Citation Errors That Get Papers Desk Rejected

A desk rejection is one of the most demoralising outcomes in academic publishing. Your paper doesn't fail peer review — it never reaches peer review. Within days of submission, it comes back with a terse note explaining that the manuscript "does not meet the journal's technical requirements for submission." Months of research and writing, dismissed at the first gate.

What many researchers don't realise is how often citation errors are the direct cause. Journal editorial offices perform a technical check on every submission before it is assigned to a handling editor or sent to reviewers. This check is partly automated and partly manual, and it specifically looks at your reference list. Wrong citation style, missing DOIs, a reference limit that's been exceeded, a mismatched author name — any of these can trigger a return without review.

The cruel irony is that these errors have nothing to do with the quality of your research. A genuinely important study can be desk rejected because its reference list uses the wrong DOI format. The research is fine. The citations are not.

This guide covers every citation error type that commonly triggers desk rejection: what it looks like, why editors catch it, and exactly how to fix it before you submit. The errors are ordered from most common to most damaging — though in practice, any one of them is enough to have your manuscript returned.

This is the single most common cause of a technical desk rejection, and it is entirely preventable. Every journal specifies a required citation style in its author guidelines — APA, Vancouver, AMA, IEEE, Chicago, or a custom house style — and submissions that use a different format are frequently returned before an editor reads a word of the manuscript itself.

The mistake happens because researchers write manuscripts for one journal, get rejected on scientific grounds, and then resubmit to a different journal without reformatting the references. A paper rejected by a psychology journal (APA format) that gets resubmitted to a clinical medicine journal (Vancouver format) with its reference list unchanged will almost certainly be desk rejected on technical grounds.

It also happens when researchers rely on a style they know well rather than checking the target journal's requirements. An engineer accustomed to IEEE format submitting to a multidisciplinary journal that requires APA will have references that look obviously wrong to an editorial assistant within thirty seconds.

Most journals impose explicit reference limits by article type, and exceeding them is grounds for immediate return. Nature typically allows 30–50 references for standard research articles. NEJM limits letters to five references. Many high-impact journals cap research articles at 40–60 references and apply these limits strictly.

The error usually happens when authors submit a manuscript that has grown organically over time without anyone specifically counting references against the stated limit. A discussion section that started with 30 citations and accumulated 15 more through revision rounds can push a manuscript well over the limit of a journal that allows only 35. The editorial office counts the references before anything else — this check takes under a minute and requires no scientific judgment.

Some journals are strict about this to the point of stating explicitly in their rejection letter that the manuscript was returned due to exceeding the reference limit, with no scientific evaluation performed. This means your research may be perfectly suited to the journal but will be returned without review until you reduce your reference count.

Most journals now require DOIs for all journal articles where a DOI exists, and many perform automated DOI validation as part of their submission system. A DOI that returns a 404 error, routes to the wrong paper, or is formatted incorrectly (wrong prefix, extra spaces, wrong URL structure) can flag your manuscript for immediate editorial review or automatic return.

This error has become significantly more common with the rise of AI writing tools. AI assistants regularly generate DOIs that are syntactically correct — they have the right format and a plausible publisher prefix — but resolve to nothing when clicked. A submission containing multiple AI-hallucinated DOIs that fail automated validation will be flagged before a human editor sees the paper.

Beyond fabricated DOIs, there are genuine formatting errors that trip up even careful researchers. APA 7th edition requires DOIs as full URLs (https://doi.org/...) while AMA requires the doi: prefix format without a URL. Submitting an APA-formatted paper to a journal that uses AMA style — or simply using the wrong DOI format within the correct style — creates an immediately visible formatting error.

Citing retracted research is a serious citation error that an increasing number of journals screen for before peer review. Journals participating in the Crossmark programme, which includes most major publishers, can query retraction status automatically during manuscript processing. A manuscript that cites multiple retracted papers may be returned with a request to remove or address these citations before review.

The problem is more common than most researchers expect. The Retraction Watch database now tracks tens of thousands of retracted papers, and studies consistently find that retracted papers continue to receive positive citations for years — sometimes decades — after retraction notices are published. Researchers who saved a PDF at the time of first access and never rechecked the paper's status may be citing work that was retracted two years ago without knowing it.

In clinical and biomedical research, this error carries particular weight. A drug safety claim cited to a retracted paper, or an efficacy figure drawn from research that was withdrawn due to data fabrication, creates not just a desk rejection risk but a potential patient safety issue that editors take very seriously.

For journals using numbered citation styles — Vancouver, AMA, IEEE — the reference list must be numbered in the exact order that sources first appear in the text. Reference [1] should be the first citation in the manuscript, reference [2] the second new source cited, and so on through to the final reference. Any gap, duplication, or out-of-sequence numbering is an immediate red flag.

This error almost always happens during revision. A manuscript that was correctly numbered at first submission gets revised: new references are added to the introduction, old references are deleted from the methods, and a section is moved from the results to the discussion. By the end, reference [8] appears before reference [4] in the text, reference [12] was deleted but the number was never reassigned, and references [19] and [20] are the same paper cited twice under different numbers.

Editorial assistants performing the technical check on a Vancouver or AMA manuscript will scan the in-text citation numbers against the reference list. A sequence that reads [1], [2], [5], [3], [7] in the first two pages signals that the manuscript has not been properly formatted for submission and will be returned.

Biomedical and clinical journals using Vancouver or AMA format require NLM (National Library of Medicine) standard journal title abbreviations. Engineering and technology journals using IEEE format require IEEE-standard abbreviations. Using full journal names where abbreviations are required, using non-standard abbreviations, or mixing abbreviated and non-abbreviated journal names throughout a reference list are all errors that get caught during technical checks.

The error often originates with reference managers. Zotero, Mendeley, and EndNote import journal metadata from databases, and that metadata frequently contains full journal names, non-standard abbreviations, or abbreviation formats that don't match the target journal's requirements. Researchers who trust reference manager output without verifying it against official abbreviation lists submit manuscripts with obviously incorrect journal formatting.

Every citation style specifies which elements are required for each source type. For journal articles in Vancouver style, that means author names, article title, abbreviated journal name, year, volume, issue, and page range — plus a DOI where available. An incomplete reference missing any one of these elements fails a technical check and, depending on the journal, may result in desk rejection.

The most commonly missing elements are page ranges (absent when a reference was added from an abstract view rather than the full paper), volume and issue numbers (omitted from references to online-ahead-of-print articles that were never updated when the paper was formally published), and DOIs (skipped because they weren't immediately visible in the database export).

Grant proposals submitted to NIH face an additional specific requirement: papers arising from NIH-funded research that are cited in the proposal must include their PubMed Central ID (PMCID). Missing PMCIDs in NIH grant applications is a compliance failure that results in the application being returned administratively — the equivalent of a desk rejection in grant review.

An in-text citation that says "(García, 2021)" should correspond to a reference list entry for García from 2021. An in-text citation that says "[14]" should point to reference number 14 in the reference list. When these don't match — a name is spelled differently, a year is wrong, a number points to the wrong entry — the mismatch is visible to any editor who spot-checks citations against the reference list.

Mismatches accumulate through the revision process. An author's name gets corrected in the reference list but not in an in-text citation. A publication year is updated in one place but not both. A section gets cut but the reference number in a figure caption doesn't get updated. None of these are hard to create — and none are hard to catch with a systematic check.

Self-citation becomes a problem when it crosses from legitimate contextualisation of your own prior work into citation padding. While there is no universal threshold, editorial offices at major journals are alert to manuscripts where a disproportionate share of the reference list consists of the submitting authors' own prior publications. Some journals explicitly state in their author guidelines that self-citation should not exceed a specified percentage of the total reference list.

The concern is twofold. First, a heavily self-cited paper signals that its literature review may be narrow or biased. Second, excessive self-citation can be used to manipulate journal impact metrics and individual h-index scores — something journal editors are increasingly sensitised to.

A manuscript from a research group with 60 references, of which 20 are to the authors' own prior work, will attract editorial scrutiny even if none of those self-citations are inappropriate individually. The aggregate pattern is the problem.

Preprints (arXiv, bioRxiv, medRxiv, SSRN) are a legitimate and valuable part of scientific communication, but their role in reference lists depends heavily on the journal and the nature of the claim being supported. Many journals — particularly in clinical medicine — explicitly prohibit citing preprints as primary evidence for key claims, or require that preprint citations be clearly labelled as such.

The error takes two forms. The first is citing a preprint that has since been peer-reviewed and published in final form without updating the citation. Using an arXiv citation when the paper has been published in a journal for two years suggests the author hasn't engaged with the current literature. The second is citing a preprint for a critical claim in a journal that explicitly requires peer-reviewed sources for such claims — which may result in the paper being returned with a request to find peer-reviewed alternatives or remove the claim.

The Pre-Submission Citation Checklist

Run through this checklist on every manuscript before submission. Completing it takes one to two hours for a standard-length paper — a fraction of the time that will be lost to a desk rejection and resubmission cycle.

Conclusion: Desk Rejection Is Preventable

Every error covered in this guide is detectable before submission. None of them require specialist knowledge to catch — they require only a structured approach and enough time to complete the check properly. The reason desk rejections for citation errors happen so frequently is not that the errors are subtle or hard to see. It's that researchers deprioritise reference checking relative to the scientific content of their manuscripts, leaving it as a last-minute task when time pressure is highest.

The editors performing these technical checks are often looking for exactly the same things you'd find if you spent thirty minutes with your reference list and a copy of your target journal's author guidelines. The difference is that when an editor finds these errors, your paper comes back. When you find them, you fix them.

Build reference verification into your manuscript preparation workflow as a non-negotiable step — not an afterthought. A clean reference list doesn't guarantee acceptance. But a broken one guarantees a desk rejection, and that's an outcome entirely within your control to prevent.