they treat as the signature of design flaws and bias rather than of hidden moderators, and the honest average is close to nothing.

What would settle it is more preregistered, third-party, adequately powered trials with moderators specified in advance. Until then, the calibrated statement is: across all studies the average effect is small (d of about 0.05 to 0.08), and the dispute is over whether that average is the right summary. One camp reads it as real effects concentrated in identifiable students and schools, diluted by weaker studies; the other reads it as a near-zero true effect inflated by weaker studies. That's different from both the poster version and the debunking version.

For your own study, the disputed part is whether being told "ability grows" changes anything. The part both camps report is that where effects appeared at all, they appeared where the surrounding environment reinforced the message. That's the same lesson as the phone in the other room: the situation does more than the resolution.

The system: a week-plan template

Everything above feeds one deliverable: a week of study that spaces, interleaves where appropriate, retrieves, and actually happens. Fill the template in for real hours, not aspirational ones.

1. Hours this week. The honest number, after work, sleep, and life.
2. Subjects or skills, each tagged with what kind of material it is: list-like (vocabulary, dates, formulas), problem types (maths, statistics, grammar rules), prose (history, theology, law), or a physical or performed skill.
3. Horizon and gap for each: when you need it, and 10 to 20 percent of that as the review gap (lesson 4). For a subject with no deadline, treat the horizon as a year, where the gap is nearer 5 to 10 percent, roughly three to five weeks.
4. Blocks. For each block: the if-then cue, the subject, the technique, and what "done" looks like ("three problems attempted and checked", not "did some stats").
5. One weekly free-recall session: a blank page per subject, no notes, then re-study only what you missed (lesson 3).
6. A drop rule. What gets cut first when the week goes wrong, decided now.

The tag in step 2 decides the technique. List-like material gets spaced retrieval and not interleaving, because Brunmair and Richter (2019) found interleaving harms word lists (g of minus 0.39).³⁶ Problem types get interleaved sets. Prose gets free recall on the way out, and self-explanation on the way in once you have enough background to explain with: Dunlosky et al. (2013) found its effects grow with prior knowledge and are uncertain without it (lesson 6),³⁷ so for the first week of a new subject, read and recall, and add the "why is that true?" prompts from the second. Skills get deliberate practice with a feedback source (lesson 7).

Worked example: five hours, three subjects

Say you have three things on: Spanish vocabulary for a trip in eight weeks, a statistics module with an exam in five weeks, and a history course you're reading for its own sake. Five hours a week, 300 minutes.

Gaps first. The exam is 35 days out, so 10 to 20 percent gives a gap of three and a half to seven days: touch statistics at least twice a week. One caution on the rule itself. It's the course's simplification of Cepeda et al. (2008), whose own five-week condition found the best single gap nearer eleven days and whose data say gaps that are too short cost more than gaps that are too long.³⁸ So twice a week is a floor for a topic you've just met, not a target to exceed for its own sake.

Now do the next one yourself before reading on. The trip is 56 days out. Apply the same rule and write down the gap, then decide how you'd split the vocabulary time.

The history has no deadline; treat it as a one-year horizon, where Cepeda et al. put the optimal gap at 5 to 10 percent, roughly three to five weeks.³⁸ So each week's chapter gets its spaced review about a month later. The shorter recall you'll see on Fridays is a check that keeps the chapter warm, not the spaced review itself.

The blocks, each with its cue and its "done":

Sunday's "criterion" is Rawson and Dunlosky's (2011) prescription: three correct recalls the first time an item is learned, then one correct recall in each later spaced session.³⁹

That's 300 minutes. Notice where interleaving lives: statistics problem types are mixed within Thursday's block, but the Spanish cards aren't, because they're a word list. Spreading subjects across the week is spacing, not interleaving; the plan uses both. Drop rule: if the week collapses, Saturday's free recall survives and Wednesday's new chapter goes, because a missed retrieval breaks a spacing gap you can't get back (lesson 4), while a missed chapter is only a week's delay.

A harder example: 45 minutes a day, on a phone, on a train

Now the wrinkle: no desk time on weekdays, two train legs of about 22 minutes, standing, on the phone that is also where your messages live. Same statistics module, same language.

What fits. Spaced retrieval of cards or the site's Review page fits a leg exactly: short items, no paper, tolerant of interruption. Typed free recall fits too: open a notes app, write what you remember of the statistics you last studied, check against the notes only afterwards. Audio fits with one rule: pause every ten minutes and recall what was said, because listening without retrieval is rereading with your ears. Self-explanation of a prose passage fits, typed as answers to "why is that true?".

What doesn't fit. Worked-example cycles need paper, sustained working memory, and few interruptions; a train supplies none, and lesson 2 says every interruption is extraneous load. Deliberate practice of a performed skill needs the skill and a feedback source. Long writing needs a keyboard. So stage the week: the commute carries retrieval and spacing, and one 60-minute desk block at the weekend carries the worked examples and the mixed problem set the week's retrieval has been preparing you for.

Ten legs and one desk block, 280 minutes:

Check the gaps. A statistics topic met on Saturday is retrieved Monday, Tuesday, Thursday and the next Saturday: gaps of two to three days, inside the exam's window, and it re-enters the mixed set every week after that. Spanish cards run three times a week with the app setting the per-card gap. Every slot has a cue: the train doors closing.

The phone is the catch: the single-tasking problem and the environment-design problem in one object. So: Do Not Disturb for the leg, the review app on the home screen, the chat apps in a folder, and the if-then plan cued by the doors. If the leg still goes to messages, that's data, and the answer is a further environment change, not more resolve.

And the 45 minutes isn't "too short to be worth it". Two retrieval blocks a day is a better week than one three-hour reread on Sunday. Roediger and Karpicke's 2006 result from lesson 1 is the reason: at one week the tested group held 56 percent against 42 percent for the restudiers, and extra rereading time doesn't change which side of that line you're on.⁴⁰

What people get wrong

• "I need long uninterrupted blocks or it's pointless." The evidence in this course is about gaps and retrieval, not block length. A 20-minute block that retrieves beats a two-hour block that rereads.
• "I'd notice if it wasn't working." Kornell and Bjork (2008) showed 72 people their own results from two study methods; 64 still said the worse one had worked better.⁴¹ Seeing the data isn't enough. The plan has to be written before you consult your judgement, which is the point of step 4 of the template.
• "I'll be more disciplined this term." Discipline is a goal intention. The literature's answer is cues, small first steps, and a changed environment, none of which require you to be a different person.
• "Sleep will consolidate it, so I can skip the review." Consolidation is real and smaller than advertised. Review anyway.

Practice

Connections

This lesson is where the course turns from findings into a routine. Lesson 1's fluency illusion explains why multitasking feels harmless and why procrastination is so easy to rationalise; lessons 3 to 7 supply the techniques the plan schedules; and the project that follows tests whether the plan worked, with a delayed self-test rather than an impression. The final test draws on all eight lessons, and the myths here will be on it.

Go deeper

• Gollwitzer, "Implementation Intentions: Strong Effects of Simple Plans", American Psychologist 54 (1999): the original statement of if-then planning, short and readable.
• Cordi and Rasch, "How robust are sleep-mediated memory benefits?", Current Opinion in Neurobiology 67 (2021): researchers honestly resizing their own finding; a model of calibrated science.
• Pashler, McDaniel, Rohrer and Bjork, "Learning Styles: Concepts and Evidence", Psychological Science in the Public Interest 9 (2008): read it for the crossover test itself, which you can apply to any future "styles" claim.
• Deans for Impact, The Science of Learning (free PDF): the one-page misconceptions table is the fastest check on any study advice you encounter.

Sources

1. Bjork, R. A., Dunlosky, J. and Kornell, N., "Self-Regulated Learning: Beliefs, Techniques, and Illusions", Annual Review of Psychology 64, 417–444 (2013). 90 percent did better after spacing; 72 percent still preferred massing.
2. May, K. E. and Elder, A. D., "Efficient, helpful, or distracting? A literature review of media multitasking in relation to academic performance", International Journal of Educational Technology in Higher Education 15, article 13 (2018). Narrative review; consistent negative associations with GPA, test performance, recall, reading comprehension; effects shown in class and while studying; students misjudge the impact. No pooled effect size.
3. Cowan, N., "The magical number 4 in short-term memory", Behavioral and Brain Sciences 24, 87–114 (2001). Working memory holds about four chunks (three to five) when rehearsal and grouping are prevented.
4. Duckworth, A. L., Gendler, T. S. and Gross, J. J., "Situational Strategies for Self-Control", Perspectives on Psychological Science 11, 35–55 (2016). Theoretical paper: situation selection and modification before temptation are argued to beat effortful resistance during it.
5. Oakley, B., A Mind for Numbers (Tarcher, 2014), chapters 5 and 6. Bridge vocabulary only: Pomodoro (credited there to Cirillo), "process not product", discomfort peaking in anticipation.
6. Gollwitzer, P. M., "Implementation Intentions: Strong Effects of Simple Plans", American Psychologist 54, 493–503 (1999). Goal intentions versus if-then plans.
7. Gollwitzer, P. M. and Sheeran, P., "Implementation Intentions and Goal Achievement: A Meta-analysis of Effects and Processes", Advances in Experimental Social Psychology 38, 69–119 (2006). 94 independent tests; medium-to-large effect on goal attainment, d = 0.65.
8. Steel, P., "The Nature of Procrastination: A Meta-Analytic and Theoretical Review of Quintessential Self-Regulatory Failure", Psychological Bulletin 133, 65–94 (2007). 691 correlations; task aversiveness, task delay, self-efficacy and impulsiveness as the strong, consistent predictors.
9. Rasch, B. and Born, J., "About Sleep's Role in Memory", Physiological Reviews 93, 681–766 (2013). Active systems consolidation during slow-wave sleep.
10. Cordi, M. J. and Rasch, B., "How robust are sleep-mediated memory benefits?", Current Opinion in Neurobiology 67, 1–7 (2021). Quoted phrase on effects being smaller and less robust than assumed.
11. Cordi, M. J. and Rasch, B., "No evidence for intra-individual correlations between sleep-mediated declarative memory consolidation and slow-wave sleep", SLEEP 44(8), zsab034 (2021). 159 participants; no within-person correlation between slow-wave sleep and consolidation.
12. Dewald, J. F. et al., "The influence of sleep quality, sleep duration and sleepiness on school performance in children and adolescents: A meta-analytic review", Sleep Medicine Reviews 14, 179–189 (2010). Sleep duration r = 0.069 (17 studies, N = 15,199); sleep quality r = 0.096; sleepiness r = −0.133.
13. Lim, J. and Dinges, D. F., "A Meta-Analysis of the Impact of Short-Term Sleep Deprivation on Cognitive Variables", Psychological Bulletin 136, 375–389 (2010). 70 articles; lapses in simple attention g = −0.78; reasoning accuracy g = −0.13, not significant.
14. Guzey, A., "Matthew Walker's Why We Sleep Is Riddled with Scientific and Factual Errors" (guzey.com, 2019). Documented errors in the book; the reason this course does not cite it.
15. Pashler, H., McDaniel, M., Rohrer, D. and Bjork, R., "Learning Styles: Concepts and Evidence", Psychological Science in the Public Interest 9, 105–119 (2008). The crossover test; virtually no evidence for it.
16. Dekker, S. et al., "Neuromyths in education: Prevalence and predictors of misconceptions among teachers", Frontiers in Psychology 3:429 (2012). 242 teachers; 93 percent of UK and 96 percent of Dutch respondents endorsed learning styles.
17. Nancekivell, S. E., Shah, P. and Gelman, S. A., "Maybe they're born with it, or maybe it's experience: Toward a deeper understanding of the learning style myth", Journal of Educational Psychology 112, 221–235 (2020). Many believers treat styles as innate, brain-based essences.
18. Rummer, R., Schweppe, J. and Schwede, A., "Fortune is fickle: null-effects of disfluency on learning outcomes", Metacognition and Learning 11, 57–70 (2016). Null across three experiments.
19. Xie, H., Zhou, Z. and Liu, Q., "Null Effects of Perceptual Disfluency on Learning Outcomes in a Text-Based Educational Context: a Meta-analysis", Educational Psychology Review 30, 745–771 (2018). 25 articles, 3,135 participants; recall d = −0.01, transfer d = 0.03.
20. Weissgerber, S. C., Brunmair, M. and Rummer, R., "Null and Void? Errors in Meta-analysis on Perceptual Disfluency and Recommendations to Improve Meta-analytical Reproducibility", Educational Psychology Review 33, 1221–1247 (2021). Coding errors in Xie et al.; transfer estimate robust, recall estimate doubtful.
21. Taylor, A., Sanson, M., Burnell, R., Wade, K. A. and Garry, M., "Disfluent difficulties are not desirable difficulties: the (lack of) effect of Sans Forgetica on memory", Memory 28, 850–857 (2020). Four experiments; no benefit.
22. Wetzler, E. L., Pyke, A. A. and Werner, A., "Sans Forgetica Is Not the 'Font' of Knowledge: Disfluent Fonts Are Not Always Desirable Difficulties", SAGE Open 11(4) (2021). N = 120, one-week delay; font effect not significant.
23. Huff, M. J., Maxwell, N. P. and Mitchell, A., "Distinctive Sans Forgetica font does not benefit memory accuracy in the DRM paradigm", Cognitive Research: Principles and Implications 7:102 (2022). Four experiments; no effect on correct or false memory.
24. Geller, J., Davis, S. D. and Peterson, D. J., "Sans Forgetica is not desirable for learning", Memory 28, 957–967 (2020). Three preregistered experiments; benefit only when no test was expected.
25. Geller, J. and Peterson, D., "Is this going to be on the test? Test expectancy moderates the disfluency effect with Sans Forgetica", Journal of Experimental Psychology: Learning, Memory, and Cognition 47, 1924–1938 (2021). Same boundary in a larger sample.
26. Eskenazi, M. A. and Nix, B., "Individual differences in the desirable difficulty effect during lexical acquisition", Journal of Experimental Psychology: Learning, Memory, and Cognition 47, 45–52 (2021). 160 participants; benefit only for high-skill spellers.
27. Mueller, P. A. and Oppenheimer, D. M., "The Pen Is Mightier Than the Keyboard", Psychological Science 25, 1159–1168 (2014). The original longhand-advantage claim.
28. Morehead, K., Dunlosky, J. and Rawson, K. A., "How Much Mightier Is the Pen Than the Keyboard for Note-Taking? A Replication and Extension of Mueller and Oppenheimer (2014)", Educational Psychology Review 31, 753–780 (2019). No consistent difference between groups.
29. Urry, H. L. et al., "Don't Ditch the Laptop Just Yet: A Direct Replication of Mueller and Oppenheimer's (2014) Study 1 Plus Mini Meta-Analyses Across Similar Studies", Psychological Science 32, 326–339 (2021). n = 142 plus mini-meta-analysis of eight studies; no reliable advantage.
30. Deans for Impact, The Science of Learning (2nd ed., 2026; 1st ed. 2015). One-page misconceptions table: 10 percent of the brain; learning styles and left/right brain; novices think like experts.
31. Nielsen, J. A., Zielinski, B. A., Ferguson, M. A., Lainhart, J. E. and Anderson, J. S., "An Evaluation of the Left-Brain vs. Right-Brain Hypothesis with Resting State Functional Connectivity Magnetic Resonance Imaging", PLOS ONE 8(8), e71275 (2013). 1,011 individuals; lateralisation is a property of particular connections, not of whole individuals.
32. Yeager, D. S. et al., "A national experiment reveals where a growth mindset improves achievement", Nature 573, 364–369 (2019). Preregistered RCT, over 12,000 ninth-graders, 65 schools; sub-hour intervention; gains among lower achievers where peer norms aligned.
33. Yeager, D. S. and Dweck, C. S., "What can be learned from growth mindset controversies?", American Psychologist 75, 1269–1284 (2020). Position A stated in its own terms.
34. Sisk, V. F. et al., "To What Extent and Under Which Circumstances Are Growth Mind-Sets Important to Academic Achievement? Two Meta-Analyses", Psychological Science 29, 549–571 (2018). d = 0.08 across 43 intervention studies.
35. Macnamara, B. N. and Burgoyne, A. P., "Do growth mindset interventions impact students' academic achievement? A systematic review and meta-analysis with recommendations for best practices", Psychological Bulletin 149, 133–173 (2023). d = 0.05 across 63 studies, N = 97,672; 0.02 in the six studies meeting their best-practice criteria.
36. Brunmair, M. and Richter, T., "Similarity matters: A meta-analysis of interleaved learning and its moderators", Psychological Bulletin 145, 1029–1052 (2019). Interleaving harms word lists, g = −0.39; used for the template's material tags.
37. Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J. and Willingham, D. T., "Improving Students' Learning With Effective Learning Techniques", Psychological Science in the Public Interest 14, 4–58 (2013). Self-explanation and elaborative interrogation depend on prior knowledge.
38. Cepeda, N. J. et al., "Spacing Effects in Learning: A Temporal Ridgeline of Optimal Retention", Psychological Science 19, 1095–1102 (2008). Optimal gap about 20 to 40 percent of a one-week delay, 5 to 10 percent of a one-year delay; optimal gap about 11 days at a 35-day delay; too-short gaps cost more than too-long ones. The course's 10 to 20 percent rule is a simplification of this ridgeline, not a figure from the paper.
39. Rawson, K. A. and Dunlosky, J., "Optimizing schedules of retrieval practice for durable and efficient learning: How much is enough?", JEP: General 140, 283–302 (2011). Successive relearning: three correct recalls, then one per spaced session.
40. Roediger, H. L. and Karpicke, J. D., "Test-Enhanced Learning", Psychological Science 17, 249–255 (2006). One-week result, 56 percent tested vs 42 percent restudied.
41. Kornell, N. and Bjork, R. A., "Learning Concepts and Categories: Is Spacing the 'Enemy of Induction'?", Psychological Science 19, 585–592 (2008). Of 72 participants with a preference, 64 chose the method that had worked worse for them.