Do I need to cite absolutely everything in a thesis?

The rule most people are taught is "cite anything that is not your own idea", which is true and unusable, because almost nothing in a literature review is your own idea and you cannot cite the whole field in every sentence. A more workable version asks what the citation is doing.

Citations do three jobs. They give credit for someone's specific contribution. They provide evidence for a claim a reader might doubt. And they point the reader to where they can follow up. If a sentence needs none of those, it probably needs no citation.

If you want a checklist instead of a principle, Penders gives one. Cite when previous work supplies the foundation, the rationale, the idea, the question, the method, or the critique behind what you are arguing (Penders, 2018, p. 2). Anything outside those six roles is usually yours. The same rule covers your own earlier papers, which Penders says you should cite whenever they are relevant and necessary to the argument (Penders, 2018, p. 3), so a chapter reworked from your own published article still needs the reference.

That resolves most of the hard cases. A definition you took from a particular author, phrased their way, needs a citation because it is their contribution. A general statement that a concept exists and is studied does not. A number needs a citation because a reader would want to check it. Your interpretation of two studies needs no citation because it is yours, and attaching one would misattribute your work.

The sentences to watch are the transitions. "This suggests that", "it therefore appears", "the field has moved towards". Those are your voice, and putting a citation at the end of them quietly hands your analytical work to someone else.

Lingard pushes back on treating citation as plagiarism insurance. In a literature review its job is to show where the knowledge came from, how it developed through disagreement, and where you stand inside that history, and stripping the citations out makes claims read as common sense or as personal opinion (Lingard, 2018, Part II, p. 133). Lingard also asks you to widen your stock of reporting verbs so the sentence says what the source actually did, while noting that this vocabulary is a flexible tool rather than a fixed rule set (Lingard, 2018, Part II, p. 134). "Argues", "reports", and "findings suggest" are not interchangeable, and picking the accurate one often does the work you were trying to do by adding a second citation. Penders makes the matching demand on your prose, calling it citing transparently: the sentence should tell the reader the source's relevance, contribution, and significance (Penders, 2018, p. 3).

What counts as common knowledge in my field?

Anything an examiner in your discipline would accept without a source and could find in any introductory textbook. The boundary moves by field and by audience, so when unsure, ask whether a reasonable specialist could dispute it. If they could, cite it.

Be careful with claims that feel common because you have read them fifty times. Repetition inside a literature is not the same as consensus, and a statement everyone repeats sometimes traces back to a single old study that no longer holds. Those are the sentences most worth checking, and they are also the ones most likely to embarrass you in a viva.

The safe default in doctoral work is to cite slightly more than you would in a journal article. Thesis examiners are testing whether you know where knowledge came from, and a supported claim is never criticized for being supported.

Getting the boundary wrong is not automatically misconduct. Sozon and colleagues, reviewing the literature on cheating and plagiarism, list improper citation, honest mistakes, and different cultural understandings of academic integrity among the causes of plagiarism, alongside deliberate deception (Sozon et al., 2024, p. 5). The reason to still take attribution seriously is in the same review, which reports that 17% of graduate students admitted cheating on tests and 40% on written assignments (Sozon et al., 2024, p. 15). Examiners read your citations against that background, so an unsourced contested claim costs you more than an extra reference ever will.

Is it possible to over-cite?

Yes. A citation after every clause slows reading and signals that you do not trust your own voice. It can also obscure your contribution, since a reader who sees a source on every sentence cannot tell which sentences are yours. Over-citing reads as anxious rather than as rigorous.

The usual cause is uncertainty about the boundary rather than genuine belief that every claim needs support. Deciding the boundary explicitly, once, and applying it consistently, produces a chapter that reads with more confidence than one where the density varies by how nervous you were that day.

There is one place to be generous without hesitation: contested claims and numbers. Nobody has ever been criticized for citing a disputed finding too carefully.