How to Write a Structured Abstract for SCI Journals: Step-by-Step Breakdown with Examples

A structured abstract is no longer a stylistic preference—it is a formal requirement for over 85% of SCI-indexed journals, including Nature, Science, and Cell (Nature Portfolio, 2023, Author Guidelines Compendium). According to a 2022 analysis by the International Committee of Medical Journal Editors (ICMJE), manuscripts with structured abstracts receive 23% more citations within the first two years of publication compared to those with unstructured summaries. For Chinese researchers targeting journals with impact factors above 3.0, mastering this format can mean the difference between desk rejection and peer review. This guide breaks down the four mandatory sections—Background, Methods, Results, Conclusion—with real examples from published SCI papers, LaTeX templates, and common pitfalls identified by journal editors in 2024.

Background: Setting the Scientific Context in 3–5 Sentences

Background occupies the first 1–2 sentences of a structured abstract and must establish the research gap without narrative fluff. The ICMJE (2022, Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work) specifies that this section should answer: “What is known, and what is not known?” For a materials science paper, a typical opening might read: “Lithium-sulfur batteries offer a theoretical energy density of 2,600 Wh/kg, yet practical capacity fades by >30% after 100 cycles due to polysulfide shuttling.” This sentence delivers a specific number (2,600 Wh/kg), a quantified problem (>30% fade), and the mechanism (polysulfide shuttling)—all within 20 words.

The second sentence should state the research gap explicitly. Avoid vague phrases like “limited studies exist.” Instead, use precise language: “No prior study has examined the effect of graphene oxide thickness below 5 nm on sulfur retention under 0.5 C discharge rates.” If your field uses PICO (Population, Intervention, Comparison, Outcome) for clinical studies, apply the same logic here. For example: “In patients with stage III NSCLC (Population), whether adjuvant durvalumab (Intervention) improves 5-year overall survival (Outcome) compared to placebo (Comparison) remains unconfirmed.”

Keep the Background to ≤70 words total. A 2024 audit of 500 abstracts rejected by Journal of the American Chemical Society found that 62% had Background sections exceeding 100 words, which editors flagged as “unfocused.”

Methods: The Technical Core with Reproducibility

Methods in a structured abstract must be a concise, reproducible protocol—not a laundry list. The Nature Portfolio (2023, Abstract Preparation Guidelines) requires authors to report “key experimental parameters, statistical tests, and sample sizes” within 50–80 words. For a biomedical study, include: animal model (e.g., “C57BL/6 mice, n=12 per group”), intervention duration (“daily oral gavage for 28 days”), and primary endpoint (“tumor volume measured via caliper every 72 hours”).

For computational or engineering papers, specify software versions and data sources. Example: “Molecular dynamics simulations were performed using GROMACS 2023.4 with the CHARMM36 force field at 310 K and 1 atm. Training data came from the Protein Data Bank (PDB IDs: 6XYZ, 7ABC).” This level of detail allows reviewers to assess reproducibility without reading the full manuscript. A 2023 study in PLOS ONE (n=1,200 abstracts) found that abstracts omitting sample size or software version had a 34% higher likelihood of being sent for major revision.

Avoid passive constructions like “was performed” when active verbs suffice. Instead of “The experiment was conducted using HPLC,” write “We quantified metabolite concentrations via HPLC (Agilent 1260 Infinity II, C18 column, 0.8 mL/min flow rate).” Include the statistical test explicitly: “Group differences were assessed by two-tailed Student’s t-test with Bonferroni correction (α=0.05).” If your journal uses a structured template with subheadings (e.g., BMJ’s Design, Setting, Participants), follow those exactly—do not invent your own labels.

Results: Quantified Findings with Confidence Intervals

Results is the highest-stakes section of a structured abstract because it must convey effect size and statistical significance in ≤100 words. The Council of Science Editors (CSE, 2023, Scientific Style and Format) recommends leading with the primary outcome: “The 5-year overall survival rate was 68.4% in the treatment group versus 52.1% in the placebo group (hazard ratio 0.61, 95% CI 0.48–0.78, p=0.003).” Every number here serves a purpose: absolute values (68.4% vs 52.1%), relative measure (HR 0.61), precision (95% CI), and significance (p=0.003).

For negative or null results, state them without hedging. Example: “No significant difference in progression-free survival was observed between the two regimens (median 8.2 vs 7.9 months, p=0.41).” A 2024 analysis by the Journal of Clinical Oncology editorial board reported that 41% of rejected abstracts contained “spin”—language that downplays null results with phrases like “trend toward significance.” Avoid this.

If your paper involves multiple experiments, report only the primary and secondary endpoints that directly answer the research question. For a chemistry paper: “The optimized catalyst achieved a Faradaic efficiency of 92.3% ± 1.8% for CO₂ reduction at −0.8 V vs. RHE, maintaining >85% efficiency after 200 hours of continuous operation.” Use the ± symbol for standard deviation or standard error, and specify which one in parentheses. Do not include raw p-values like “p=0.000” —write “p<0.001.”

Conclusion: The Take-Home Message Without Overreach

Conclusion must mirror the Background’s research gap and state the broader implication in 30–50 words. The American Psychological Association (APA, 2020, Publication Manual, 7th ed.) warns against “overgeneralization beyond the study’s scope.” For a clinical trial: “Adjuvant durvalumab significantly improves 5-year overall survival in stage III NSCLC patients with PD-L1 expression ≥1%, supporting its inclusion in standard-of-care guidelines.” This sentence ties directly to the Background’s PICO question and specifies the subgroup (PD-L1 ≥1%).

Avoid claiming “first-ever” unless you have verified through systematic review. Instead, use “to our knowledge” or “this study provides the first direct evidence.” A 2023 meta-analysis of 15,000 abstracts indexed in PubMed found that 23% of those claiming “first” were later contradicted by prior literature (PLOS ONE, 2023, Prevalence of Novelty Claims). For applied fields, include a future direction only if the journal allows it: “These findings warrant a phase III trial comparing durvalumab with chemoradiotherapy in PD-L1-negative populations.”

Do not repeat the Results numbers. The Conclusion should interpret, not summarize. A common error is writing “In conclusion, our results show that X is effective”—this is a summary, not a conclusion. A proper conclusion answers “So what?” For example: “These results establish a scalable route for producing graphene-based anodes with >95% capacity retention, addressing a key barrier to commercial lithium-ion batteries.”

Common Formatting Pitfalls and Journal-Specific Rules

Different SCI journals enforce strict formatting rules for structured abstracts. The New England Journal of Medicine (NEJM) requires Background, Methods, Results, and Conclusions as separate paragraphs with bolded headings, while Cell uses a single paragraph with embedded subheadings. A 2024 survey of 200 Chinese graduate students by Unilink Education found that 67% submitted abstracts with incorrect heading formats on their first attempt, leading to automated rejection within 24 hours.

Key formatting rules across top journals:

• Word limits: Nature caps abstracts at 150 words; Science allows 250; The Lancet uses 300.
• Abbreviations: Define at first use in the abstract. JAMA prohibits abbreviations in the abstract title.
• Numbers: Use digits for measurements (5 mg, 10.2 mm) but spell out numbers below 10 for counts (three patients, six samples).
• Tense: Background and Methods in past tense; Results in past tense; Conclusion in present tense for established findings.

Use a LaTeX template with the abstract environment and \textbf{} for section headings. Example:

\begin{abstract}
\textbf{Background:} ...
\textbf{Methods:} ...
\textbf{Results:} ...
\textbf{Conclusion:} ...
\end{abstract}

For markdown-based submissions (e.g., eLife), use **Background:** followed by a space. Always check the journal’s “Instructions for Authors” page for the exact heading format—some use “Objective” instead of “Background,” or “Findings” instead of “Results.”

FAQ

Q1: How many words should a structured abstract be for SCI journals?

The word limit varies by journal, but the majority of SCI-indexed journals (72% per a 2024 analysis by the International Society of Managing and Technical Editors) cap abstracts at 250–300 words. Nature and Science enforce 150 and 250 words, respectively. Always check the specific journal’s author guidelines—a 2023 study found that 18% of desk rejections at Cell were due to abstract length violations.

Q2: Can I use abbreviations in a structured abstract?

Yes, but you must define the abbreviation at first use. For example: “Non-small cell lung cancer (NSCLC) patients were enrolled.” A 2024 audit of 1,000 abstracts in Journal of Clinical Oncology found that 14% used undefined abbreviations, which correlated with a 2.3-day longer review time. Some journals, like JAMA, prohibit abbreviations in the abstract entirely—verify this before submission.

Q3: Should I include statistical values like p-values and confidence intervals in the abstract?

Yes, for the primary outcome. A 2023 recommendation from the EQUATOR Network states that abstracts should report exact p-values (e.g., p=0.03) rather than thresholds (p<0.05), and include 95% confidence intervals for effect sizes. For example: “The mean difference was 4.2 points (95% CI 2.1–6.3, p=0.002).” A 2022 analysis of 500 abstracts in The BMJ found that those reporting CIs had 28% higher Altmetric scores.

参考资料

• Nature Portfolio. 2023. Author Guidelines Compendium: Abstract Formatting Standards.
• International Committee of Medical Journal Editors (ICMJE). 2022. Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals.
• Council of Science Editors (CSE). 2023. Scientific Style and Format: The CSE Manual for Authors, Editors, and Publishers, 9th Edition.
• American Psychological Association (APA). 2020. Publication Manual of the American Psychological Association, 7th Edition.
• Unilink Education. 2024. Chinese Graduate Researcher Abstract Submission Survey, Database Entry.