How to Write an Abstract
Compress five essential elements — background, objective, methods, results,
and conclusion — into 200 to 300 words. Write the abstract last after
completing the entire paper by extracting key points from each section.
Include specific numerical findings in the results, and always check the
target journal's format requirements first, as 80% of readers decide whether
to read the full text based on the abstract alone.
Why Does the Abstract Matter?
The abstract is the first thing visible after the title in database searches. Reviewers, editors, and researchers all judge a paper's value by reading the abstract. A well-written abstract increases the likelihood of citation, creates opportunities for conference presentations, and shapes the first impression during the review process.
Conversely, a weak abstract buries good research. If the key results are missing, the language is too vague, or the content is listed without structure — readers will not proceed to the full text.
What Are the Types of Abstracts?
Unstructured Abstract
Written as a single paragraph. Commonly used in humanities, social sciences, and some natural science journals.
Structured Abstract
Divided by subheadings (Background, Methods, Results, Conclusions). Standard in medicine, health sciences, psychology, and other empirical research fields.
| Type | Length | Fields | Features |
|---|---|---|---|
| Unstructured | 150-250 words | Humanities, social sciences, engineering | Free-form narrative |
| Structured | 200-350 words | Medicine, health, psychology | Subheading divisions, systematic |
Always check the journal's guidelines first. Abstract type, word count limits, and keyword requirements vary by journal.
What Should an Abstract Include?
Regardless of type, an abstract should contain the following five elements.
1. Background — 1-2 sentences
Concisely present the context and need for the research. Answer "Why is this study needed?" without repeating the entire introduction.
❌ "Recent advances in AI technology have brought many changes to the field of education."
✔ "Although the adoption of AI tutoring in university courses has been expanding, its effect on academic achievement has not been sufficiently verified empirically."
2. Objective — 1 sentence
Clearly state what this study aims to investigate.
"This study examined the effect of AI-based personalized tutoring on university students' academic achievement in statistics."
3. Methods — 2-3 sentences
Summarize the research design, participants, and data collection and analysis methods. Include only the information needed to interpret the results.
"A total of 312 statistics students from four universities in Seoul were randomly assigned to an experimental group (AI tutoring) and a control group (traditional instruction), and their academic achievement and learning motivation were compared over one semester. ANCOVA was used to analyze between-group differences after controlling for prior scores."
4. Results — 2-3 sentences
Present the main findings with specific numbers. Include statistical significance.
"The AI tutoring group's final exam scores were significantly higher than those of the control group (M=82.4 vs 76.1, p<.01, d=0.45), and learning motivation also showed a significant difference (p<.05)."
5. Conclusion — 1-2 sentences
Present the implications of the findings. Avoid overgeneralization.
"AI personalized tutoring is effective in improving academic achievement in university statistics courses and suggests potential as an individualized learning support strategy for large-enrollment classes."
When Should You Write the Abstract?
Write it last, after the paper is complete. Extracting the core of each section after the entire paper is finished produces the most accurate abstract.
- From the introduction — extract background and objective
- From the methodology — extract research design, participants, analysis methods
- From the results — extract main findings and figures
- From the discussion/conclusion — extract implications
- Combine and refine — check that the flow is natural and word count is within limits
Entering your paper's structure and key results into Nubint AI's Draft Writer agent can generate an abstract draft. Use the generated draft as a starting point and refine it to fit the context of your research.
Selecting Keywords
Most journals require 3-6 keywords below the abstract.
- Avoid words already in the title — the purpose is to broaden search discoverability
- Use standard terminology recognized in your field
- Consult controlled vocabularies such as MeSH (medicine) or ERIC Thesaurus (education)
- Avoid overly broad keywords ("education") and overly narrow ones ("University A in Gangnam, Seoul")
Abstract Characteristics by Field
| Field | Abstract Type | Key Differences |
|---|---|---|
| Medicine/Health | Structured (required) | Background-Methods-Results-Conclusions subheadings required; clinical trials include registration number |
| Psychology | Structured (APA) | Objective-Method-Results-Conclusions; emphasis on effect sizes and confidence intervals |
| Education | Unstructured/structured mixed | Must specify research context (school level, participants) |
| Engineering/CS | Unstructured | Emphasis on technical contributions and performance metrics; includes benchmark results |
| Humanities | Unstructured | Argument-centered; emphasizes key claims and interpretation over numbers |
| Business/Economics | Unstructured | Specifies research design and dataset; includes practical implications |
Tips for English Abstracts
Whether for a thesis or journal submission, an English abstract is almost always required.
- Tense: Background and objective in present/present perfect ("This study examines..."), methods and results in past tense ("Data were collected...", "Results showed..."), conclusion in present tense ("These findings suggest...")
- Subject: Start with clear subjects like "This study," "The results," or "We." Avoid vague passives like "It is thought that..."
- Reporting numbers: Follow APA guidelines — report p-values, effect sizes, and confidence intervals precisely (e.g., "F(2, 147) = 4.56, p = .012, η² = .06")
For overall paper writing guidance, see the How to Write an Academic Paper guide. For proofreading methods, see the How to Proofread Your Paper guide.
Checklist
- ☐ Does it include all five elements: background, objective, methods, results, and conclusion?
- ☐ Does it comply with the journal's word count limit?
- ☐ Does the results section include specific figures?
- ☐ Can a reader understand the entire study from the abstract alone?
- ☐ Does the abstract contain no information not in the main text?
- ☐ Are abbreviations spelled out on first use?
- ☐ Are there no citations or reference numbers? (These should not appear in abstracts)
- ☐ Are the keywords appropriately selected?
Common Mistakes
| Mistake | Solution |
|---|---|
| Ending with "we studied..." without results | Always include main findings and figures |
| Repeating the entire introduction | Compress the background to 1-2 sentences |
| Vague language ("meaningful results") | Provide specific numbers and statistical significance |
| Writing the abstract before the paper | Write it last after completing the paper |
| Exceeding the word count | Check journal guidelines, remove unnecessary modifiers |
Summary
The abstract is a miniature version and standalone text of your paper. Include background, objective, methods, results, and conclusion — completely, specifically, and concisely. The two most important principles: write it last after finishing the paper, and always include numbers in the results. A well-written abstract improves discoverability in database searches and increases both conference presentation opportunities and the likelihood of citation.