A reliable research paper summary template keeps the research question, methods, sample or data, comparison, findings, uncertainty, and limitations in separate fields. Add a page, section, table, or figure location to every consequential note. Then write a short synthesis that states what the study reports without turning one result into a universal conclusion. Use the paper as the source for checking methods, numbers, and interpretation, not a generated summary.
Publication and evidence note: Heni Hazbay creates Summarise Visually and may benefit from downloads or subscriptions. AI assistance supported drafting and editorial review; source-dependent claims were checked against the cited primary or authoritative sources. The template, claim ledger, and fictional library example are original editorial assets, not captured app outputs or validation results. Product statements are limited to current first-party project review and the capability boundaries stated below.
Copyable research paper summary template
Copy the full version for close reading. Use the compact version later when you have already completed and checked the source record.
PAPER IDENTITY
Full title:
Authors:
Journal / conference / repository:
Year:
DOI or stable URL:
Version read:
Date accessed:
Sections and supplements actually read:
RESEARCH PURPOSE
Problem or context:
Research question, hypothesis, or main argument:
Why the authors say it matters:
Source location:
METHODS
Study or paper type:
Design:
Population, sample, dataset, or materials:
Selection / inclusion details:
Comparison or control:
Measures and outcomes:
Analysis described by the authors:
Source location:
RESULTS
Primary finding:
Measure, direction, and size as reported:
Uncertainty or precision reported:
Important secondary or null findings:
Relevant table / figure / section:
INTERPRETATION
Authors' interpretation:
My interpretation, clearly labeled:
Alternative explanation mentioned:
What the study does not establish:
LIMITATIONS
Limitations stated by the authors:
Transfer or generalization boundary:
Missing information or unresolved question:
Conflict, funding, preregistration, or data note if relevant to my task:
CHECKING LEDGER
Claim:
Paper location:
Status: Supported / Partial / Ambiguous / Wrong
Correction:
SHORT SYNTHESIS
In 3–5 sentences: question, method, primary result, important limit, and why it matters.Do not force every paper into the same empirical-study fields. A review, methods paper, theoretical article, qualitative study, case report, and commentary have different evidence structures. Write “not applicable” when a field genuinely does not apply; write “not found in the sections read” when the paper may contain it elsewhere.
Why a section-aware template matters
Nature Education’s Scientific Papers overview explains the functions of common parts of a scientific paper, including the abstract, introduction, methods, results, and discussion. Those sections answer different questions. The introduction frames the problem; the methods describe what was done; the results report findings; the discussion interprets them. A summary that blends them can turn an author’s interpretation into an observed result or omit the design that limits the conclusion.
The abstract is useful for orientation but should not automatically become your final summary. It is already a selective account written by the authors. Check the methods and results before relying on claims that matter. Record which sections you actually read so an abstract-only note does not look like a full-paper review six months later.
Fill the template in a deliberate order
1. Establish provenance and scope
Record the exact version. A preprint, accepted manuscript, and final publication may differ. Save the DOI or stable URL and note supplementary material you consulted. If your task requires citation, follow the relevant style or institutional rules rather than treating this worksheet as a formatted reference.
Write the research question in your own words, then attach the location that supports your reading. If the paper states several aims, do not compress them into one vague purpose. Identify the primary aim and list secondary aims separately.
2. Read methods before interpreting findings
Record the design, sample or data, comparison, measures, and analysis at the level needed for your task. Preserve selection rules and time periods when they affect who or what the result represents. For experiments, distinguish assignment from mere comparison. For observational work, do not casually replace “associated with” by “caused.” For qualitative work, use the design and analytic terms the paper itself supports.
If you do not understand a method, say so. Do not translate it into a familiar but different procedure. Add the term and source location to the unresolved list for a qualified instructor, methods reference, or domain expert.
3. Record results with their measure and conditions
A result needs context. Write what was measured, in whom or what, over which period, and relative to which comparison. Copy exact numbers only after checking the table, figure, or result section. Preserve direction, units, uncertainty, and whether the finding was primary, secondary, exploratory, or null when the paper makes that distinction relevant.
Avoid adjectives such as “large,” “important,” or “successful” unless the paper defines the criterion and your summary attributes it appropriately. Statistical significance, practical importance, and certainty are not interchangeable labels.
4. Separate discussion from results
Record the authors’ interpretation in its own field. Then add limitations they state and boundaries you need to inspect. A discussion can propose mechanisms or applications that were not directly measured. Your short synthesis should make that difference visible.
When the authors cite earlier work, follow the citation if your task depends on the earlier finding. One paper’s description of another is a secondary account, not a substitute for reading the original where appropriate.
5. Write the synthesis last
Use three to five sentences:
- State the question or purpose.
- Describe the design, sample or data, and comparison in one compact sentence.
- State the primary finding with the relevant measure and condition.
- Preserve the most important uncertainty or limitation.
- Explain why the paper matters for your specific task without implying a broader conclusion than the study supports.
Read the synthesis against the completed fields. Every sentence should trace to the paper or be labeled as your interpretation.
Worked fictional example
Consider a fictional study that compares two reminder formats for returning borrowed library items. The example below is invented to demonstrate the fields. It is not a real study, not a claim about reminder effectiveness, and not a Summarise Visually output.
| Template field | Fictional entry |
|---|---|
| Research question | In the fictional library, is Format A associated with more on-time returns than Format B during the four-week study window? |
| Design | Comparison of two assigned reminder formats in a simulated dataset. |
| Sample or data | 200 fictional loan records created for the example. |
| Primary result | The authored example states a difference in on-time returns; no real estimate is presented here. |
| Authors’ interpretation | The fictional authors propose that shorter wording may be easier to notice. |
| Important limit | The example does not establish behavior in a real library or beyond the fictional window. |
| Source location | Methods paragraph 2; Results table 1; Discussion paragraph 1 in the imagined paper. |
| Verification status | Structure demonstrated; no empirical claim made. |
Notice that the interpretation does not become the measured result. The proposed mechanism is labeled as a proposal, and the boundary remains in the short record. This is the discipline the template is designed to encourage.
A claim-checking ledger
For each consequential sentence in your summary, complete one row:
| Summary claim | Paper location | Evidence type | Status | Required correction |
|---|---|---|---|---|
| Exact sentence from your draft | Section, page, table, or figure | Reported result / author interpretation / my inference | Supported / Partial / Ambiguous / Wrong | Restore measure, condition, uncertainty, or attribution |
“Supported” means the source supports the wording and scope you used. “Partial” means your sentence omitted a condition, comparison, or limitation. “Ambiguous” means the available text does not settle the interpretation. “Wrong” means the sentence conflicts with the paper. Correct the summary from the paper rather than asking another generated response to adjudicate it.
The AI summary accuracy guide provides a broader method for checking omissions, contradictions, attribution, and source alignment.
Using Summarise Visually as a bounded first pass
Current project review supports supplied selectable text through a bounded iOS PDF extension handoff and selected summary modes. If the paper’s PDF exposes selectable text and you are permitted to process it, a careful workflow is:
- Save the paper identity and untouched source first.
- Share the eligible PDF through the route available in the installed app version.
- Use a selected summary mode only to create a candidate map or prose pass.
- Compare the candidate with the paper’s section structure.
- Fill the methods, sample, results, uncertainty, and limitations fields directly from the paper.
- Check every consequential sentence in the ledger.
The reviewed route is not evidence of OCR for scanned pages, a home-screen document picker, whole-page visual understanding, table or equation interpretation, reference validation, statistical analysis, or scientific-validity checks. Figures, tables, equations, appendices, and supplements require direct inspection. A generated summary cannot tell you that a method is sound or a conclusion justified merely because the prose is coherent.
For a complete note-building workflow after the template is filled, read how to turn a research paper into study notes. For a reading-order strategy, use how to read a research paper efficiently. If the source is a chapter rather than a paper, use the textbook chapter notes template.
Limits and verification
This template improves source alignment, but it does not evaluate a paper on its own.
- A summary is not peer review. It does not establish research quality, validate citations, reproduce analyses, or determine whether conclusions are justified.
- Abstract-only reading is limited. Record the sections read and inspect methods, results, tables, figures, limitations, and supplements when your task depends on them.
- Paper types differ. Adapt the fields without inventing a sample, control, or statistical result that the paper does not contain.
- Causal language needs support. Preserve the design and the authors’ wording; do not upgrade an association or interpretation into a demonstrated cause.
- Numbers need direct checking. Verify units, denominators, comparison groups, time periods, direction, and uncertainty in the source location.
- Generated text can blend roles. Keep reported results, author interpretation, and your inference in separate fields.
- Product evidence is bounded. The reviewed PDF handoff supports supplied selectable text, not OCR, whole-page visual understanding, document picking, scientific validation, or automatic citation checking.
Before relying on the summary, confirm the exact paper version, sections read, design, sample or data, comparison, primary measure, uncertainty, author-stated limits, and source location for every consequential claim. Ask a qualified domain or methods expert when interpretation carries material consequences or exceeds your expertise.
Sources and evidence
- Nature Education, “Scientific Papers”, for the functions of common scientific-paper sections that inform this section-aware template.
- First-party project review supports only the bounded selectable-text PDF extension handoff and selected-mode statements above. No OCR, document picker, whole-page visual analysis, statistical checking, citation validation, or scientific-quality assessment is claimed.
- The template, compact grid, ledger, and fictional library example are original editorial assets created for this guide. They are not Nature materials, research findings, or captured app outputs.