Summarize a research paper's claim and limits
Five parts: the claim, the method, how much the evidence actually supports it, the authors' own caveats, and what it does not show.
Last updated
Fill in the blanks
Assembled in your browser — nothing you type is stored or sent anywhere.
Your prompt
Model-agnostic — works in Claude, ChatGPT, Gemini or any AI chat.
Why this prompt works
- →Separating "evidence strength" from the authors' own confident framing forces an independent read of the method, instead of parroting the abstract's tone as a conclusion.
- →Restricting "limitations" to what the authors themselves admit keeps that section verifiable against the source, rather than blending in the model's own doubts unlabeled.
- →The "what it does NOT show" section targets how papers actually get misused — overreached in write-ups, not in the paper's own careful language — and "not stated" beats guessing when the text is thin.
Part by part
- the single main finding, in one sentence
- Forces the claim to be singular. Papers report many results, and the summary is useless if it does not identify which one the paper is about.
- sample size, design, what was measured
- Three specifics rather than a general description of method. Each is a number or a fact, so each is checkable against the text.
- not how confident the abstract sounds
- The most important instruction in the prompt. Abstracts systematically overstate, and a summary that inherits their confidence is worse than the paper.
- don't add limitations they didn't mention
- Keeps the limitations section evidential. Invented caveats are indistinguishable from real ones once the summary is separated from the paper.
- What it does NOT show
- The section that prevents the usual misuse. Most misreporting comes from stretching a real finding, not from inventing one.
- not stated
- An explicit marker for missing detail, so a gap in the paper does not become an inference in the summary.
Common mistakes
Summarise this paper.Produces a condensed abstract, which inherits the abstract's confidence and omits the method. The result reads authoritative and gives you no basis to judge whether the finding is solid.
What are the key takeaways from this study?Takeaways framing invites implications rather than findings, so the output is what the result might mean for practice. That is precisely the stretch the last section exists to identify.
Summarise this paper and explain why it matters.Asks for significance, which the paper may not support. A model will supply it regardless, producing a confident claim about importance that no part of the text establishes.
Questions
Why separate evidence strength from the claim?
Because a finding and its reliability are different facts. A real result from forty undergraduates and a real result from a preregistered trial of four thousand read identically in a summary that omits the method.
Why does the abstract overstate?
Because it is written to be read and cited, and the incentives reward a clean claim. The method and limitations sections routinely qualify what the abstract asserts plainly, which is why summarising from the abstract alone is unsafe.
Can a model judge study quality?
Roughly. It can notice a small sample, an absent control group or a correlational design described causally. It cannot assess whether the statistics were applied correctly, and it will not spot a subtle methodological problem.
What is the not-shown section actually for?
It is the guard against the most common misuse of research, which is extending a narrow finding to a broad claim. Naming the stretch explicitly is far more useful than a caveat about generalisability.