Find Where a Claim Was Actually Tested

Find Where a Claim Was Actually Tested

Search the sentences authors write when citing, to find where your claim was tested rather than repeated.

scite search evidence null-results literature-review citations

What it is for

A search technique that goes after the sentences authors write when citing, rather than titles and abstracts, including deliberately searching the negative phrasing that surfaces null results. Use it when you need to know whether a claim was tested, not just discussed. It prevents an argument built on a widely repeated claim that traces back to a single unopened study.

The skill

Abstract search finds papers about your topic. Searching the
sentences other authors wrote while citing finds the papers where
your specific claim was put to a test — including results buried in
studies whose title is about something else entirely.

Search the way people write citations, not the way they write
titles:

  - Use mid-sentence phrasing: "reduced mortality in", "no
    significant difference in", "failed to replicate", "in contrast
    to", "consistent with".
  - Pair the intervention with the outcome, not with the field.
  - Try the negative form on purpose. "did not reduce" and "no
    association between" surface the null results that title search
    hides completely.
  - Vary the noun. Authors cite the method, the cohort, the model
    and the effect by different names than the abstract uses.

For every sentence you find, do this before it goes near a draft:

  - Open the citing paper and read the sentence in its paragraph.
  - Open the cited paper and confirm it says what the citing author
    said it says. Citing authors misdescribe their sources
    constantly, and the misdescription then gets copied onward.
  - Note whether the citing paper is reporting its own result or
    summarising somebody else's.

Keep the sentence and both papers together in your notes. A citing
sentence without its citing paper is an orphan quote you will not
be able to reconstruct in a month.

What this prevents: building an argument on a claim that looks
well-supported because it is repeated often, when every sentence
repeating it traces back to one study nobody reopened.

How to install it

1. Open the tool and start a new conversation, project or template.
2. Paste the text below wherever the tool keeps standing instructions
   — usually called system prompt, instructions, or persona.
3. Save it so it applies to every run rather than a single message.

Did it work for you?

Where it came from

Skills app Original