The Purpose of an Abstract
Of everything in your paper, the abstract gets read most, and for many readers it is all they read. It turns up in search results, indexing databases and reading lists, typically detached from the full text. Treat it as the face of the paper when you write it.
This detachment explains all the guidance below. An abstract has to be self-sufficient: it should have no citations, no abbreviations left undefined, no mention of "Table 2", and no sentence that relies on the reader having seen the introduction.
Its purpose is to help a reader judge, within a minute, whether the paper deserves reading. It is not a teaser, and holding back the finding to build curiosity just drives readers away.
Required Contents
- The problem — the question the work tackles and the reason it matters
- The approach — what you did, described well enough for its suitability to be judged
- The principal findings — the real results, including the key figures
- The implication — the consequences for the field or for practice
The most frequent flaw is an abstract that sets out the problem and the method but leaves out the findings. Readers come to learn what you discovered.
A Recipe, Sentence by Sentence
- Context — a single sentence that sets the area and the problem.
- Gap — a single sentence on what remains unknown or unsettled.
- Aim — a single sentence saying what the study intended to achieve.
- Method — a sentence or two covering design, data and analysis.
- Results — two or three sentences giving the real findings and figures.
- In Summary — one sentence on what it means.
- Implication — an optional sentence on what ought to follow.
Leave it until the end, once the paper is complete. An abstract written before the results are fixed nearly always promises something the paper fails to deliver. Revisit it after your last edit to the body, so that the two still agree.
Structured vs Unstructured
A structured abstract carries explicit labels — Background, Methods, Results, Conclusions — and is frequently seen in health and social science. An unstructured abstract is one paragraph that covers the same content with no labels, and is typical of engineering, computing and the humanities.
Follow the format named by the target journal. What the abstract must cover is the same either way; just the signposting is different.
Length and Language
The usual range set by journals is from 150 to 300 words. Look at the submission portal for the requirement in force, instead of trusting a figure you saw somewhere else.
- Put what you did and found in the past tense, and general truths in the present tense
- Choose the active voice whenever it makes the sentence shorter
- Give the full form of an abbreviation when it first appears, or do without it
- Cut the warm-up: "This paper aims to investigate" can generally be replaced by the finding itself
The Parts of a Research Paper
Usual proportions for an empirical paper — guidance rather than a fixed rule.
- BackgroundOutline the gap in 1–2 sentences.
- ObjectiveHypothesis or question under investigation.
- MethodsDesign plus data plus tools used.
- ResultsImportant findings, given with numerical values.
- In SummaryWhat it means.
- Key Terms4–6 searchable terms.
An abstract that holds all five parts needs roughly 150–250 words, and writing it last makes that easiest. Full rules: JMVRA Publication Guidelines.
Mistakes to Avoid
| Mistake | Fix |
|---|---|
| No results reported | State the real findings along with figures |
| Citations within the abstract | Take them out; the abstract must be self-sufficient |
| Undefined abbreviations | Spell out or omit |
| Mentions of figures or tables | Delete them; readers may lack the full text |
| Claims more forceful than the paper supports | Keep the abstract in line with what the evidence backs up |
| Background taking up half the words | A sentence or two of context suffices |
Worked Example
Weak: "This paper aims to study the important problem of energy consumption in buildings. A method is proposed. Results are presented and discussed. The proposed approach is found to be effective."
None of these sentences says anything: there is no setting, no method, no numbers and no finding.
Stronger: "Heating accounts for a large share of energy use in older institutional buildings, but retrofit decisions are usually made without measured baseline data. We instrumented four university buildings over two heating seasons and compared measured consumption against the models used to justify their retrofits. Measured savings were consistently below modelled projections, with the gap widening in buildings with mixed occupancy patterns. Retrofit appraisals that rely on modelled savings alone may overstate returns, and we recommend a measured baseline period before commitment."
Final Checks
- Understandable by itself, without pointing to the full text
- Gives the findings themselves rather than just noting that there are some
- Inside the word limit set by the journal
- Free of citations, undefined abbreviations and figure references
- Each claim backed up in the paper's body
- Keywords are picked separately — see choosing keywords
At-a-Glance Checklist
- Self-sufficientClear without the paper at hand
- Includes figuresMain results given, not merely hinted at
- Citation-freeAnd no abbreviations left undefined
- Agrees with the paperNothing appears here that the text lacks
- Inside the limitWord count as set by the portal
What Abstracts Stress in Different Fields
| Field | Emphasis |
|---|---|
| Engineering | Method, performance results |
| Health sciences | Structured headings, outcomes |
| Social sciences | Context, sample, implications |
| Humanities | Argument and contribution |
Abstract and Keywords as a Pair
- Place key terms in the opening two sentences — this helps search
- Avoid pointless keyword repetition — put fresh terms in the keyword list
- Echo the title — use the same terms
- Say it out loud — it should be clear on its own
Polishing Your Abstract Until Every Sentence Earns Its Place
An abstract is often the only part a busy reader sees. These editing habits help you refine a draft so it stands alone and represents the full paper honestly.
- Write it last. Draft the abstract after the paper is complete, so it reflects what you actually found rather than what you originally planned to find.
- Test it in isolation. Give the abstract alone to a colleague from a neighbouring field and ask what the study did and found. Their confusion shows exactly which sentence needs work.
- Prefer concrete verbs. Replace vague phrases such as 'is discussed' with active statements about what the study measured, compared or demonstrated. Specific wording signals confidence and saves space.
- Keep outside material out. Do not cite references, introduce new abbreviations or include results absent from the body. The abstract must be consistent with the paper in every respect.
- Choose keywords from reader language. Select terms people would type when searching for your topic, including a common alternative spelling where relevant, rather than invented or overly narrow labels.
Write the abstract after the paper, let an outsider test it and remove anything the main text does not support.
Questions Researchers Often Ask
What length should an abstract have?
Journals mostly request 150 to 300 words. For JMVRA, the limit stated on the submission portal is the one that counts, and it overrides any general figure.
Should the abstract be structured or unstructured?
Do what the journal asks. A structured abstract is divided under headings (Background, Methods, Results, Conclusions), while an unstructured one handles the same moves within a single paragraph.
Is it allowed to cite references in an abstract?
It is best avoided. Abstracts are read by themselves — in databases and search results — and there the references cannot be followed.
Do abstracts need citations?
Generally not. An abstract ought to be complete without references.
Are numbers allowed in an abstract?
Yes; main results stated with figures are often valuable.
More Researcher Resources
This guide belongs to the JMVRA Researcher Resources library, which offers practical help that authors and researchers can check, free of promotional claims.
Ready to Publish? Submit to JMVRA
JMVRA is aligned with UGC's Suggested Parameters for Choosing Peer-Reviewed Journals.
JMVRA is a Transparent, Peer-Reviewed, Open Access Journal (Refereed Journal). Following the discontinuation of the UGC and UGC-CARE journal listing, and in alignment with the latest UGC-CARE recommendations and guidelines, JMVRA is aligned with UGC’s Suggested Parameters for Choosing Peer-Reviewed Journals.
Where to go next on jmvra.org
Key Terms
Subjects and search phrases this article addresses.
Related Reading
Further reading on the same theme from the JMVRA blog.
Indexing a New Journal: The Realistic Order
Read the guide RelatedMake Your Research Paper Searchable in Google Scholar: A Stepwise Guide
Read the guide RelatedWhat Is a Scopus Journal? A Complete Guide for Researchers and Academics
Read the guide RelatedJournal Indexing in Plain Language
Read the guide RelatedWhy Is JMVRA's Impact Factor 11.25? Google Scholar H-Index Explained
Read the guide RelatedJMVRA's Transparency Statement in Practice
Read the guide From the blogEngineering Papers International Reviewers Accept
Read the article From the blogTurning a Project Report Into a Journal Paper
Read the article From the blogWriting a Review Article and Picking the Right Journal
Read the articleKeep Reading on the JMVRA Blog
Find this guide on the JMVRA journal blog, a collection of practical guidance on publishing, indexing, metrics and research integrity that favours verifiable content over promotion.