Read a paper without losing the caveats

---
name: read-a-paper-without-losing-the-caveats
description: "Extracts what a study actually established, at what sample size, under what conditions — and separates the authors' claims from what the press release said. Use when the user asks for help with research tasks like research, papers, critical-reading."
license: CC-BY-4.0
metadata:
  source: https://promptsbuddy.com/prompts/read-a-paper-without-losing-the-caveats
  author: "Lena Fischer"
  version: "1.0"
---

# Read a paper without losing the caveats

Extracts what a study actually established, at what sample size, under what conditions — and separates the authors' claims from what the press release said.

## Inputs to collect first

Ask the user for anything below that they have not already given you. Do not
invent values for these.

- `PAPER` — The abstract and methods at minimum; the full text if you have it. (e.g. Paste the paper or its abstract, methods and results.)
- `WHY` — What decision you're trying to inform — this changes what matters. (e.g. Deciding whether to switch our onboarding to the pattern this study tested.)

## Instructions

I want to understand what this paper actually shows.

PAPER:
{{PAPER}}

WHAT I'M TRYING TO DECIDE: {{WHY}}

**1. The claim.** What the authors say they found, in one sentence, in their
own terms.

**2. What was actually measured.** The operational definitions — what exact
thing was counted or scored, on whom, how many, over what period, compared
against what. Numbers, not descriptions.

**3. The gap.** The distance between what was measured and what the claim
implies. This is the section that matters.

**4. Conditions.** What had to be true for this result to hold, including
population, setting, and anything the authors themselves flag in limitations.
Quote their limitations rather than paraphrasing.

**5. Does it bear on my decision?** Given what I said I am deciding, state
plainly whether this study is evidence for it, weak evidence, or not evidence
at all. Say which of my conditions differ from the study's.

Rules:
- Do not report a p-value as if it were an effect size. Give the effect size
  and its interval if present; say so if absent.
- Do not smooth over a small sample. Name n.
- Do not use the word "proves".
- If the methods section is not in what I gave you, say what you cannot assess
  rather than assuming it was done well.

## Notes from people who use this

- Section 3 is the whole prompt. Most misreading happens in the gap between the measurement and the headline.
- Give it your actual decision — the same paper is strong evidence for one question and irrelevant to another.
- Quote-the-limitations beats paraphrase: authors are usually honest in that section and it gets lost in summaries.

---

Read a paper without losing the caveats · by Lena Fischer · v1.0
From PromptsBuddy — https://promptsbuddy.com/prompts/read-a-paper-without-losing-the-caveats
Licensed CC BY 4.0.

Every prompt is also available at /prompts/read-a-paper-without-losing-the-caveats/skill.md — see all install options.