How to Summarize an Academic Paper

What is a summary — and what isn't

A summary condenses a source into its essential argument, evidence, and conclusions. It is written in your own words, in the third person, and it describes what the paper says without evaluating whether it is right.

A summary is not a book report ("this paper was really interesting"), a paraphrase of a single passage, or an opinion piece. Evaluation goes in a separate critique or annotation. Keep the two jobs distinct.

The IMRAD structure

Most empirical papers — in the sciences, psychology, medicine, and economics — follow the IMRAD structure: Introduction, Methods, Results, and Discussion. Knowing the structure tells you where each piece of a summary lives.

Humanities and theoretical papers rarely follow IMRAD — but they usually still have an argument, evidence, and a conclusion, just under different section headings. The summary principles below work for both.

Read in the right order, not cover to cover

Reading an academic paper from first word to last is slow and inefficient. Use this order instead:

1. Title and abstract. Roughly 80% of what you need is in the abstract. Read it twice.
2. Introduction — first and last paragraphs only. The first paragraph sets up the problem; the last usually states the paper's contribution.
3. Discussion — first two paragraphs. The authors restate their main finding and its significance at the top of the discussion.
4. Results — figures and tables. Read captions first, then skim the text for the effect sizes behind the figures.
5. Methods — only if a specific design detail matters. For a summary, a single sentence ("a cross-sectional survey of 412 adolescents") is usually enough.
6. Limitations — at the end of the discussion. Often the most important paragraph for a reader deciding whether to trust the paper.

The five-sentence template

For reading notes, annotated bibliographies, and most literature review entries, five sentences are enough. Each sentence has one job.

1. Topic and question. What is the paper about, and what question does it ask?
2. Approach. How did the authors investigate the question?
3. Main finding. What did they discover?
4. Interpretation. What do the authors claim that finding means?
5. Relevance. Why does this paper matter — to the field, or to your own project?

Notice what the summary does not include: the reviewers' names, the funding source, the year the data were collected, or any of the ancillary analyses. A summary preserves the spine of the paper, not its skeleton.

Summary length by context

Summaries in annotated bibliographies

An annotated bibliography entry has a specific shape: a full citation, followed by a paragraph that usually combines a summary and a short evaluation. The summary portion should cover question, method, and finding; the evaluation portion adds a sentence on strengths, limitations, and relevance to your project.

Summarising theoretical and review papers

Theoretical papers, review articles, and humanities scholarship don't fit neatly into IMRAD. Replace "methods" and "results" with argument and evidence:

1. Claim. What is the paper's central argument?
2. Evidence. What kinds of evidence does it bring — textual analysis, prior studies, historical sources?
3. Framework. What theoretical frame organises the argument?
4. Contribution. How does it revise or extend previous scholarship?
5. Relevance. Why does it matter to your project?

The shape is the same — five sentences, third person, author's voice not yours. Only the content categories change.

Common mistakes

Summarising section by section

"The introduction says X. The methods say Y. The results say Z." produces a summary that is both long and uninformative. Summarise the paper's argument, not its sections.

Copying the abstract

The abstract is a summary written by the author. Copying it is plagiarism; lightly rewording it is patchwriting. Read the paper, then write the summary without the abstract in front of you.

Mixing in your own opinions

A summary reports the authors' position. Your agreement or disagreement belongs in a separate evaluation paragraph, not inside the summary itself.

Losing the effect size

"The intervention helped" is not a summary — it's a conclusion without evidence. Where the original paper reports an effect size, percentage change, or odds ratio, keep one concrete number in your summary.

Over-hedging

"The authors seem to suggest that maybe…" signals that the summariser has not understood the paper. If a claim is hedged in the original, echo the hedge; if it isn't, don't invent one.

Quick summary

A strong academic summary is a compressed argument, not a section walkthrough. Once you can answer "what did they do, what did they find, and what does it mean?" in three clean sentences, the rest of the summary is just wrapping.