Write Scientific Abstract

---
name: write-scientific-abstract
description: Use when drafting or revising a scientific abstract for a journal article, conference submission, or thesis
source: IMRAD format — ANSI/NISO Z39.14; APA Publication Manual 7th ed.; CONSORT/PRISMA reporting standards
tags: [biology, scientific-writing, abstract, imrad]
verified: true
---

# Write Scientific Abstract

Compose a structured scientific abstract that accurately represents the study within word limits using IMRAD logic.

## Why This Is Best Practice

**Adopted by:** PubMed/MEDLINE indexing standards, over 3,000 journals in the Elsevier/Springer/Wiley portfolios, NIH grant applications, APA-style journals, CONSORT and PRISMA systematic review guidelines.

**Impact:** Structured abstracts increase retrieval accuracy by 30% and reader comprehension by 25% vs. unstructured abstracts (Hartley 2004, BMJ study); 75% of readers read only the abstract — it must stand alone.

**Why best:** IMRAD mirrors the logical flow of scientific reasoning: background → question → method → finding → implication. Structured sections prevent omission of critical elements.

Sources: ANSI/NISO Z39.14-1997; APA 7th ed. §2.9; CONSORT 2010 item 17; PRISMA 2020 item 2.

## Steps

1. **Check journal requirements** — note the word limit (usually 150–300 words), section structure (structured vs. unstructured), and any mandatory fields (trial registration, funding).

2. **Write Background/Objective (1–2 sentences)** — state the knowledge gap and the specific aim or hypothesis. Do not write general field introduction.

3. **Write Methods (2–4 sentences)** — specify study design, subjects/samples (species, n, key characteristics), intervention or exposure, and primary outcome measure.

4. **Write Results (2–4 sentences)** — report the primary outcome with effect size and 95% CI or p-value; include one or two key secondary findings. Use exact numbers, not "significant" alone.

5. **Write Conclusion (1–2 sentences)** — state what the results mean for the field; do not overstate beyond what the data support.

6. **Add Keywords** — list 4–6 MeSH terms or journal-specified controlled vocabulary terms below the abstract.

7. **Self-check completeness** — verify: Is every claim in the abstract supported in the full text? Are all statistical figures accurate? Does it make sense without reading the paper?

8. **Cut to word limit** — remove adjectives, condense methods, shorten background; never cut results or conclusions.

## Rules

- Every number in the abstract must match the full text exactly — discrepancies are a common retraction trigger.
- Do not cite references in the abstract unless the journal explicitly allows it.
- Do not introduce abbreviations unless used ≥2 times in the abstract; spell out on first use.
- Avoid phrases like "novel," "first ever," or "groundbreaking" — let the data speak.

## Common Mistakes

- **Methods too vague** — "samples were analyzed" without specifying how makes the abstract unverifiable.
- **Results without numbers** — "levels were elevated" without values fails readers who use abstracts for meta-analysis.
- **Conclusion overreach** — claiming causation from an observational study or generalizability beyond the study population.
- **Background too long** — spending >25% of words on background leaves insufficient space for results and conclusion.

## When NOT to Use

- When writing a lay summary or press release (use plain-language summary format instead)
- When the journal requires a structured abstract with different section headers — adapt steps to match their required fields
- When writing a conference abstract with a specific template that diverges from IMRAD (follow the template)