Skip to content

Note-taking

Principle · All principles
Evidence · 19 claims (9 for, 10 mixed) · 32 studies (10 causal, 9 review, 7 quant-synthesis, 3 theoretical, 2 associational, 1 qualitative), q1–q4 · 10 of 32 report an effect size · 9 claims rest on one study

Conditional relationship

This page owns making a separate, selective record while information arrives (a lecture, a video, a demonstration, a discussion, a reading) and using that record afterwards. For a learner meeting a dense stream of new material (novice or intermediate on this content, whatever their general study skill) that they will need later for a test, a discussion, a problem or a piece of writing, notes that select, condense and restate the key ideas in the learner's own words, organized so the structure is visible, and then reviewed by recalling from them rather than rereading, are expected to produce better immediate-performance, conceptual-understanding and delayed-retention on that material than either no notes or notes that copy the stream word for word. Where the learner cannot yet select and organize at the pace of the stream, a partial outline supplied by the teacher (guided notes) is expected to make the notes more complete and accurate, and is withdrawn as the learner's own notes come to capture the key points.

The relationship is conditional on four things. Pace and density: when the stream is faster than the learner can select and restate, note-taking competes with listening and tends to fall back on transcription. What the learner does in the notes: copying adds little; restating, linking and questioning are the expected active ingredients. Later use: notes that are never reviewed or used are only an encoding activity, and their record value is lost. Learner expertise and skill: a learner who already takes selective notes gains little from a supplied outline and may lose time to it.

What the evidence tests, and what it does not. One claim tests a note-taking support directly: guided notes against unstructured or outline-only notes, on note accuracy and course performance (a meta-analysis read from its abstract, and one quasi-experiment with 65 undergraduates). One claim tests the medium: laptop against longhand notes, where a first set of experiments favoured longhand on conceptual questions and a direct replication did not reproduce it. No claim in the wiki tests taking notes against not taking notes, or generative notes (paraphrase, summary, question) against verbatim notes with the device held constant, or any named format (Cornell, outline, mapping) against another. The transformation and review steps below rest on neighbouring claims about self-explanation, summarization and retrieval practice, carried to note-taking and labelled as such.

How the converted neighbours sit beside this: Annotating holds marking the text itself (highlighting, margin notes) while reading; this page holds a separate record, most often of transient input such as speech. Graphic Organizers holds the visual form of a structure and who builds it; graphic notes are that case applied during a stream. Chunking holds the size and boundaries of units; pauses for note-taking are placed at those boundaries. Self-Explanation and Retrieval Practice hold the two activities this page asks learners to do with their notes, and their models are not repeated here. Self-Monitoring holds checks during a task; a "questions" column in notes is one such check. Clear Structure holds making an organization explicit, of which a guided-notes outline is one form.

Default design, while the relationship is untested

The earlier page's guidance, made concrete as a default a designer can act on and revise. Each step says what supports it.

  1. Tell learners what the notes are for, before the stream starts. Name the later use (a quiz next week, a discussion tomorrow, an essay, a procedure to carry out) and the two or three things the notes must capture for it (the claims, the steps, the evidence, the vocabulary). Untested proposal; from the earlier page's requirement of "a clear note-taking purpose".
  2. Start with guided notes for learners new to the content or to note-taking. Give a handout or on-screen outline with the headings and the order of the session, and blanks for the key ideas, definitions, steps and examples, roughly one blank per key point and none for incidental detail. Do not supply a full transcript: the blanks are what keep the learner listening for the point. Evidence: Guided Notes Improve Note Accuracy [+M] (meta-analysis, abstract only, strongest for school-age students; one quasi-experiment with undergraduates). The blank density and "not a transcript" are untested proposals from that claim's Discussion.
  3. Model selective notes once, early. In the first session, take notes on a 3–5 minute segment in front of learners, thinking aloud about what to keep, what to abbreviate and what to skip, and then show the same segment transcribed in full beside your notes. Untested as a note-taking step; the nearest evidence is that taught summarization improved summary quality against untaught practice, though not delayed recall or a reading test (Summarization Effective With Training [~M]).
  4. Build pauses into the stream. Stop every 8–12 minutes of talk or video, at a conceptual break, for 2–3 minutes in which learners complete gaps, fix what they missed and write one question about what is unclear. Untested proposal; the earlier page's "divided attention" constraint, with boundaries placed as Chunking describes.
  5. Ask for restatement, not copying. At each pause, or at the end, ask for one or two sentences in the learner's own words: what the main point was, and why a step or claim follows. Carried evidence: prompting learners to explain material to themselves improved immediate understanding and problem solving by a moderate average amount across task types and levels (Prompting learners to self-explain improves understanding and problem solving on immediate tests by a moderate average amount, with little evidence yet on delayed or classroom gains. [+S]). Untested inside note-taking. Do not ask for a summary of each of several sources before a task that requires combining them (step 9's caution).
  6. Review by recalling from the notes within a day or two, not by rereading them. Learners cover the notes (or the right-hand column of a Cornell page), answer the questions or headings in the margin from memory, then check and correct the notes. Carried evidence: testing beat restudy at delays of a day or more (Retrieval Practice Improves Long Term Retention [+S], 61% against 40% recall at one week in one study; meta-analyses g = 0.50 and 0.51); untested as a way of reviewing one's own notes.
  7. Use the notes in the next task. Have learners bring the notes to a discussion, a problem set or a draft and use them there, so a gap shows up while it can still be fixed. Untested proposal from the earlier page ("notes are most valuable when learners revisit, revise, compare, or use them"). Associational evidence only: in one tutoring-system study of 28 college students, those who complied more with a prompt to review their notes showed larger learning gains (Higher compliance with MetaTutor's review-notes prompt is associated with larger proportional learning gains [+W]).
  8. Check a sample of notes for errors, not only for completeness. Collect or glance at a few learners' notes after the first sessions, against a key of the main points, and correct a wrong point publicly, since a mistaken note will be reviewed and relearned. Untested proposal; from the earlier page's caution that poor notes "reinforce confusion".
  9. Withdraw the outline across sessions. A default arc for a course or unit:
  10. Sessions 1–3: guided notes with blanks, the modelled segment in session 1, pauses every 8–12 minutes, a restatement prompt at each pause, a 10-minute recall review at the start of the next session.
  11. Sessions 4–6: headings only (no blanks), pauses kept, learners write their own restatements; the review becomes a 5-question self-test from their notes.
  12. Session 7 onward: learner-built notes in a format they choose, a short checklist (main points, one example each, one question), pauses only for dense segments; review on their own, with a weekly check.
  13. Moving on: move a learner to the next stage when their notes capture most key points on the key (a proposal: about four in five) on two consecutive sessions; move back a stage if a dense or unfamiliar topic drops them below that. Learners who already meet it in session 1 skip to stage three (expertise reversal [~M]). The arc, the timings and the threshold are untested proposals.
  14. Graphic notes: supply the frame before asking learners to build one. If the content has a definite structure (a cycle, a hierarchy, a comparison), give the frame or a partly completed diagram first; ask learners to build their own only once they can fill a supplied one accurately. Evidence: with college readers of a textbook passage, author-provided organizers beat learner-built ones on transfer in three experiments, and retention did not differ (Learner-constructed graphic organizers are not shown to outperform provided ones: the one direct test, with college readers, favoured provided organizers on transfer [~M]); carried from reading to note-taking. How to fade the frame is modelled in Graphic Organizers.
  15. Do not ban laptops on the evidence; ask for paraphrase on any device. If typing produces transcripts, set the instruction "no full sentences from the speaker; your words only" and keep the pauses. Evidence: Laptop note-takers transcribed more verbatim and did worse on conceptual questions than longhand note-takers in one set of experiments, but a direct replication found no consistent difference in test performance [~M]; the replication found no consistent difference between longhand, laptop and even no notes, so the device is not the lever; verbatim copying is the suspected one.

When a condition is not met: if there is no time for review (step 6), say so and treat the notes as an encoding activity only; if learners cannot keep up even with guided notes, slow the stream or shorten segments before adding more note-taking; if the material is short and will be handed out in full, note-taking may not be worth its cost, and a recall task on the handout may do more.

Fitting the design to a situation

The three facts that most change the decision:

  • The learner's current notes on material like this. A learner who transcribes, or captures little, needs an outline, modelling and pauses; a learner whose notes already catch the main points needs only a purpose and a review routine. If unstated, ask: give a 5-minute segment and look at the notes. Do they hold the main points, in whose words, and is anything wrong?
  • Pace, mode and control of the stream. Live speech the learner cannot pause, a video they can pause, and a text they can reread need different supports. If unstated, ask: can learners stop or replay the input, and how fast and dense is it?
  • What the notes will be used for, and when. A recall test next week, a discussion tomorrow, a procedure on the job, and an essay combining several sources call for different notes and a different review. If unstated, ask: what will the learner do with these notes, how long after, and will they have them in hand?
If the brief says… Then change… Basis
Learners are advanced on the content, or already take selective notes Skip guided notes and modelling; state the purpose, keep one recall review; let them choose the format Instructional guidance that helps novices can become redundant or counterproductive as expertise grows. [~M] (review of cognitive-load studies; not tested on note-taking aids)
Learners are child (primary) or adolescent with little note-taking experience Guided notes with few, well-cued blanks; pauses every 5–8 minutes; teacher models notes on the board for several sessions, not one; review in class Guided Notes Improve Note Accuracy [+M] (meta-analysis reports the benefit "particularly for school-age students", abstract only); timings untested
Learners are adult in a work setting (training, briefings, meetings) Notes are for doing something: a one-page template with "what to do", "when", "who to ask", "one question"; review is applying the note on the next real task, followed by a check with a colleague or supervisor untested proposal
Setting is online-self-paced video Put the pause points into the video (a prompt screen at each break); give a downloadable outline for the first modules; end each module with a short recall quiz learners answer before reopening their notes Pauses and outline: untested here. Quiz: Retrieval Practice Improves Long Term Retention [+S] (carried to reviewing notes)
Mixed mode: some learners in the room, some remote, or some on a phone only Share the outline in a form that works on a phone (a short list of headings with blanks, not a wide table or PDF); say the pause aloud and post it in the chat; ask remote learners to post their one-line restatement in the chat so everyone sees some; send the slides afterwards but not before the session untested proposal
One teacher with a large lecture group Guided-notes outline for everyone; two or three pauses per hour with a restatement written to a neighbour or a poll; collect a random sample of notes after the first week to check for errors and address the commonest one in the next lecture Guided Notes Improve Note Accuracy [+M] (one quasi-experiment with two intact undergraduate classes); pauses and sampling untested
Scarce facilitator time (volunteer tutors, a busy shift, one brief check-in) Spend the facilitator's minute on one thing: a look at a learner's notes against the key points, naming one missing or wrong point; put the rest in the materials (outline, pause prompts, self-test questions at the end) untested proposal
A single session only Guided-notes outline, two pauses with a restatement prompt, and a 5-minute recall at the end from covered notes; do not plan the fade; do not claim retention untested proposal
Learners have low literacy, a learning disability, or are working in a second language Fewer and shorter blanks, with word banks or first letters; allow symbols, sketches and first-language words in notes; give the full slides or a transcript after the session so missed content is recoverable; allow audio recording where it is permitted untested proposal; the earlier page's "working-memory constraints" learners, and Hughes & Suritsky (1994, Key Sources, no claim page) on university students with and without learning disabilities
The brief names its own routine or template (Cornell notes, a school-wide organizer, a company note template) Keep it; change only what the evidence points at: add a review in which the learner covers the notes and recalls before checking, and one restatement prompt per section Retrieval Practice Improves Long Term Retention [+S] and self-explanation [+S], both carried to notes; no claim compares note formats
Learners take notes on laptops or tablets, or the brief asks whether to ban them Do not ban on this evidence; instruct paraphrase, keep pauses, and check a sample of notes for verbatim copying; allow devices for learners who write slowly or with difficulty Laptop note-takers transcribed more verbatim and did worse on conceptual questions than longhand note-takers in one set of experiments, but a direct replication found no consistent difference in test performance [~M] (replication found no consistent difference)
The goal is writing or argument from several sources Do not ask for a separate summary of each source; ask for notes organized by the question (a column per claim, with which source says what), then write from those Summarization Effective With Training [~M] (295 adults: summarizing each of five texts gave weaker integrated essays than rereading; carried from summaries to notes)
Stakes are high (a licensing exam, a safety procedure) Do not rely on learners' notes as the record: supply a checked reference version after the session, keep the learner's notes for recall practice against it, and correct errors before the assessment untested proposal
Notes will be shared with others (class notes posted for absent peers, meeting minutes for a team) Separate the two: a shared record written to a checklist and checked by a second person, and each learner's own notes kept for their own recall; do not treat reading someone else's notes as a substitute for taking and reviewing one's own untested proposal
Adults whose notes become a job aid (a handover, a shift, a procedure they will use again at work) Replace fill-in-the-blank guided notes with a reusable domain grid the learner keeps, built from the categories the brief or the workplace already uses (if none are given, for a handover: facts, actions due, what is uncertain, what to report), with a pre-printed prompt for any item whose omission is dangerous, modelled once with a think-aloud, then used on a real or recorded case; check it by having the learner brief a partner from the grid, not by reading back. In one session of about 75 minutes: 10 minutes modelling the grid on a short clip, 20 minutes filling it on the full case, 10 minutes comparing grids in pairs, 20 minutes briefing a partner from it (with sentence frames such as "The priority is… because…" for learners working in a second language), 10 minutes fixing the grid for reuse untested; the guided-notes claim [+W] tests teacher-made gapped notes with students, carried here only as "give the structure", not the worksheet form
The notes are a take-home aid, and the brief asks for an evaluation that does not feel like a test (older adults, community workshops, practical skills) Judge the notes by use, not recall: the learner follows their own notes, on their own device, to do the task again while a partner watches (no swapping unfamiliar devices); fix any step the partner cannot follow. Keep any retrieval as optional self-checking, not a quiz untested proposal

Six of the fourteen rows cite a claim; none of them tests note-taking against no note-taking. Guided notes (two rows) is the only direct note-taking test; the device row rests on a contested pair of studies; the rest carry a neighbouring finding (expertise reversal, retrieval practice, self-explanation, summarization) to note-taking and say so. Eight rows are untested proposals.

Observation, state and explanation

What can be observed is the notes themselves under stated conditions (the stream, its pace, the outline given, the prompts, the time allowed), what the learner does with them later (what they review, how, how long after), and performance on a probe. "Understands the lecture", "processed it deeply" and "is a weak note-taker" are inferred states. Complete notes show what the learner wrote down; they do not show that it was understood, and sparse notes do not show that it was not.

Response Plausible explanations Observation that could change the interpretation
Notes are long and close to verbatim Transcribing because the pace leaves no time to select; no criteria for what matters; a belief that complete notes are good notes; typing speed invites copying Slow the stream or add a pause and see whether the notes become selective; ask the learner to mark the three most important lines and explain why.
Notes are sparse Already knows the material; lost the thread; cannot write fast enough; is relying on slides arriving later Ask two recall questions on the segment without notes; ask what they would need to add for next week's task.
Notes are complete but recall a week later is poor Notes were never reviewed; reviewed by rereading only; the notes were copied without being understood Ask how and when they reviewed; give one recall review from covered notes and retest a matched item.
Notes contain a wrong point Misheard; misunderstood; copied a peer's error; the explanation was ambiguous Ask the learner to explain the point aloud; check whether others' notes carry the same error, which points to the teaching rather than the learner.
Guided-notes blanks are filled correctly, but the learner cannot explain the ideas The blanks were filled by matching words from slides; the outline did the selecting; the learner was listening only for blanks Ask for a one-sentence restatement of a section without the handout; compare with a section that had no blanks.
Learner stops taking notes after a few sessions Notes were not used in any later task; the format does not suit them; they gave up under the pace Ask what they do instead; make the next task need the notes and watch whether note-taking returns.

These rows are proposals, untested as diagnostics. A probe (a recall question, a request to explain) is itself practice and changes what it measures, so record each one and keep explanations open until an observation separates them.

Evidence and qualifications

  • Guided Notes Improve Note Accuracy [+M]: a meta-analysis of guided notes (instructor-prepared lecture handouts with blanks) concluded they increase note-taking accuracy and academic performance, particularly for school-age students (q4, read from the ERIC abstract; no pooled effect size could be verified). A quasi-experiment with 65 undergraduates in two intact classes found the guided-notes class produced better lecture notes at the delayed posttest and better learning at immediate and delayed posttests than an outline-only class (q2, abstract only, no effect size). This is the one claim that tests a note-taking support directly. It supports step 2 and the fading plan's starting point; it does not show how many blanks, how long to keep them, or how guided notes compare with learner notes once skill is in place. The claim's Discussion raises expertise reversal as a reason to withdraw the outline; that is reasoning, not a finding. Marker kept at M because both entries are abstract-only with no effect size.
  • Laptop note-takers transcribed more verbatim and did worse on conceptual questions than longhand note-takers in one set of experiments, but a direct replication found no consistent difference in test performance [~M]: three experiments with college students found laptop note-takers took more notes, transcribed more verbatim, and did worse on conceptual (not consistently factual) questions (q3, abstract only). A direct replication with two experiments and added eWriter and no-notes groups found no consistent difference between any of the groups, and smaller differences once students studied their notes; its internal meta-analysis found a small non-significant effect favouring longhand (q3, abstract only). This is the limiting claim for the page. It is the one recorded test that includes a no-notes group, and notes did not reliably beat no notes on the test used. It qualifies the model in two ways: the device is not shown to matter, and the assumed benefit of taking notes at all is not shown here. The claim's title still states the original finding; its own Discussion calls it fragile.
  • Summarization Effective With Training [~M]: 58 developmental community-college students randomly assigned to about 1.25 hours a week for five weeks of training in a text's hierarchical structure produced better summaries than an active-reading control (73% against 14% captured at least 70% of required information), with only a marginal effect on one-week delayed recall and no transfer to a standardized reading test (q2, unpublished conference paper). In a second study, 295 adults asked to summarize each of five texts wrote weaker integrated essays than adults who reread (q3). It supports teaching learners to condense (step 3) and warns against summary-per-source notes for synthesis tasks. Its title overstates the evidence (the second study runs against it), and it duplicates Training in text structure improves college students' summaries, but summarizing has not been shown here to improve recall or writing more than rereading or other active reading [~M], which carries the same two entries.
  • Prompting learners to self-explain improves understanding and problem solving on immediate tests by a moderate average amount, with little evidence yet on delayed or classroom gains. [+S]: a meta-analysis of 69 effect sizes found self-explanation prompts improved learning by g = .55 across task types, subjects and levels (q4, abstract only); a mathematics meta-analysis found small-to-moderate immediate gains with little evidence after a delay or in classrooms. It is the basis for asking learners to restate and explain in their notes (step 5). It is carried: no entry tests explanation written into notes, and the effect is for immediate tests.
  • Retrieval Practice Improves Long Term Retention [+S]: repeated testing beat repeated study at one week (61% against 40%) though not at five minutes; two meta-analyses put testing over restudy at g = 0.50 and 0.51, the advantage growing with delay. It is the basis for reviewing notes by covering and recalling (step 6). Carried: none of its entries uses learners' own notes as the study material.
  • Learner-constructed graphic organizers are not shown to outperform provided ones: the one direct test, with college readers, favoured provided organizers on transfer [~M]: in three experiments with college readers of a textbook passage, author-provided organizers produced better transfer than learner-built ones (d = 0.24 ns, 0.43, 0.84), with no difference in retention; a meta-analysis found constructing concept maps associated with a larger benefit than studying them, but against different comparisons. It qualifies graphic note-taking: building the structure oneself is not shown to be better, and may cost more than it gives for readers new to it. Carried from reading a passage to taking notes.
  • Higher compliance with MetaTutor's review-notes prompt is associated with larger proportional learning gains [+W]: one study of 28 college students in a computer-based learning environment; those who complied more with the prompt to review notes had larger proportional learning gains (ηp² = .16). Associational and small; learners who comply may differ in other ways. It is consistent with step 7 and does not test it.

Learners, materials and outcomes differ across these claims (school-age and undergraduate lecture notes, college readers, community-college summarizers, laboratory recall), so do not rank them by effect label. None measures a whole note-taking routine (purpose, outline, pauses, restatement, review) against ordinary lecture attendance.

The evidence by question

Further evidence, not yet read against this model

Claims this page cited before it was rewritten, and others found while rewriting it. None has been read against the model above as closely as the claims in the previous section.

Objective and learner-valued goal

The designer's objective is usually performance on a named later task (a test at a stated delay, a discussion, a written product, a procedure) on the material the notes cover. Note quality is a means, not the outcome: complete notes the learner never uses meet no objective. Elicit separately what the learner wants from their notes: something to revise from before an exam, a record so they need not listen closely, a way to stay awake, a reference for the job. Agreement and divergence are both observations. A learner who values a complete record may transcribe, and will resist pauses for restatement unless they also get the slides; a learner who values passing the exam will take up recall review once they see it raise their scores on a practice quiz.

For example, an adult in a workplace induction may want a checklist they can use on their first solo shift, while the designer's objective is recall of the safety policy on a quiz. Notes built around the quiz headings serve the designer; a one-page "what I do, when, who I ask" built from the same session serves the learner. Record both, and check the second on the shift, not the quiz.

What would revise this model?

The expectation should weaken if studies with the device held constant find that notes written in the learner's own words do no better than verbatim notes on delayed or conceptual tests, or if more studies like the replication above find no reliable advantage of taking notes over not taking them once both groups study the same material. If guided notes improve note accuracy but not later performance, the outline is a record-keeping aid and its place in the model should shrink to that. If recall review from one's own notes does no better than rereading them, step 6 should go. If learner-built notes, after training, beat supplied outlines on transfer, the fading plan should move faster. Do not protect the model by calling every failure "poor notes" or "not reviewed" after the fact.

Note completeness, note accuracy, immediate performance, delayed retention, transfer and use in the task the learner cares about are separate claims. The present evidence establishes no blank density, pause interval, review delay or hand-over criterion for note-taking.

Claims

  • Graphic Organizers — graphic note formats can make structure and relationships more visible.
  • Handouts/Online Guides/Visual Reading Aids — guided notes and visual cues can improve note quality.
  • Journaling — both practices externalize thinking, but note-taking is usually more selective and task-focused.
  • Pre-Reading Questioning — questions can shape what learners attend to and record.
  • Annotating — marking the text itself while reading; this page holds a separate record, most often of speech or video.
  • Chunking — where to place pauses for note-taking: at the boundaries of meaningful units.
  • Self-Explanation — the restatement and "why" prompts that turn notes from copying into explanation.
  • Retrieval Practice — reviewing notes by covering and recalling rather than rereading.
  • Clear Structure — a guided-notes outline is one way of making the organization of a session explicit.
  • Self-Monitoring — the "one question" at each pause is a check on one's own understanding during the task.

Examples

Key Sources

  • Makany, T., Kemp, J., & Dror, I. E. (2009). Optimising the use of note-taking as an external cognitive aid for increasing learning. British Journal of Educational Technology, 40(4), 619-635. https://doi.org/10.1111/j.1467-8535.2008.00906.x
  • Robinson, D. H., Katayama, A. D., Beth, A., Odom, S., Hsieh, Y. P., & Vanderveen, A. (2006). Increasing text comprehension and graphic note taking using a partial graphic organizer. The Journal of Educational Research, 100(2), 103-111. https://doi.org/10.3200/JOER.100.2.103-111
  • Hughes, C. A., & Suritsky, S. K. (1994). Note-taking skills of university students with and without learning disabilities. Journal of Learning Disabilities, 27(1), 20-24. https://doi.org/10.1177/002221949402700105
Page metadata
Field Value
type principle
id note-taking
title Note-taking
description For a learner meeting a dense stream of new material they will need later, selective notes restated in their own words, supported by a teacher-supplied outline that is withdrawn as skill grows and reviewed by recalling from them, are expected to beat no notes or verbatim notes; one claim tests guided notes, and none tests note-taking itself against no notes.
status review
canonical True
generated by claude/unspecified on 2026-10-05
trust unverified — no one has confirmed this page's content
sources 3 — makany-2009, robinson-2006, hughes-1994

Frontmatter is the page's machine-readable record — see the Schema & Guide.