Quoting is the one part of academic writing where a small mechanical slip — a missing page number, an unmarked omission, an unattributed paraphrase that drifts too close to the original wording — can turn into a misconduct finding. This guide covers the mechanics of quoting correctly across the major style conventions, and the parts of quotation practice that a generic citation-generator page won’t tell you: what happens when you quote a study participant instead of a published author, when copyright law limits how much you’re allowed to reproduce, and how quotation errors escalate into plagiarism findings.
Quote, paraphrase, or summarise?
Before formatting a single quotation mark, decide whether you should be quoting at all. The three ways of incorporating a source are not interchangeable, and choosing the wrong one is a common weakness reviewers flag in scientific and social-science writing.
- Quoting reproduces a source’s exact wording. Reserve it for cases where the original phrasing itself is the evidence: a precise definition, a legal or regulatory clause, a striking or contested claim you need to analyse word-for-word, or — in qualitative research — a participant’s own words, where the voice is the data.
- Paraphrasing restates a source’s idea in your own words and sentence structure, at roughly the same level of detail, with a citation. This is the default for reporting prior findings, methods, and background in most scientific and social-science writing.
- Summarising condenses a source’s overall argument or findings into a much shorter form, with a citation. Use it for literature-review coverage of work that’s relevant but not being closely analysed.
Over-quoting is a recognised weakness in scientific writing, not a stylistic preference. A manuscript built from strung-together quotations reads as under-processed — it shows that the author hasn’t fully digested the source material, and it shifts analytical work onto the reader. Style guides and academic writing centres consistently advise using direct quotation sparingly outside of textual, legal, or qualitative-data analysis, and paraphrasing or summarising everywhere else. A rough rule many journals apply informally: if you can restate the point accurately without losing anything essential, paraphrase it.
Quotation mechanics: short quotes vs. block quotes
The three style families most common in research writing — APA, MLA, and Chicago — agree on the underlying logic (short quotations run in-line with quotation marks; long quotations are set off as an indented block without quotation marks) but differ on the exact length threshold and formatting.
| Style | Threshold for a block quote | Block quote formatting |
|---|---|---|
| APA (7th ed.) | 40 words or more | Indented 0.5 in. from the left margin, double-spaced, no quotation marks; citation follows the final punctuation. |
| MLA (9th ed.) | More than 4 lines of prose (or 3 lines of verse) | Indented 0.5 in., double-spaced, no quotation marks; parenthetical citation follows the final punctuation. |
| Chicago (author-date and notes-bibliography) | Roughly 100 words or more, or 5+ lines | Indented as a block, typically single-spaced with a blank line before and after; no quotation marks. |
Two mechanical points trip authors up repeatedly:
- Don’t double up. The indentation itself signals a quotation in block format — adding quotation marks around an indented block is a formatting error, not extra emphasis.
- Citation placement differs by context. For a short, in-line quotation, the citation sits inside the closing punctuation of the sentence (or per your style’s convention). For a block quote, the parenthetical citation follows the quotation’s own closing punctuation, outside it, with no period after the parenthesis.
Illustrative example (composite, not a real citation)
A short, in-line quotation in APA style: Smith and Delgado define the concept succinctly, noting that “replication is not repetition; it is an independent test of a claim’s robustness under new conditions” (Smith & Delgado, 2023, p. 14).
The same source quoted at length would be set as a block quote once it crosses roughly 40 words, with the citation following the final period:
Replication is not repetition; it is an independent test of a claim’s robustness under new conditions. A replication that uses identical materials and procedures tells us whether an effect can be reproduced under the same operational choices, while a conceptual replication tells us whether the underlying phenomenon holds when those choices vary. (Smith & Delgado, 2023, p. 14)
This example is written to illustrate formatting only; “Smith & Delgado, 2023” is not a real, citable source.
Ellipses, brackets, and [sic]: marking what you changed
Any time you alter quoted text — by cutting a portion, inserting a clarifying word, or reproducing an error — the change must be visibly marked. Silent alteration of a direct quotation, even a minor one, is treated as a quotation-accuracy problem, and in some cases as a form of misrepresentation.
- Ellipses (…) mark an omission. Use three spaced periods for an omission within a sentence; when the omission spans a sentence boundary, most styles (APA, MLA) still use three periods, placed after the preceding sentence’s own closing punctuation. Never use an ellipsis to omit words in a way that reverses or meaningfully changes the source’s meaning — that crosses from abridgement into misquotation.
- Square brackets [ ] mark an insertion or substitution — typically to make a quotation grammatically fit your sentence, or to clarify an ambiguous pronoun or term. Example: the original “it was rejected by the committee” becomes “[the proposal] was rejected by the committee” if quoted out of its original context.
- [sic] (Latin for “thus”) follows an error in the original — a misspelling, grammatical mistake, or factual slip — to signal that the error belongs to the source, not to you as the transcriber. It is not a tool for editorial commentary or sarcasm; use it sparingly and only where the error might otherwise look like a transcription mistake on your part.
- Emphasis. If the source itself uses italics or bold for emphasis, reproduce it as-is. If you add emphasis that wasn’t in the original (to highlight a specific phrase for your argument), you must flag it — conventionally with a bracketed note or a footnote reading “emphasis added” (and, conversely, “emphasis in original” when you’re confirming the source’s own formatting is unchanged).
Citing quotations without page numbers
Page numbers are a print-era convention that doesn’t always map onto web pages, video, audio, or interview data. Each major style has a substitute locator:
- Web pages without pagination: cite a paragraph number (¶ or “para.”) if the source numbers its own paragraphs or sections, or a section heading, if neither is available. Some styles accept omitting a locator entirely for short, unpaginated web sources, but a heading or paragraph count is more precise and more defensible if a reviewer asks you to verify the quotation.
- Video and audio: cite a timestamp (e.g., 12:47 or 00:12:47) marking where the quoted words occur, in place of a page number.
- Interviews and oral history: cite by date and, where a transcript exists, a transcript page or line number; where no public transcript exists, note that the material comes from a personal communication or primary interview record you conducted or have access to.
- E-books without fixed pagination: use a chapter or section identifier, or the platform’s location marker, and note in your reference list that pagination varies by device/edition.
Whatever locator you use, the goal is the same one page numbers serve: letting a reader find the exact passage you quoted, not just the source it came from.
Quoting research participants: consent and anonymisation
Quoting a study participant in a qualitative paper is not the same act as quoting a published author, and treating it as though it were is a common source of avoidable ethical exposure. A participant’s words are personal data, and using them as a direct quotation carries obligations that quoting a journal article does not:
- Consent must cover quotation specifically. General consent to “participate in the study” does not automatically authorise publishing a participant’s verbatim words. Many institutional review boards and research ethics committees expect the consent process (and the consent form itself) to state explicitly whether direct quotes may be used in publications and presentations, and whether the participant will have an opportunity to review quotes attributed to them before publication.
- Anonymisation of quoted material goes beyond removing a name. A verbatim quote can contain enough contextual detail — a job title, a rare diagnosis, a specific location, an unusual life event — to make a participant identifiable to people in their own community even with names and obvious identifiers stripped out. This is a well-recognised risk in small-sample qualitative research and in studies of small or tightly networked populations, and it typically requires deliberately altering or generalising identifying details within the quote itself, not just replacing the name with a pseudonym or participant number.
- Attribution conventions differ from citing a published author. Participants are usually identified by pseudonym, role, or a participant ID (e.g., “Participant 7,” “a mid-career nurse in the intervention arm”) rather than by real name, and the attribution convention should be established and applied consistently, and disclosed in your methods section.
- Withdrawal of consent applies retroactively where feasible. If a participant withdraws consent after providing a quotable statement but before publication, that withdrawal generally needs to be honoured for material not yet published, consistent with your ethics approval and applicable consent framework.
These obligations sit alongside, not instead of, the general informed-consent framework covering study participation — see CASRAI’s guide to the components of informed consent and its related coverage of the four principles of informed consent for the underlying consent requirements, and the dictionary entry on informed consent for the core definition. For the technical side of protecting identity in quoted or reported material, see CASRAI’s guide to data anonymisation in research.
Secondary citation: quoting a quote
Sometimes a source you’re reading quotes or paraphrases another source you don’t have direct access to. Citing that original source as if you’d read it yourself — rather than citing it “as cited in” the source you actually read — misrepresents what you verified. All three major style families provide a secondary-citation format for exactly this situation (commonly rendered “as cited in” or “qtd. in”), but every one of them also advises using it only when the original source is genuinely unavailable, not as a shortcut.
The reasons to avoid secondary citation where possible:
- Quotations and paraphrases get distorted as they pass through intermediate sources — a claim can be simplified, exaggerated, or subtly altered each time it’s re-quoted, and errors compound.
- You cannot verify the accuracy of a quotation, its context, or its page locator without the original.
- Reviewers and editors increasingly flag chains of secondary citation as a sign the literature review wasn’t done from primary sources.
Where the original source is out of print, in an inaccessible archive, or in a language you can’t verify, secondary citation is legitimate — just cite it explicitly in that form (e.g., “Original Author, Year, as cited in Secondary Author, Year”) rather than presenting the original as something you consulted directly.
Copyright and permissions for quoted material
Quotation length that’s fine under a style guide’s block-quote threshold is not automatically fine under copyright law — these are two separate constraints, and researchers sometimes assume clearing one clears both.
- Fair use (US) and fair dealing (UK/Commonwealth) are narrower than “cited properly.” US fair use is a multi-factor balancing test (purpose, nature of the work, amount used, market effect) with no fixed word-count safe harbor, despite persistent informal rules of thumb. UK and Commonwealth fair dealing, under the Copyright, Designs and Patents Act 1988, is purpose-limited rather than a general balancing test — permitted purposes include non-commercial research, private study, criticism, review, and quotation, and a use outside those listed purposes isn’t covered regardless of how little text was taken. Reproducing a long prose excerpt, an entire poem, a full figure, or a complete table is far more likely to exceed either doctrine than a short prose quotation used to support analysis.
- Figures and tables reproduced from another publication almost always need explicit permission, distinct from the quotation allowances that cover short text excerpts — publishers generally treat reproducing a figure or table as a reuse request, not a citation.
- Publisher permission workflows exist for exactly this. Most major publishers route reuse requests through a licensing service — Elsevier, Springer Nature, Wiley and others commonly use RightsLink — which generates a permission record and, where required, a specific credit line you must reproduce alongside the reused material. See CASRAI’s dictionary entries on copyright, copyright transfer agreements, and publication permission for the underlying mechanics.
- When in doubt, request permission rather than rely on fair use/fair dealing — particularly for anything beyond a short prose excerpt, and always for figures, tables, and any material you intend to reproduce in full.
How quotation errors become integrity findings
Quotation problems don’t only look like sloppy formatting — several of them map directly onto categories of research misconduct or publication-ethics violation that journals and institutions actively screen for:
- Unmarked verbatim text is textual plagiarism, regardless of whether it was reproduced deliberately or through careless paraphrasing that stayed too close to the source’s original wording and sentence structure. See CASRAI’s dictionary entry on textual plagiarism and the broader taxonomy of plagiarism types for how similarity-detection tools and editors distinguish acceptable paraphrase from disguised copying.
- Reusing your own previously published text without disclosure — including quoting your own earlier work at length as if it were new — falls under self-plagiarism/text recycling, which journals evaluate differently from plagiarising another author’s work but still treat as a disclosure failure.
- Misquotation — silently altering a quotation’s wording, or quoting selectively (e.g., trimming a quotation via an unmarked ellipsis in a way that reverses its meaning) — is a distinct integrity concern from plagiarism: the issue isn’t unattributed copying but misrepresenting what a source actually said. It can surface in post-publication corrections, letters to the editor, or formal integrity investigations when a reader traces a quotation back to its original and finds it doesn’t match.
- Fabricated or unverifiable quotations — attributing words to a source, or a study participant, that cannot be verified against a real transcript or document — are treated as a fabrication concern, the most serious tier of research misconduct findings.
- AI-generated paraphrasing tools carry their own quotation-adjacent risk: a tool that “rewrites” a passage closely enough that the result is still substantively the source’s structure and phrasing, just with synonyms swapped in, can still register as plagiarism under similarity-detection review. See CASRAI’s guide to AI paraphrasing tools and plagiarism risk.
The practical safeguard against all of the above is the same one that makes quotations useful in the first place: quote sparingly, quote exactly, mark every change, and always keep a locator precise enough that someone else can find the exact passage and confirm it says what you say it says.
Frequently asked questions
How many words can I quote before I need a block quote?
It depends on the style: APA’s threshold is 40 words, MLA’s is more than four lines of prose (three lines of verse), and Chicago’s is roughly 100 words or five-plus lines. Below the threshold, quote in-line with quotation marks; at or above it, format as an indented block without quotation marks.
Do I need permission to quote a short passage from a journal article?
A short prose excerpt used to support analysis and properly cited is generally covered by fair use (US) or fair dealing (UK/Commonwealth), which both recognise quotation for criticism, review, and research as a permitted purpose. Longer excerpts, whole figures, whole tables, or reproducing a substantial portion of a short work (like a poem) usually require the publisher’s explicit permission regardless of citation.
Can I quote something from an interview I didn’t record?
Only if you have a reliable basis for the quotation’s accuracy — contemporaneous notes, a transcript, or a recording. Journals and integrity reviewers can ask you to substantiate a quoted statement; “reconstructed from memory” is a weak basis for presenting text in quotation marks, and you should consider paraphrasing (with appropriate hedging) instead if you can’t verify exact wording.
Is it plagiarism if I cite the source but don’t use quotation marks around copied text?
Yes. A citation tells the reader where an idea came from; quotation marks (or block formatting) tell the reader that the exact wording is copied. Citing a source without marking copied text as a direct quotation is a recognised form of plagiarism — sometimes called patchwriting when phrases are lightly altered but sentence structure stays close to the original — even though a citation is present.
What’s the difference between “as cited in” and a direct citation?
“As cited in” (or “qtd. in”) tells the reader you read the quotation in an intermediate source, not the original — use it when you genuinely cannot access the original work. Citing the original source directly asserts that you consulted it yourself; doing so when you actually only read a secondary source’s quotation of it misrepresents your research process.
