How to Read an Academic Article Efficiently: The Three-Pass Method, With a Reading Log

November 6, 2026 · Study skills

A reading list of twelve journal articles can feel impossible in a week, and many students respond in one of two ways: they read every word of the first two articles and run out of time, or they skim everything and remember nothing. Both are passive. Researchers read dozens of papers a month, and the ones who manage it do not read faster; they read selectively, in stages, with a question in mind. This guide adapts one well-known method for students and adds a simple way to record what you read.

Know how an article is built

Most empirical articles in the sciences and social sciences follow the same skeleton, often called IMRaD: Introduction (the problem and why it matters), Methods (what the authors did), Results (what they found) and Discussion (what it means, and its limits). Around it sit the title, the abstract, the conclusion and the reference list. Humanities articles are usually built as a continuous argument instead, but they still have an introduction that states the claim and a conclusion that sums it up.

Knowing the skeleton lets you go straight to the part that answers your question. If you need to know whether a finding is solid, the methods matter most. If you need the big picture for an introduction, the abstract, the first pages and the discussion may be enough.

The three-pass method

In a short paper called "How to Read a Paper", published in 2007 in ACM SIGCOMM Computer Communication Review, the computer scientist Srinivasan Keshav describes the approach he had used for years. The key idea is to read a paper in up to three passes instead of ploughing from the beginning to the end, with each pass building on the one before. It was written for researchers, but suits any long reading list.

Pass 1: the ten-minute triage

Keshav's first pass is a quick scan that should take about five to ten minutes:

1. Read the title, the abstract and the introduction carefully.
2. Read the section and subsection headings, and nothing else in the body.
3. Read the conclusions.
4. Glance over the references, noting any you already know.

At the end of the pass you should be able to answer what he calls the five Cs: category (what kind of paper is it?), context (which other work and theories is it related to?), correctness (do the assumptions look valid?), contributions (what does it add?) and clarity (is it well written?). For a student, the most useful outcome is a decision: keep, park or drop. Many articles on a long list only need this first pass.

Pass 2: the one-hour read

The second pass is for articles you will actually use. Read with more care, but skip the fine detail such as proofs or the full statistical models. Keshav suggests looking closely at figures, diagrams and graphs, checking for example whether axes are labelled and results come with error bars, and marking unread references that look worth following up. He puts this pass at up to an hour. At the end, you should be able to summarise the main thrust of the article, with its supporting evidence, to someone else.

That last test is the one to use. Close the article and explain it aloud, or in three sentences on paper. If you cannot, you have not finished the second pass, however long you have spent on it.

Pass 3: the deep read

The third pass is for the few articles at the centre of your project: the study your dissertation builds on, or the paper you are asked to critique. Keshav describes it as trying to virtually re-create the work, making the same assumptions as the authors, so that you can spot both its innovations and its hidden weaknesses. He estimates it can take four or five hours for a beginner and about an hour for an experienced reader. Most undergraduate essays need this level of reading for two or three sources at most.

When an article feels too hard

Keshav lists the usual reasons a paper stays unclear after the second pass: the subject is new to you, the terminology and acronyms are unfamiliar, the authors use a technique you do not know, the paper is badly written, or it is simply late and you are tired. Some practical responses:

• Read the background first. A textbook chapter, a review article or your lecture notes on the topic will make the specialist article far easier.
• Keep a glossary. Write down each unfamiliar term with a one-line definition in your own words.
• Read the discussion before the results. The authors' own interpretation can show you what the tables are meant to demonstrate; then check whether the tables really show it.
• Ask a specific question. "What does this regression show?" gets a better answer from a tutor or a study group than "I don't understand this paper".

Read actively, not just faster

The UNC Learning Center's guide to reading effectively describes two common passive habits: reading every word while copying pages of notes, and casually skimming. Neither, it says, leads to a high level of comprehension or retention. Active reading means making your brain do something with the text: previewing it, asking questions before you start, and pausing to summarise as you go.

The oldest version of this idea is SQ3R, introduced by the psychologist Francis P. Robinson in his 1946 book Effective Study: Survey, Question, Read, Recite, Review. The first pass of the three-pass method is essentially the survey step; turning headings into questions before you read gives the second pass a purpose. If you highlight, write a word in the margin saying why.

Keep a reading log

The most common reading mistake is not slow reading but forgetting: three weeks later, you remember that an article was relevant but not why. A short log for every article you keep solves this. Use the same template each time:

Write the findings in your own words and keep quotations clearly marked, so you never mistake a copied phrase for your own sentence later. The same template feeds directly into an annotated bibliography, and our guide to note-taking methods explains how to organise notes from lectures and readings together.

A reading plan for a week

Say you have ten articles to get through for a seminar or an essay. A realistic plan looks like this:

• Session 1 (about 90 minutes): first pass on all ten, about eight minutes each. Sort them into keep, park and drop.
• Sessions 2 to 4 (about an hour each): second pass on the four or five articles you kept, one or two per session, with a reading log for each.
• Session 5: third pass on the one or two articles at the heart of your argument.
• Last session: reread your logs, not the articles, and draft your plan.

Spread the sessions over the week rather than in one long day; shorter blocks keep your attention fresher. Our guide to time management for students shows how to fit them around lectures and work.

A note on AI summaries

AI tools can produce a summary of an article in seconds, and some students use them as a first pass. Treat any such summary as unverified: tools can misstate findings or invent details, and a summary cannot tell you whether the methods are sound. Check every point against the article itself before you rely on it, never cite a source you have not read, and follow your institution's rules on AI use, which vary from course to course.

Sources

• Keshav, S. (2007). How to read a paper. ACM SIGCOMM Computer Communication Review, 37(3), 83–84. PDF.
• The Learning Center, University of North Carolina at Chapel Hill, Reading Textbooks Effectively.
• Robinson, F. P. (1946). Effective Study. Harper & Brothers.

More study skills guides