How to write an abstract

The abstract is the most-read 250 words of your paper — often the only ones. It is also a genre with rules: five moves, in order, no suspense. Here is how to write one that gets the paper read, cited, and accepted.

The five moves of every abstract

Across fields and formats, effective abstracts make the same five moves in the same order: context (one sentence on why the topic matters), gap or question (what is unknown or contested), method (what you did, with the design and sample), results (what you found — with the key numbers), and implication (what it means for the field or practice). In 250 words that is roughly one to two sentences per move, with results getting the largest share.

The most common failure is writing a table of contents instead of a synopsis: "We discuss X and present findings on Y" tells the reader nothing. Abstracts state findings — "X increased Y by 23% (p < .001)" — because readers, reviewers, and search engines all decide from the abstract whether the paper contains anything.

Structured, unstructured, and conference formats

Medical and health journals mostly require structured abstracts with labeled sections (Background / Methods / Results / Conclusions — check the venue, the headings vary); most other fields use a single unstructured paragraph making the same moves invisibly. Conference abstracts are their own genre: often submitted before the work is complete, judged competitively, and sometimes the only thing ever published — so they run more promotional, with the contribution stated harder and earlier.

Write the abstract last, from the finished paper — the version written before the results are in always shows it. For structured formats, draft each labeled section from the corresponding section of the paper; numbers in the abstract must match the text exactly, a detail copy editors and reviewers check.

Write for search engines too

Abstracts are indexed by Google Scholar and every database; the title and abstract are effectively your paper’s SEO. Use the field’s standard terminology (the words people search, not your novel coinage), get the key concept into the title and the first two sentences, and choose keywords that complement rather than repeat the title — that combination maximizes the queries the paper can match.

Spell out an abbreviation at first use even in the abstract, keep citations out of it (most venues forbid them), and stay inside the word limit — submission systems truncate or reject, and a truncated abstract is what gets indexed.

A pre-submission checklist

Thirty seconds against this list catches most abstract problems:

  • All five moves present, in order — context, gap, method, results, implication.
  • At least one concrete result with numbers, not "results are discussed".
  • Numbers identical to those in the paper.
  • No citations, no undefined abbreviations, no "this paper will…" (write in past/present, not future).
  • Within the word limit with a few words to spare.
  • The first sentence works standalone — it is what shows in search results.

Where Kahubi fits

Kahubi drafts abstracts from the finished manuscript, not from thin air: the agent reads your paper, pulls the actual numbers from your results, follows the five-move structure (or your venue’s labeled sections), and writes in the voice learned from your publications. Word limits are enforced at draft time, and the abstract updates when the results do.

Last updated 2026-08-07.

Frequently asked questions

How long should an abstract be?
Journal abstracts commonly cap at 150–300 words (250 is the most frequent limit); conference abstracts range from 150 words to two pages depending on the venue. The limit in the author guidelines wins — submission systems enforce it mechanically.
What is the difference between an abstract and an introduction?
The abstract is the whole paper in miniature, including results and conclusions; the introduction motivates and frames but reveals no findings. Reusing introduction sentences in the abstract wastes its scarce words on preamble.
Can I submit a conference abstract before the results are final?
Fields differ: in much of medicine and psychology, preliminary-results abstracts are normal and expected to firm up by the meeting; in others, claims you cannot yet support invite awkward questions. Never state as found what you have not found — frame it as "preliminary analyses indicate" and update before presenting.

Related

Abstracts drafted from the finished paper

Real numbers, five moves, your voice, inside the word limit. Free plan included.