A useful journal club presentation does more than summarize a paper. It helps a group follow the research question, inspect the study design, connect each major claim to evidence, test the authors' interpretation, and discuss what remains uncertain.
Start by confirming your club's format. Then build a claim-to-evidence worksheet before you build slides. Use the paper as the source of record, choose an appraisal tool that fits the study design, and preserve every uncertainty you cannot resolve.
This guide includes a copyable presentation plan, an original worksheet, a fictional worked example, and practical AI preparation tips. It does not provide medical advice or decide whether a paper should change clinical care, policy, or research practice.
Confirm the assignment before you make slides
Journal clubs do not all use the same format. Some expect a short discussion guide. Others expect a formal slide presentation. UAB's journal club guide recommends checking how the local group normally runs journal club before applying a general structure. Stanford's Community College Outreach Program similarly emphasizes the audience, the paper's figures, and the discussion rather than a universal slide count.
Ask five questions first:
- What is the time and format? Confirm the presentation length, discussion time, file format, and whether slides are expected.
- Who is the audience? Note what background they already have and what context they need.
- Was the paper assigned or selected? If you choose it, confirm the selection criteria with the organizer.
- Which appraisal framework is expected? Use the course, lab, or professional program's required method when one is specified.
- What may be shared with an AI tool? Check institutional policy, course rules, article license, research restrictions, and the approved tool before uploading a paper or notes.
If the organizer's instructions conflict with this guide, follow the organizer.
Copy this claim-to-evidence worksheet
Do not begin by copying the abstract into slides. First separate what the paper asked, what it did, what it observed, what the authors concluded, and what the evidence leaves unresolved.
Start with this paper-level information:
- Paper citation
- Why this paper is being discussed
- Audience and presentation format
- Research question in your own words
- Study design as reported
- Funding and conflicts of interest reported by the paper
- Important missing or unclear information
- Study-design-specific checklist used
- Questions to verify with the organizer or a subject expert
Then repeat this block for every major claim:
- Claim:
- Source location: page, section, table, or figure.
- Method or comparison:
- Reported result:
- Authors' interpretation:
- My appraisal question:
- What remains unproven:
- Discussion question:
The worksheet is not a score. Its purpose is to make the evidence path visible. Write not reported, unclear, or not found when the paper does not answer a question. Do not fill a gap with an assumption.
Make the worksheet usable on a phone
Avoid a wide spreadsheet with one column per claim. On a small screen, keep the paper-level details in one note and give each claim its own vertically stacked card or note. Keep the source location near the reported result so it remains visible without horizontal scrolling. If you later move the worksheet to a desktop, preserve those labels rather than compressing them into unlabeled columns.
Use an appraisal tool that fits the study design
The Critical Appraisal Skills Programme publishes different checklists for randomized trials, systematic reviews, qualitative studies, cohort studies, diagnostic studies, case-control studies, and other designs. CASP describes its checklists as teaching tools and advises against reducing the appraisal to a single score.
The EQUATOR Network organizes reporting guidelines by study type. A reporting guideline helps you check whether the authors reported the information expected for that kind of study. It is not automatically a verdict on whether the study is valid or useful.
Keep three judgments separate:
- Reporting: Is the information needed to understand the study present?
- Method: Does the design and execution address the stated research question?
- Interpretation: Do the conclusions stay within the evidence reported?
A missing detail may indicate incomplete reporting. It does not, by itself, prove that the underlying work was performed incorrectly. A fully reported method can still have important limitations. Record the difference.
Build the presentation around nine jobs
The sequence below describes presentation jobs, not a mandatory slide count. Combine or split them to match your group's format.
1. Identify the paper
Show the full citation, publication context, and why the group is discussing it. Avoid using journal prestige or author reputation as a substitute for appraisal.
2. Give only necessary background
Show the minimum concepts the audience needs to follow the research question. Avoid a broad literature review that consumes the discussion.
3. State the research question
Name the population or setting, comparison where relevant, and outcome or phenomenon studied. Do not rewrite the question to make it stronger than the paper does.
4. Map the design
Show groups, sequence, timing, data sources, and the reported analysis. Avoid decorative diagrams that hide missing details.
5. Walk through the central evidence
Take one figure, table, or result at a time. Explain the method that produced it and the claim it supports. Do not read every number or panel without a purpose.
6. Appraise the evidence
Connect strengths, limitations, uncertainty, and design-specific questions to the results. Avoid a generic limitations list that could apply to any paper.
7. Compare claim and conclusion
Separate what the results support, what the authors conclude, and your appraisal of the gap. Do not present your interpretation as the authors' statement.
8. Lead the discussion
Prepare three to five questions the group can answer from the paper and its context. Avoid questions so broad that nobody knows where to begin.
9. Close with an evidence boundary
State the supported takeaway, the largest unresolved issue, and a useful next question. Do not make a clinical, policy, or research recommendation you are not qualified or authorized to make.
Stanford's journal-club guidance suggests taking figure-level notes on the aim, method, result, your interpretation, the authors' interpretation, and how the figure contributes to the paper's main question. Use that as a check on every central evidence slide.
Use the paper's figures only when your use is permitted. Cite the source on the slide, keep labels legible, and do not alter a figure in a way that changes its meaning. If you redraw a study flow for clarity, label it as your own representation and check it against the paper.
See the worksheet in a fictional example
The example below is synthetic. It does not describe a real study, institution, result, or recommendation.
Imagine a paper that compares two peer-feedback formats in an undergraduate writing course. The authors report that students using structured prompts completed more revision categories than students using an open comment box. The paper concludes that structured prompts improve the revision process.
A weak presentation repeats that conclusion as fact. A stronger worksheet separates the layers:
- Research question: Did a structured feedback prompt change the revision categories completed in this course?
- Evidence: The paper's reported comparison of completed revision categories between two course sections.
- Authors' interpretation: Structured prompts improved the revision process.
- What the evidence supports: The measured revision-category outcome differed in this reported setting.
- What remains uncertain: Whether the groups were comparable, whether the measure reflects writing quality, and whether the result transfers to other courses.
- Discussion question: Does the reported outcome justify the broader phrase "improved the revision process"?
The presentation can now show the design, the central result, and the gap between the measured outcome and the broader conclusion. The group has a concrete question to discuss without the presenter inventing a verdict.
Prepare questions that point back to evidence
Use questions that help the group inspect a specific choice or inference:
- Does the study design fit the research question?
- Which result carries the paper's central claim?
- What alternative explanation remains plausible?
- Which reported limitation most changes the interpretation?
- Is an important population, setting, measure, or time horizon missing?
- Does the conclusion go beyond the measured outcome?
- What additional evidence would resolve the largest uncertainty?
- Which part of our interpretation should be checked with a subject or methods expert?
Avoid asking only, "What did everyone think?" A specific question gives the group an evidence path to follow.
Use AI to organize, not to decide
An AI assistant can help turn permitted notes into a draft worksheet, identify where a claim appears in the provided paper, propose a slide order, or surface questions for human review. It should not replace reading the paper, manufacture a citation, decide that a study is valid, or make a clinical recommendation.
Before using any AI system, follow your instructor's, institution's, publisher's, and research group's rules. Stanford Libraries advises students to confirm that AI use is permitted, verify generated material, and keep personal or confidential information out of submitted context. UC Berkeley prohibits unapproved use of confidential, proprietary, unpublished-research, and other sensitive information in generative AI tools.
Use only a permitted paper or approved notes. Do not upload unpublished research, participant data, protected health information, student records, reviewer materials, confidential manuscripts, licensed database content, or a colleague's notes unless the applicable rules and approved system explicitly allow it.
Try this preparation prompt:
Using only the permitted paper and notes I provided, prepare a draft claim-to-evidence worksheet. For each major claim, give the page, section, table, or figure where the support appears. Keep the reported result, the authors' interpretation, and possible appraisal questions separate. Write "not found" when the source does not support a statement. Do not create citations, infer missing methods or results, reproduce long passages, score the paper, provide medical advice, or recommend a change to practice. End with a verification checklist and three discussion questions for my review.
Treat the return as a navigation aid. Open every cited location and verify it yourself. Check numbers, denominators, units, labels, study design, and whether the wording came from the paper or from the assistant. Ask a qualified instructor, mentor, librarian, statistician, or subject expert when the appraisal turns on expertise you do not have.
If your program requires disclosure of AI use, record the tool, purpose, inputs, and material changes according to that policy.
Rehearse the evidence path
Before presenting, check the deck in the order the audience will see it:
- The research question appears before detailed results.
- Every central claim leads to a readable figure, table, or source location.
- The paper's interpretation and your appraisal are visibly distinct.
Not reported,unclear, andnot establishedremain visible where appropriate.- Funding and conflicts reported by the paper are included where relevant.
- Figures are legible and properly cited.
- The discussion questions can be answered by examining the paper and its context.
- The ending states a bounded takeaway, not a treatment or policy directive.
- The presentation fits the assigned time and leaves room for discussion.
Practice explaining one central figure without reading the slide. Then practice saying what that figure does not establish. That second sentence often creates the most useful discussion.
Keep the follow-up separate from the verdict
After the journal club, save the group's questions, points to verify, and any agreed follow-up. Do not turn a comment made in discussion into an established finding. Label suggestions, decisions, and unresolved questions separately.
If the group decides that an expert should review a statistical method or domain-specific implication, record the owner and next step. The presentation is an educational artifact, not a final adjudication of the paper.
Where Finta can help
Finta Documents can keep approved papers, notes, and working materials together. Aurora can help organize supported context into a draft worksheet or presentation outline for review. Availability depends on file support, permissions, connected context, and the tools available in the workspace.
Finta does not replace a literature database, institutional repository, citation manager, statistical expert, instructor, research supervisor, or clinical decision process. Keep the original paper authoritative and verify the draft before presenting.
To organize and find your course materials, use Build an AI Second Brain for Class Notes and Documents.
If you are presenting your own project's progress, observations, and open question rather than appraising someone else's paper, use Lab Meeting Presentation: Prepare Your Update with an AI Teammate.
Prepare one paper, not a new system
Start with the paper you have been assigned and the format your group expects. Complete the claim-to-evidence worksheet, choose the right appraisal framework, and build the smallest presentation that lets the audience follow the evidence and discuss the uncertainty.
Sources
- Stanford Community College Outreach Program, How to Give a Journal Club Presentation
- UAB Physician Scientist Development Office, Step-by-Step Approach to Presenting at Journal Club
- Critical Appraisal Skills Programme, CASP Checklists
- Critical Appraisal Skills Programme, FAQs
- EQUATOR Network, What Is a Reporting Guideline?
- Stanford Libraries, Generative AI in Undergraduate Research: The Basics
- UC Berkeley, Appropriate Use of Generative AI Tools