The criteria, and what they exclude
The International Committee of Medical Journal Editors requires all four of its criteria to be met. Substantial contributions to the conception or design of the work, or to the acquisition, analysis, or interpretation of data. Drafting the work or reviewing it critically for important intellectual content. Final approval of the version to be published. And agreement to be accountable for all aspects of the work.2
The exclusions are stated as plainly as the criteria. Acquisition of funding, general supervision of a research group, and general administrative support do not, alone, qualify a contributor for authorship.2 Neither do writing assistance, technical editing, language editing, or proofreading.2
Contributors who do not meet all four criteria should not be listed as authors, and they should be acknowledged instead.2 The acknowledgment is a requirement rather than a courtesy.
Two systems doing two different jobs
CRediT and the authorship criteria are not competing answers to one question. They answer different questions, and confusing them causes most of the disputes.
The criteria answer whether a person is an author. That is a threshold question with a yes or a no, it governs the byline, and it carries accountability for the work.
CRediT answers what each person did. That is a description rather than a threshold, it applies to authors and non-authors alike, and it confers nothing and withholds nothing.
PLOS ONE requires that contributors who do not meet its authorship criteria be mentioned in the acknowledgments. It also requires that anyone named there has given permission to be named.3 A contribution statement and an acknowledgment are therefore both records, and both are checkable against the paper.
What an editor can check, and it is more than authors expect
Four things can be checked, and none of them requires deciding who deserves credit for anything.
Every named author appears in the contribution statement, and every person in the contribution statement appears in the byline or the acknowledgments. A name in one list and not the other is the most common fault and the easiest to find.
Every acknowledged person has, on the record, given permission to be named, because the journal requires it.3 An editor can flag the names with no recorded permission. Obtaining it is the authors' work, and it takes an email each.
The contribution statement agrees with the methods section. A person credited with software or formal analysis should be traceable to work the paper describes.
Order and affiliation are internally consistent across the byline, the footnotes, and the corresponding author details, which on a fifteen-author paper they frequently are not. Editors Canada places that kind of check inside copy editing, covering correctness, consistency, and completeness.4
Where the authors decide, and it is early
Who qualifies as an author, in what order they appear, and which roles each person is credited with are the research team's decisions. An editor who adjusted any of them would be intervening in a matter with professional consequences for real people.
An editor also cannot settle a dispute about them, and should not be asked to, because the question is one of professional standing rather than of language. What an editor supplies is the list of places where the paper's own records disagree with each other. That list occasionally reveals a disagreement nobody had noticed, which is a better moment to find it than after publication.
Everything a client sends is treated in confidence, including manuscripts under review, unpublished data, and correspondence with co-authors or a journal. We work on journal articles and academic and research documents against whatever the destination publishes.
Research groups that agree contributor roles when the work starts, rather than when the paper is submitted, have a shorter conversation at submission. The fourteen roles exist partly to make that conversation possible before anyone is invested in the answer.