🔬 Research
Research Paper Analysis Prompt: Critical Appraisal Worksheet
Critically appraise one study from its pasted sections: design, methods checks, whether results back the claims, bias, reporting gaps and author questions.
0Reviews
Prompt
Act as a research methods reviewer who critically appraises one empirical paper for a reader who has to decide how much to trust it. You work only from the sections the user pastes, you quote them when you judge them, and you never invent citations, numbers, or details the paper does not report. Inputs: - Pasted sections (abstract, methods, results, limitations): [PastedSections] - Study design as stated by the authors: [StudyDesign] - The question I need this paper to help answer: [ReviewQuestion] - Reporting guideline to check against, if known (CONSORT, STROBE, PRISMA, or other): [ReportingGuideline] - Who will read my analysis: [ReaderLevel] Generate: 1. Main claim: one sentence in plain words, then the exact sentence from the paper it comes from. 2. Study frame: population, intervention or exposure, comparison, outcomes, and follow up, each marked Reported or Not reported. 3. Methods check table: sample and recruitment, allocation, blinding, outcome measures, missing data, and analysis. Columns: Item, What the paper says (quote), Concern, Severity (low, medium, high). 4. Results against claims: for each claim, the numbers reported and whether they support it, noting effect sizes, confidence intervals, and dropouts. 5. Bias and confounding: the three biggest threats, each tied to a quote. 6. Reporting gaps: items from ReportingGuideline that the pasted sections do not cover. 7. What this means for ReviewQuestion: how much weight to give the paper and why, in language that suits ReaderLevel. 8. Questions for the authors: five specific questions. Rules: - If something is not in the pasted text, write Not reported instead of guessing. - Do not add outside studies or references. - Separate what the paper shows from your judgment of it.
Instructions
Paste the abstract, methods, results and limitations, not only the abstract. Name the reporting guideline if you know it, and check every quote in the table against the paper. Use the worksheet to learn and appraise, not to write graded work, and verify any citation against the original source.
Generated Output
This image was generated using the prompt above.
Examples
Example Input
PastedSections: Title: A 10 minute daily mindfulness app and exam anxiety in first-year nursing students: a randomized trial. Abstract: 120 first-year nursing students at one university were randomized to a mindfulness app (n=60) or a waitlist (n=60) for 6 weeks before final exams. Anxiety was measured with the GAD-7. The app group showed a greater reduction in anxiety (mean difference -2.1 points, 95% CI -3.6 to -0.6). We conclude the app reduces exam anxiety in nursing students. Methods: Students volunteered after a class announcement. Allocation used a computer-generated list. Participants knew their group. GAD-7 was self-completed online at baseline and week 6. Analysis included students who completed both surveys. Results: 96 students completed week 6 (app 44, waitlist 52). App users opened the app on a median of 19 of 42 days. Limitations: single site, short follow up. StudyDesign: randomized controlled trial. ReviewQuestion: Should our nursing school recommend this app to all first-year students before exams? ReportingGuideline: CONSORT. ReaderLevel: the associate dean, not a statistician.
Example Output
1. MAIN CLAIM A 6 week mindfulness app lowered exam anxiety more than waiting. From the paper: "We conclude the app reduces exam anxiety in nursing students." 2. STUDY FRAME - Population: 120 first-year nursing students at one university. Reported. - Intervention: 10 minute daily mindfulness app for 6 weeks before final exams. Reported. - Comparison: waitlist. Reported. - Outcome: GAD-7 anxiety score at week 6. Reported. - Follow up after exams: Not reported. 3. METHODS CHECK TABLE | Item | What the paper says | Concern | Severity | |---|---|---|---| | Recruitment | "Students volunteered after a class announcement" | Volunteers may be more motivated than typical students | Medium | | Allocation | "computer-generated list" | Concealment of the list Not reported | Medium | | Blinding | "Participants knew their group" | Self-reported anxiety can shift when people know they got the app | High | | Outcome measure | "GAD-7 was self-completed online" | Standard scale, but self-report only | Low | | Missing data | 96 of 120 completed (app 44 of 60, waitlist 52 of 60) | 27% dropout in the app group against 13% on the waitlist | High | | Analysis | "included students who completed both surveys" | Completers only, not intention to treat | High | 4. RESULTS AGAINST CLAIMS - Claim: the app reduces exam anxiety. Reported: mean difference -2.1 GAD-7 points, 95% CI -3.6 to -0.6, so the interval does not cross zero. But the result comes from completers only, and more app users dropped out. If the dropouts did worse, the true effect could be smaller. Partly supported. - Use: median 19 of 42 days, so most students used it on fewer than half the days. The paper does not link use to results. 5. BIAS AND CONFOUNDING 1. Attrition bias: "96 students completed week 6 (app 44, waitlist 52)." 2. Performance and detection bias: "Participants knew their group" with a self-reported outcome. 3. Selection: "Students volunteered after a class announcement," so results may not carry over to all first-year students. 6. REPORTING GAPS (CONSORT) - Allocation concealment and who ran the randomization. - Sample size calculation. - Reasons for dropout by group. - Intention to treat or sensitivity analysis. - Any harms or adverse effects. 7. WHAT THIS MEANS FOR YOUR QUESTION This is one small trial at one school, with open group assignment and uneven dropout. It suggests the app might help some students, but it is not strong enough on its own to recommend it to every first-year student. A reasonable next step is offering the app as an optional resource and tracking your own students' results. 8. QUESTIONS FOR THE AUTHORS 1. How was the allocation list concealed from the people enrolling students? 2. Why did 16 app group students drop out, and how did their baseline GAD-7 scores compare? 3. Do the results hold in an intention to treat analysis? 4. Did students who used the app more days show larger changes? 5. Were anxiety scores measured after final exams?