Desk Screening Automation for Journals
Desk screening is the pass a journal makes before peer review, and it is largely a completeness protocol: does the submission contain what the journal requires, do its references hold up, is anything missing that an editor would otherwise chase by email. Automation can run that pass first. It cannot judge novelty or scientific validity.
Editorial access is limited to partner journals
Where the automated pass sits in the protocol
Running a completeness pass at receipt, before assignment, changes what the editor is reading for: the mechanical questions are already answered, so the read is about scope and merit. The pass produces a report, never a decision, and it says nothing about novelty or scientific validity. Sequence matters more than the tool. A screening protocol that runs the automated pass after an editor has already read the manuscript saves nothing, because the expensive step has happened. Run at receipt, the report arrives with the submission, and the handling editor opens a file that already has its mechanical findings listed with locations. Most editorial offices already have a screening protocol written down somewhere. What varies is how much of it survives a busy week, and that is the gap an automated pass closes: not a better protocol, the same protocol applied every time.
A screening protocol is the part of a journal's process that is easiest to write down and hardest to keep. Most editorial offices have one: a document listing what should be verified before a manuscript is assigned to a handling editor. What actually gets verified depends on who opened the file and how much of the week was left when they did.
Order of operations
Where the automated pass sits decides whether it is worth anything. Run after an editor has read the manuscript, it confirms what they already know and saves nothing, because the expensive step already happened. Run at receipt, it changes the read: the file arrives with its mechanical findings listed and located, and the editor's first pass is about scope and contribution rather than about whether the bibliography is intact. Same tool, different value, purely as a function of position in the workflow.
Two piles, not one list
A screening report is only useful if the office can split it. Some findings are correspondence: a funding statement that never made it into the file, an identifier that does not resolve, a figure cited in the results that is not in the document. Those become a return letter, and the letter can be written from the report rather than from a second read of the manuscript. Other findings are context for the editorial read: a bibliography leaning heavily on one venue, references clustered in a narrow window of years, a high proportion of preprints. Those are not grounds for anything on their own. They are things an editor would want to have noticed before deciding where the manuscript goes.
A written protocol and an applied protocol are two different things
A screening protocol that drifts does not fail loudly. It fails quietly and unevenly: the same document, the same list, and a different set of things actually verified depending on who opened the file and how much of the week was left when they did. The office ends up unable to say which submissions were screened against the whole protocol and which were screened against the half of it someone remembered, which means it cannot state what its own screen covers. That is the cost, and it is not recoverable later by reading more carefully, because nobody knows which files need the second look. An automated pass does not write a better protocol. It removes the variable of who was on duty.
Where the protocol stays human
Scope fit, novelty, and scientific validity are not on the list of things a screening pass can decide, and no honest description of one will put them there. The pass reports on presence, consistency, and resolution. Everything that requires reading the argument stays with the editor, which is the point: the protocol exists so that editorial attention is spent on the part of the manuscript only an editor can evaluate.
Also see: Manuscript screening for journal editors | For journals and editorial teams | CheckMyManuscript vs iThenticate
The screening pass, step by step
Run at receipt, before assignment
The full check set runs on the submitted file, so the report is attached to the manuscript by the time a handling editor is chosen.
Every submission gets the same reference check
Citation and bibliography matching, missing entries, duplicates, and identifier validation run identically regardless of who is on duty.
Findings arrive with locations, not impressions
Each flag names what was found and where, so a return letter to the author can be written from the report instead of from a second read.
Separates the email items from the editorial ones
A missing funding statement or an unresolvable DOI is correspondence. A bibliography leaning on one venue is context for the editor's read.
Checks figures and tables against the text
Visuals cited in the text but absent from the file, visuals present but never cited, reused or missing captions, and numbering out of sequence.
The protocol survives a change of editor
A guest editor or a new associate editor inherits the same screen, so consistency does not depend on how long someone has been in post.
Checks relevant to this topic
Reference and identifier validation
References checked for missing fields, duplicates, and DOIs or URLs that do not match the work cited. The findings that most often surface late, in a reviewer report.
Figure and table consistency
Cross-checks the text against the visuals: cited but absent, present but uncited, captions missing units or sample sizes, and numbering that has drifted out of order.
Abstract and title compliance
Required abstract components, structured-abstract subheadings where the journal uses them, word limits, and titles too vague to describe the scope.
Section structure and headings
Section presence and sequence, heading hierarchy and depth, headings that do not describe their content, and capitalisation inconsistency across the document.
Funding statement
Whether the funding declaration is present, whether grant numbers and agency names are included, and whether multiple funders are formatted consistently. Funding is the only mandatory declaration the engine checks.
See all 95 checks
The screening pass is not limited to these: the full catalogue also covers language, acronyms, keywords and author metadata.
Desk Screening Automation for Journals questions
In the self-serve product, whoever uploads the manuscript gets the report, which for a journal usually means one editorial assistant screening and forwarding by email. For partner journals the pass runs on the intake as a whole: submissions go in as a batch, every report lands where the whole team can see it, and a manuscript screened once is settled rather than re-screened by the next editor who touches it. Team access itself is covered on the journals page.