Manuscript rubric
Download: Word · PDF · Back to assignments
Foundations of Clinical Research · Level 1: From Clinical Question to Publication
Organisation: Neudata (#ClearDataClearImpact) Total: 100 points: the manuscript itself 90 + peer review and response 10 Use with: manuscript_brief.md, peer_review_instructions.md, the results pack for the question the participant chose (Q1 to Q5) and the STROBE checklist
The marks are formative. They help each participant improve their paper and show instructors which writing skills need more time. Mark the revised manuscript (submitted with the response letter 2 weeks after Session 8). If no revision arrives, mark the draft and give 0 for criterion 10.
Mark the writing, not the analysis. The instructors did the analyses. We gave participants the results; the question is whether the participant reported and interpreted them correctly, clearly and honestly.
How to score
- Open the results pack for the participant’s question next to the manuscript and check the main numbers against it.
- For each criterion, choose the level that best describes the work and give its points. If the work sits between two levels, choose the lower one and note what would lift it.
- Write one comment for each criterion that loses points: what to change and where (page/line).
- Add the points. Use the bands at the end only to guide the overall comment.
Levels: Excellent: complete, correct, specific. Good: mostly complete and correct, minor gaps. Needs work: present but vague, incomplete or partly incorrect. Missing: absent or wholly incorrect.
A. The manuscript (90 points)
Front matter and Introduction
| Criterion | Excellent | Good | Needs work | Missing |
|---|---|---|---|---|
| 1. Title and abstract (10) | Title names population, exposure, outcome and design. Structured abstract of 250 words or fewer; Results give the main effect with 95% CI; conclusion matches the design. 10 | Minor gaps (e.g. design not in the title, CI missing for one estimate). 7 | Abstract unstructured, over length, missing the main number, or conclusion stronger than the data. 3 | Absent. 0 |
| 2. Introduction (10) | Funnel in about 3 paragraphs: what is known (cited), the gap, a clear aim that matches the chosen question (Q1 to Q5). About 350 words. 10 | Clear but long, or the gap is only implied. 7 | Mostly general background; aim missing or different from the analysis. 3 | Absent. 0 |
Methods and Results
| Criterion | Excellent | Good | Needs work | Missing |
|---|---|---|---|---|
| 3. Methods (15) | Design, setting, participants (eligibility), variables with definitions (exposure, outcome, confounders), data sources and cleaning, bias, study size, statistical methods (crude and adjusted models, missing data, software), the simulated-data statement. Enough detail to repeat. 15 | One or two STROBE items thin (e.g. confounders not justified). 10 | Vague (“data were analysed statistically”); methods do not match the results pack. 5 | Absent. 0 |
| 4. Results (15) | Participants first (flow, n analysed); descriptive data; main result as effect size with 95% CI (crude and adjusted); secondary or sensitivity results; numbers with denominators; no interpretation; numbers agree with the results pack. 15 | Correct but with minor omissions (e.g. crude estimate missing) or one inconsistency. 10 | p-values without effect sizes; interpretation in Results; several numbers disagree with the pack. 5 | Absent. 0 |
| 5. Tables and figures (10) | Table 1 plus one or two tables or figures that stand alone: informative title, units, n per column, abbreviations explained, sensible decimals; the figure type suits the data (e.g. Kaplan-Meier, box plot, forest plot). 10 | Clear, with minor formatting gaps. 7 | Cannot be read without the text; wrong figure type (e.g. dynamite plot for skewed data); duplicated text. 3 | None. 0 |
Discussion and reporting quality
| Criterion | Excellent | Good | Needs work | Missing |
|---|---|---|---|---|
| 6. Discussion (15) | Six moves, in order: main finding in one or two sentences; comparison with other studies (incl. Paper 2); possible explanations; strengths and limitations (confounding, missing data, simulated single dataset, generalisability); implications for practice and research; conclusion. 15 | Most moves present; limitations brief. 10 | Repeats the Results; no comparison; limitations missing or token. 5 | Absent. 0 |
| 7. Calibrated language (5) | “Associated with” throughout; uncertainty acknowledged; no “trend” for non-significant results; conclusion no stronger than the design allows. 5 | One overstatement. 3 | Causal claims (“fluid causes death”) or spin. 1 | Not applicable. 0 |
| 8. STROBE adherence (5) | Checklist completed with page numbers; items not applicable are justified; the manuscript actually contains what the checklist claims. 5 | Completed; a few items claimed but thin. 3 | Incomplete or not matching the text. 1 | Not submitted. 0 |
| 9. References, length and presentation (5) | 10 to 20 relevant references in consistent Vancouver style; main text about 2,000 words (plus or minus 10%); template used; CRediT statement, acknowledgement of the course and AI-use statement present. 5 | Minor problems (a few formatting errors, slightly over length). 3 | Inconsistent or unverifiable references; far over or under length; required statements missing. 1 | Not applicable. 0 |
Manuscript subtotal: ______ / 90
B. Peer review and response to reviewers (10 points)
| Criterion | Excellent | Good | Needs work | Missing |
|---|---|---|---|---|
| 10. Review given and response to reviewers (10) | The participant’s review of one classmate’s draft is specific, constructive and uses the checklist. The response letter answers every comment point by point and politely (the comment, the reply, then the change with page/line). Disagreements are explained with reasons, and the revised manuscript shows the changes. 10 | Good review and response, but some answers vague (“done”) or one comment not addressed. 7 | Review generic or rude; response defensive, or changes claimed but not made. 3 | No review and no response. 0 |
Total: ______ / 100
Guide to overall comment
| Total | Overall comment |
|---|---|
| 80 to 100 | A well-reported paper. The structure would pass a journal’s technical check. |
| 60 to 79 | A good draft. Revise the flagged sections. |
| 40 to 59 | The core is there, but Results reporting or the Discussion needs substantial work. |
| Below 40 | Major revision needed. Meet an instructor to go through the results pack again. |
Common problems to look for (and comment on)
- Only the adjusted OR reported, so readers cannot see that confounding by severity shrank the association (or only the crude OR).
- “Positive fluid balance increased mortality”: causal language from a cohort.
- p-values without effect sizes or 95% CIs; “p = 0.000”; percentages without denominators.
- Methods describe analyses that are not in the results pack, or Results report analyses not described in the Methods.
- Missing data not mentioned (especially lactate), or the complete-case vs imputation comparison left uninterpreted (Q4).
- Mean ± SD for a skewed outcome such as ICU length of stay or ventilator-free days (Q3, Q5).
- Hazard ratio interpreted as a risk ratio; Kaplan-Meier tick marks read as deaths (Q1).
- The simulated-data statement or the AI-use statement missing.
- Discussion that repeats the Results or ignores Paper 2.