How much self-citation is acceptable?

Self-citation has a bad reputation it partly deserves and mostly does not. In a doctoral thesis, or in a paper extending your own earlier work, citing yourself is not merely acceptable but required, because failing to do so would misrepresent where the work came from. A programme of research builds on itself, and hiding that is worse than showing it.

What gives the practice its reputation is the other kind, where citations are added to inflate a metric or to advertise. That version is visible to readers because the citations do no work in the sentences they appear in. A reader can tell the difference, and so can a reviewer.

If you want numbers to compare yourself with, the largest study of the question gives ranges rather than limits. Ioannidis, Baas, Klavans and Boyack measured the top 100,000 authors in Scopus and found a median career-long self-citation rate of 12.7%, with half of those authors falling between 8.6% and 17.7% (Ioannidis et al., 2019, p. 2). For 2017 alone the median dropped to 9.2%, with an interquartile range of 4.8% to 14.7%. These figures describe heavily cited senior researchers rather than graduate students, so read them as context and not as a quota you need to hit.

The same dataset shows where scrutiny starts. Roughly 8,600 of those 100,000 authors exceeded 25% self-citation across their careers, and about 1,100 exceeded 40% (Ioannidis et al., 2019, p. 2). The authors stop short of calling any of that misconduct. They recommend examining extreme cases one by one instead of deleting self-citations from metrics wholesale. Treat a proportion above a quarter as a reason to check what each of your citations is doing, not as evidence that something is wrong.

For a thesis by publication, self-citation is structural. Your chapters cite each other, and they cite the papers those chapters became. State clearly at the start how the published work relates to the thesis, since examiners will want that mapping and providing it removes any suggestion of concealment.

There is one situation that needs care: citing your own work for a claim it does not support. The scrutiny you apply to other people's papers applies to yours, and the fact that you wrote it does not exempt it from being checked.

Penders draws the line by function. Self-citation is legitimate when it documents incremental iterative advancement, and it becomes a problem when it turns into salami slicing, where one study is split across several papers to raise publication and citation counts (Penders, 2018, pp. 2-3). Omission carries its own cost. Leaving out your own relevant earlier work implies that the current paper originated an idea, question, method or critique that it did not.

Will an editor or reviewer flag it?

They can, and some journals screen for it. What draws attention is a pattern rather than a number: several of your own papers cited for claims that do not need them, or citations to your work that the argument does not use. A high proportion in a narrow specialism is normal and usually passes without comment.

Coercive citation, where a reviewer or editor asks you to add citations to their own work without a substantive reason, is the mirror image of the problem and is treated as a publication ethics issue. If it happens, you are entitled to decline and to explain why in your response letter.

For anonymized review, replace self-citations that would identify you with a neutral form for submission and restore them afterwards, following the journal's instructions.

Norms are local before they are disciplinary. Penders observes that referencing practice varies between laboratories, departments and fields, so molecular biology papers routinely place several references after a single statement while research ethics and science policy expect each reference to be given contextual explanation (Penders, 2018, p. 5). Ioannidis and colleagues make the matching point from the data side, warning that comparisons across fields are tenuous because citation densities differ and databases treat self-citations inconsistently (Ioannidis et al., 2019, p. 1). Check what your own subfield does before deciding your rate is high.

How do I cite my own unpublished work?

Cite it with its actual status. A manuscript under review is cited as unpublished or under review, a preprint is cited as a preprint with its identifier, and a thesis chapter is cited as part of the thesis. Never cite in-progress work as though it were published.

Status changes during a long project, so keep these citations in one place and check them at final submission. A reference list containing three papers described as under review that appeared eighteen months ago looks careless.

Where the work is genuinely inaccessible to a reader, consider whether you need the citation at all. If the claim matters, summarize the evidence in the text so a reader can evaluate it without waiting for publication.