Judge a Paper by Who Argued With It

Judge a Paper by Who Argued With It

Use citation statements, not citation counts, to decide whether a paper can carry the claim you need.

scite citations evidence literature-review replication academic-writing

What it is for

A reading order for citation evidence: contrasting statements first, then supporting ones tested for whether they replicate or merely repeat, then independence of the citing authors, then dates. Use it before any paper enters your argument. It prevents leaning on a finding that later work failed to reproduce, on the strength of a healthy-looking count.

The skill

Never quote a citation count as evidence of anything. Open the
citation statements and read the sentences.

Before you rely on a paper:

  1. Read the contrasting statements first, however few there are.
     All of them. Two contrasting sentences from a serious
     replication attempt outweigh forty supporting mentions.
  2. Read the supporting ones next, and ask of each: did this
     author test the claim, or repeat it? Most citations repeat. A
     wall of support can be one original result echoing.
  3. Look at who is citing. Same lab, same senior author, same
     consortium is not independent support.
  4. Look at the dates. Support that stops five years ago, followed
     by silence, is itself a signal.

Treat the supporting / contrasting / mentioning labels as a machine
classification, because that is what they are. They are derived
from the citing sentence, and citing sentences are hedged, partial,
or about a different result in the same paper. Read the sentence
and decide for yourself. Re-label it in your own notes when you
disagree.

Zero contrasting statements means nobody has published a challenge
that got classified as one. It does not mean the finding held up. A
recent paper has almost no statements of any kind — that is age,
not quality.

Write your verdict in one line before moving on:

  "<Paper>: N citations, X contrasting — read all X, they concern
   <the sample / the dose / the analysis>. Using it for <narrow
   claim> only."

The failure this prevents: citing a paper that a later and larger
study failed to replicate, because the counts looked healthy and
you never opened the sentences underneath them.

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