Cornell Notes Cornell Notes: A Page That Turns Into a Quiz.

The page, region by region

Everything in the method follows from the page, so it is worth getting the proportions right before the first lecture. The figures here come from the handout Cornell's Learning Strategies Center gives its students.

Two details matter more than they look. First, the two columns add up to 8.5 inches, the width of a US Letter sheet, so the cue column takes a little under a third of the page. That is wider than a margin and narrower than the notes, which is the point: enough room for a question, not enough to rewrite the lecture. Second, the summary is per page, not per lecture. The handout puts the space at the bottom of each page for the notes on that page, so a lecture that fills four pages gets four summaries.

Some printed templates add a strip across the top for the course, the date and the topic. The handout does not show one, but it costs nothing and makes a pile of loose pages much easier to sort. If you use A4 paper rather than US Letter, keep the proportions rather than the inches: roughly 30 percent of the width for the cues and the rest for the notes.

The five steps, as Cornell teaches them

Cornell's handout names five steps. Most explanations stop at the layout, but the steps are where the learning happens, and three of the five take place after the lecture is over.

1. Record, during the lecture. Write in the note-taking column only. The handout asks for "telegraphic sentences": short and stripped of filler words, but still sentences rather than single words, so that they make sense a week later.
2. Questions, as soon after class as you can. Read your notes and write questions about them in the cue column, level with the notes they refer to. The handout gives its reasons: writing questions makes you work out what the notes mean and how the ideas relate, keeps one lecture connected to the next, helps you remember, and leaves you with exam questions ready for later.
3. Recite, when you study. Lay a sheet of paper over the note-taking column. With only the cues in view, answer each one out loud and in your own words, then uncover the notes and check. This is the step that turns the page into a test.
4. Reflect. Ask yourself bigger questions about the material. The handout's examples include what the facts signify, what principle they rest on, how you could apply them and how they fit with what you already know.
5. Review, every week. Spend at least ten minutes each week going over all your previous notes, not just the newest page. The handout's promise is modest and specific: regular review keeps much of the material available both now and at exam time.

You will find other versions of the five steps online, with different names or a different order. This page follows Cornell's own handout, which says it is adapted from the 2001 edition of How to Study in College. Cornell's center also teaches the method in the two short videos above, and in a publicly available online module linked from its Cornell notes page.

A worked example: one lecture, start to finish

Here is one page for an introductory psychology lecture on memory and testing. The content is real: it comes from two studies this guide cites later, so the example doubles as a preview of why the method works.

That is what the right-hand column looks like at the end of the hour: short sentences, no paragraphs, a few numbers. The same evening, the cue column gets filled in, one question per cluster of notes, written level with the notes it refers to.

Two things to notice. Every cue is a question with a checkable answer, so covering the right column and answering the left is a genuine test rather than a glance at a list of topics. And the reflection asks questions the lecture did not answer. That is the point of the fourth step: it pushes past what was said toward what it means for you, which here is a change in how to study for Friday.

A template you can copy

You do not need special paper or a special app. Here are three ways to set up the page, from a notebook to a plain text note.

On paper

1. Rule a vertical line 2.5 inches (about 6.4 cm) from the left edge, from the top of the page down to where the summary will start.
2. Rule a horizontal line 2 inches (about 5 cm) above the bottom edge, across the full width of the page.
3. If you like, add a line at the top for the course, the date and the topic.
4. Rule several pages before class. Drawing lines while the lecture starts is how people give up on the format in the second week.

In a document or a notes app with tables

Insert a two-column table with the left column at roughly 30 percent of the width, which matches the 2.5 to 6 ratio on the handout. Add one full-width row at the bottom for the summary. Copy this structure:

As a plain text note

If your notes app has no tables, put the cues first and the notes below them. That order makes self-testing easy: you read the questions at the top without the answers in view, and scroll down to check.

• Title: Course · Lecture topic · Date
• Cues: one question per line, added after class
• Notes: short sentences, one idea per line, written in class
• Summary: two to four sentences, added after class
• Reflect (optional): what this means, how it connects, what you will do differently

On a laptop, a tablet or in a notes app

Cornell notes were designed for paper, and Cornell's own center currently tells students that handwriting beats typing. The evidence for that is less settled than the advice suggests, and the method works on a screen if you keep its intent.

The handwriting question

The study usually cited is Mueller and Oppenheimer (2014). Across three studies, students who took notes on laptops did worse on conceptual questions than students who wrote by hand, and the authors traced the gap to laptop users transcribing lectures word for word instead of reframing them in their own words. A direct replication by Morehead, Dunlosky and Rawson (2019) saw some trends in the same direction, but performance did not consistently differ between the groups, including a group that took no notes at all, and a meta-analysis of the replications found small, non-significant effects favoring longhand. The authors concluded that it is premature to say which method is better.

The useful part of the first study is the mechanism it points to: verbatim transcription. Typing makes it easy, and the Cornell format pushes against it, because a note-taking column full of transcript leaves nothing to turn into questions. If you type, type in short sentences and paraphrase on purpose.

On a laptop

A two-column table in any document editor reproduces the page, but in a narrow window the cue column gets cramped. Two arrangements work better on screens. One alternates blocks: a question in bold, then the notes that answer it. The other, from the template above, puts all the questions at the top and the notes below. For self-testing the second is easier, because the answers stay out of view until you scroll.

On a tablet with a stylus

A tablet with a stylus is the closest thing to paper. You can use a ruled Cornell page as the background of a notebook and write by hand, and covering the notes is as simple as laying a sheet of paper, or your hand, over the right side of the screen. The trade-off is search: unless your app recognizes handwriting, the pages are harder to search than typed text, which matters when you are hunting for one idea across a term.

In a notes app

Most notes apps have no column layout, and that is fine. The method needs three things: a place for notes taken in the moment, a separate place for questions written later, and a summary. Headings do the job. What a notes app adds is reach across lectures: a search that finds every page where a concept came up, so you can study a topic rather than a date. That is the part paper cannot do, and it is the main reason to keep a study setup on your Mac even if the pages themselves start on paper.

Reviewing with the cue column

The cue column is a stack of flashcards that writes itself as you go. How you use it decides most of what the method is worth.

The research case for this step is the strongest in the whole method. In Roediger and Karpicke's 2006 study, students read prose passages and then either restudied them or took free-recall tests without feedback. On a final test five minutes later, restudying came out ahead. Two days or a week later, the students who had been tested retained substantially more, even though restudying had made students more confident. A 2013 review by Dunlosky and colleagues assessed ten common study techniques and gave its high utility rating to two: practice testing and distributed practice. Reciting from the cue column is a form of practice testing, and the weekly review spreads it out over time.

A recite session, step by step

1. Cover the note-taking column. A sheet of paper, a folded page or your hand. On a screen, scroll so that only the questions show.
2. Answer each cue out loud, in your own words. Saying a full answer is harder to fake than nodding at a page. A vague sense of knowing rarely survives being said aloud.
3. Uncover and check, one cue at a time. Mark the ones you missed or only half answered. A small dot next to the cue is enough.
4. Go round again on the marked cues only. The ones you answered well can wait until next week.
5. Finish by reading the summary. It puts the answers back into the shape of the lecture.

For the weekly review, start with the oldest pages. The handout asks you to review all your previous notes, and the old pages are the ones most at risk of being forgotten by the exam. If you only read cues and check answers, skipping anything you get right, ten minutes covers more pages than you might expect.

Variations for reading, meetings and video lectures

Pauk built the page for live lectures. The same split, notes now and questions later, carries over to other kinds of input with small changes.

Reading a textbook or a paper

Do not take notes on every paragraph as you read; that drifts into copying. Read a section, look away, and write what you remember in the note-taking column. Then look back and fill the gaps. That first attempt to write from memory is already a small act of recall. Write the questions when you finish the chapter, and a summary at the bottom of each page as usual. Page numbers next to each cue make it easy to go back to the source.

Meetings

In a meeting, the note-taking column holds the discussion. Afterwards, the cue column holds what you need to act on rather than study: decisions, owners, deadlines and open questions. The summary is the line you would send to someone who missed the meeting. This version changes the most from Pauk's design, because you rarely need to recall a meeting from memory; you need to find what was decided. If your meeting notes live in Mindly, the meeting notes workflow shows how they stay searchable.

Online and recorded lectures

Recorded lectures remove the pressure the method was built for: you can pause. That sounds like an advantage, and it becomes a trap when pausing turns into transcribing. A better rhythm is to watch a segment of a few minutes without writing, pause, and write the notes for that segment from memory. Put the timestamp in the cue column next to each block, so a question you cannot answer later points straight to the minute where the answer is. If the platform offers a transcript, treat it as a reference to check against, not as your notes.

What research says, and the honest limits

The research splits into two kinds: studies of the techniques the method is built from, which are strong, and studies of the Cornell format itself, which are fewer, smaller and mostly classroom-based.

The parts are well supported

Recite and Review line up with the two techniques Dunlosky and colleagues rated high utility, practice testing and distributed practice, which they found benefit learners of different ages and abilities across many kinds of test, including in real classrooms. The same review rated summarization low utility: it helped some students on some tasks, but only under limited conditions. That is a useful corrective for the summary band. It is worth writing, as an index and a check on your understanding, but it is the part of the page with the weakest evidence as a study technique on its own.

The whole system, less so

The studies we found that test the Cornell format directly mostly compare it with regular instruction or with taking no notes, and report gains. The more telling result is what happens against another structured method. In Ahmad (2019), three classes of first-year secondary school students learning English as a foreign language spent 12 weeks with regular instruction, with Cornell notes, or with a different note-taking strategy called REAP. Both note-taking groups scored significantly higher on critical reading than the control group, and there was no significant difference between the two. One study is not a verdict, but it fits the wider picture: structure and processing help, and the Cornell layout is one good way to get them, not the only one.

Handwriting versus typing

As covered above, the longhand advantage reported in 2014 did not hold up consistently in the 2019 replication. A reasonable reading of the two together: how you process the lecture matters more than the tool, and verbatim transcription is the habit to avoid on any device.

Where Cornell notes fall short

• Problem-heavy subjects. In mathematics, physics or programming, much of a lecture is worked examples and derivations. They fit awkwardly into short sentences, and the useful cue is often "solve this" rather than a question. One workable adjustment is to keep the cues and the summary and give the notes column over to worked problems.
• It costs time after class. Three of the five steps happen outside the lecture. The method assumes you will spend that time, and it gives little back if you do not.
• Pages do not connect. Each page stands alone. A concept that comes up in week 2 and again in week 9 lives on two pages with nothing linking them. Methods built around small, linked notes, such as atomic notes, handle that better.
• The geometry fights small screens. A fixed 2.5 to 6 split suits a sheet of paper. On a phone or in a narrow window it usually has to become a sequence of headings instead.
• Fast, dense lectures. When the lecturer moves faster than you can condense, the notes column fills with fragments. The practical answer is to write less in the moment and fill the gaps from the slides or a classmate soon afterwards, while you still remember the lecture.

After the page: progressive summarization and AI

A Cornell page is built to get you through a course. Some of what it holds is worth keeping far longer, and that is where it meets other methods, and where the question of AI comes up.

How it connects to progressive summarization

Tiago Forte's progressive summarization compresses a note in layers each time you come back to it: bold the best passages, highlight the best of the bold, then summarize in your own words. A Cornell page already has two such layers built in. The note-taking column is a first compression of the lecture, and the summary band is a summary in your own words, written the same day. The difference is timing. Cornell does all of its distilling in the week of the lecture, with an exam in view; progressive summarization does it gradually, and only when a note proves useful again.

The two work well in sequence. Use Cornell notes while a course is running. When it ends, take the few pages you expect to need again, for a thesis, a later course or your job, and carry their summaries and best cues into your long-term notes, where progressive summarization can keep refining them. The rest can stay in the binder. If you are setting up that long-term system now, our page for students shows how the pieces fit.

Can AI do this?

Partly, and the part it can do is the part that matters least. An AI model given a lecture transcript can produce a tidy summary and a list of questions in seconds. What it cannot do is the remembering. The point of the cue column is not to own a list of questions; it is the moment you cover the notes and find out what you can produce on your own. In the retrieval research above, the benefit came from students doing the recall themselves. A list of questions you never answer from memory is just another document to reread.

Writing the questions is part of the learning too. Cornell's handout gives its reasons for the second step, that writing questions makes you work out meanings and relationships and strengthens memory, and an AI that writes them for you skips that work. The summary is a different case. It is the region with the weakest research support, so handing it to an AI costs the least. Even then, AI summaries leave things out, often the detail you will later need; our piece on what gets compressed away in AI summaries looks at how.