Sciwrite Publish

---
name: sciwrite-publish
description: >
  Navigate publication: choosing a journal, the submission process, responding to
  reviewers after a revise-and-resubmit, performing a peer review yourself,
  authorship decisions and order, plagiarism and self-plagiarism, and spotting
  predatory journals. Based on Stanford's Writing in the Sciences. Use this skill
  whenever someone mentions submitting a paper, a decision letter, reviewer
  comments, a rebuttal or response letter, "revise and resubmit", "my paper was
  rejected", being invited to review a manuscript, who should be an author or in
  what order, whether a passage counts as plagiarism, reusing their own text from
  a previous paper, an unsolicited invitation from a journal or conference, or
  whether a journal is legitimate. Also use it when someone asks how to choose
  where to submit. Do NOT use it for writing the manuscript itself
  (sciwrite-manuscript) or for grant applications or responding to grant or
  fellowship reviewers (sciwrite-proposals).
---

# Publication, ethics and peer review

Answer in whatever language the user writes to you in.

One number is worth keeping in mind throughout, because it reframes what the rest
of this plugin is for: roughly **60 % of reviewer criticisms concern the quality
of the writing and the tables and figures**, and only about 40 % the quality of
the science. Good writing and good data presentation are not decoration on top of
getting published — they are most of it.

---

## Choosing where to submit

Do this **before you write**, not after. It determines formatting, length,
reference style, and often the framing.

- Identify who actually cares about these data. That may not be the most
  prestigious venue; a niche journal with the right readership often serves the
  work better.
- Aim high, but be realistic. If you are junior, ask someone senior where the
  paper belongs.
- Read the online instructions for authors in full. Length limits, reference
  counts and formats, table and figure requirements.
- Pull a published paper from that journal and copy its table formatting,
  footnote symbols and section structure. This costs ten minutes and makes the
  submission look professional.
- Check the journal is legitimate — see `references/predatory-journals.md`.

---

## The submission process and what the letter means

Submission itself takes a couple of hours of online form-filling, plus time to
collect conflict-of-interest and copyright forms signed by every author. Budget
for it.

Four possible outcomes:

| Outcome | Reality |
|---|---|
| Accepted as is | almost never happens |
| Accepted with minor revisions | rare, and excellent |
| **Rejected, resubmission possible** ("revise and resubmit") | **this is what you are aiming for** |
| Rejected outright | submit elsewhere — but fix it first |

**Revise-and-resubmit letters read like rejections.** They open with something
like *your manuscript is not acceptable for publication*, and then, further down,
invite you to revise and resubmit. That second part is the actual message: the
editor thinks the paper is publishable if you address the concerns. Most papers
that eventually appear in journals go through this. Editors *want* a round of
revision, because it is what makes the paper defensible.

**On an outright rejection**, the temptation is to send it elsewhere unchanged.
Do not. You do not owe the reviewers a response, but the problems they found are
real and the next set of reviewers will find them too. Fix them first.

**After acceptance**, review the proofs carefully. Errors are introduced during
formatting and layout almost every time.

---

## Responding to reviewers

A response letter is its own genre. See `references/response-letter.md` for the
template and worked phrasing.

The essentials:

- **Be polite and appreciative**, and mean it by the time you write. Your first
  instinct on reading criticism will be defensive. That is normal; get past it
  quickly, because the faster you can treat criticism as information rather than
  as an attack, the easier your career gets.
- **Respond point by point, numbered**, mirroring the reviewers' numbering. For
  each: what you did, and exactly where in the manuscript it now appears.
- **You do not have to make every change.** Small and inconsequential requests:
  just do them. Requests you genuinely disagree with: explain your reasoning,
  respectfully. Reviewers are sometimes wrong.
- **Include a tracked-changes copy.** Most journals require it, and reviewers
  check that the changes were actually made.

One thing worth knowing, because it changes how you read a bad review: **when a
reviewer misidentifies a problem, there is usually still a problem.** A reviewer
who thinks your analysis is wrong may have been misled by a poor explanation of a
correct analysis. In one course example, a reviewer appeared to misunderstand the
background literature — and the real cause was that the authors had numbered
their references wrong, attributing a fact to the wrong paper. The reviewer was
wrong about the diagnosis and right that something was broken.

---

## Doing a peer review

If you are invited, accept. Editors actively want junior reviewers: they are
closer to current methods, and they are more careful — review quality
demonstrably deteriorates the longer someone has been doing it. You will also see
the back end of publication, read other reviewers' comments on the same paper,
and discover that your judgement is better than you thought.

Full method, tone guidance and structure: `references/peer-review.md`.

The two things that matter most:

**Tone.** You are looking for problems, so a critical tone creeps in without your
noticing. Criticise the work, not the authors. *The authors should delete Table 5.
Not only is it completely irrelevant, but it reveals their utter lack of
statistical understanding* and *Table 5 contains unnecessary information; for
example, a Pearson's correlation coefficient may not be appropriate here. The
authors should consider revising or omitting the table* are the same criticism.
Somewhere at the far end of the process is a graduate student whose confidence is
on the line.

**You are not a copy editor.** Do not itemise every typo. If the language needs
work, say so once, give one or two examples, and recommend copy editing.

---

## Authorship

Settle it **before** the paper is written, especially if you are the one drafting.
Run the list and the order past everyone early.

**Who gets authorship.** The criterion is that an author takes **public
responsibility for the content**. That is a high bar: if the paper is later found
to be flawed, or is retracted, every author owns that.

Junior researchers tend to over-include, feeling obliged to add every senior
person they consulted. That is backwards — most people do not want to be on a
paper they were not substantially involved in, precisely because of the
responsibility. They will not be offended. The acknowledgements section exists
for people who helped without meeting the bar.

**Order.** First author: usually the person who wrote the draft and did the work,
often the junior researcher. Last author: usually the senior investigator who
oversaw it — also a position of stature. Between them, order by contribution; if
contributions are genuinely equal, alphabetical (or reverse-alphabetical) is
acceptable. Dual first authorship is available when two people contributed
equally. Very large collaborations can be cited as a group.

**Acknowledgements** carry funding sources (always cite your fellowship or grant)
and people who contributed materials, statistical advice or feedback.

**Ghost and guest authorship** are the two practices to know about and refuse.
See `references/authorship.md`.

---

## Plagiarism

Most people know not to steal paragraphs. The boundary is further out than they
think. See `references/plagiarism.md` for the full treatment, including
self-plagiarism and a worked example of the subtle kind.

The short version: cutting and pasting a sentence, or part of one, is plagiarism.
So is pasting someone's text and rearranging or substituting words — those changes
are cosmetic, and the thought is still theirs. So is borrowing a definition from
Wikipedia. Either put it fully in your own words, having understood it well
enough to write it from scratch, or quote it and cite it.

The practical defence is procedural: when collecting material, paste quotations
**inside quotation marks**, and when you write, work from your own understanding
rather than from someone else's text on screen.

---

## Reference files

| File | Read it when |
|---|---|
| `references/response-letter.md` | drafting a response to reviewers |
| `references/peer-review.md` | reviewing someone else's manuscript |
| `references/authorship.md` | deciding authors and order; ghost and guest authorship |
| `references/plagiarism.md` | judging whether something crosses the line; self-plagiarism |
| `references/predatory-journals.md` | vetting a journal or an unsolicited invitation |

For checking a draft against its own bibliography before submitting, the
`paper-similarity` skill (if installed) runs a local text-overlap analysis
against the cited references. This plugin does not duplicate it.