You changed every third word, rearranged a clause, and put the citation at the end. That is not paraphrasing. It has a name, researchers have been studying it since 1993, and it is the most common source-use problem in student writing by a wide margin.

It is called patchwriting, and the reason it matters is that most people doing it believe they are doing the right thing. There is no intent to deceive anywhere in the process, which is exactly why it survives all the way to submission.

Where the term comes from

Rebecca Moore Howard coined patchwriting in 1993 to describe a specific behavior: copying from a source and then deleting some words, altering the grammar, or substituting synonyms one for one. Merriam-Webster's write-up of the term traces the same origin.

What made Howard's framing useful was the refusal to call it simple cheating. She described it as a developmental stage: what writers do when they are working with material they do not yet fully understand. You cannot restate an argument in your own words if the argument has not become yours yet, so you do the next best thing and rearrange the author's.

How common it actually is

The Citation Project examined 174 researched papers from first-year students across 16 US institutions, coding how each citation used its source. The results are worth sitting with:

In 94% of citations, students were working from just one or two sentences of the source.
70% of citations came from the first or second page of the source.
52% of papers contained at least one instance of patchwriting.
In an earlier close study of student papers, not one summarized a source.

Read together, those numbers describe something other than a discipline problem. Students were mining sources for usable sentences near the front rather than reading them through, which is a research-skills problem that then shows up as a citation problem. The published analysis makes the same argument: the pattern points at how source use gets taught, not at student character.

What it looks like

Take this source sentence:

Coastal cities have delayed flood defenses for decades because the costs arrive immediately while the benefits appear only in a future that no serving official will be held responsible for.

Patchwriting:

Coastal towns have postponed flood protection for decades since the expenses come at once while the advantages emerge only in a future no current official will be accountable for (Author 8).

Cited, reworded, and still patchwriting. The clause order is untouched, the "because X while Y" logic is untouched, and the sentence breathes exactly like the original.

A real paraphrase:

The political incentives run backwards on flood defense: an official who funds it pays the full price during their own term and hands the reward to a successor decades away, which is why coastal cities keep postponing the work (Author 8).

The second version required understanding the mechanism, not finding synonyms. That is the entire difference.

Is it plagiarism?

It depends on your institution, and this is genuinely unsettled ground.

Some policies name patchwriting explicitly and treat it as a form of plagiarism regardless of citation. Others handle a cited patchwrite as a citation error to be corrected rather than misconduct to be punished, and reserve the plagiarism label for uncited work. Howard's own position, and that of many writing instructors, is that penalizing a developmental stage teaches nothing.

What is not in dispute: uncited patchwriting is plagiarism everywhere. If the citation is missing, the discussion about intent and development does not begin.

So the practical answer for anyone submitting work is that the citation protects you from the serious version, and it does not make the passage acceptable writing. Both things are true at once.

What it is not

Three things get confused with patchwriting, and telling them apart makes the concept usable.

Not a short quotation. The author's exact words inside quotation marks, with a citation, is simply correct practice. Patchwriting is what you get when those quotation marks come off and a few words get swapped so the sentence stops looking like a quote.

Not shared technical vocabulary. Some phrases have no synonyms worth using. "Randomized controlled trial" is what the thing is called, and writing around it produces worse prose, not more original prose. Fields have fixed terms and using them correctly is competence.

Not agreeing with your source. Restating a claim you accept, in genuinely your own structure, with a citation, is ordinary scholarship. The problem was never sharing the author's conclusion. It is borrowing the author's sentence while presenting it as your writing.

It is not only a student problem

Patchwriting gets discussed as something first-years do, which lets everyone else off too easily. It shows up in professional journalism, in policy reports, and in published academic work, usually for the same reason: someone was writing to a deadline about material they had skimmed.

The professional cases tend to be more damaging precisely because the defense that works for a student, that they were still learning to handle sources, is unavailable to someone paid for expertise. A staff writer who reshapes a competitor's paragraph and files it under their own byline has done the same mechanical thing as a nervous undergraduate, minus the developmental excuse.

If you write anything sourced for money, the same test applies. Could a reader put your paragraph beside the source and see one behind the other?

Why checkers catch it

Matching software looks for runs of consecutive words shared between documents, so a passage where you replaced isolated words still contains plenty of intact three and four word runs between your substitutions. That is enough for the software, which is why patchwritten passages surface in reports that their authors expected to come back clean. If you want the mechanism in detail, it follows directly from how checkers turn text into fingerprints.

The useful consequence: patchwriting is one of the few writing problems a machine reliably flags before a human sees it. Scan a draft, and the passages that come back yellow are usually the ones where your rewriting stayed too close.

How to stop doing it

The habit is caused by writing with the source visible. Every fix is a version of separating those two moments.

Read the whole source before you write anything from it. The Citation Project's finding about first pages is the tell. If you only ever read page one, page one's sentences are all you have to work with.
Close the source before you write your version. You cannot copy the syntax of a sentence you cannot see.
Summarize more than you paraphrase. Compressing an argument into two sentences is nearly impossible to fake, and it is the skill the research found missing.
If a phrase is worth keeping exactly, quote it. A short quotation with quotation marks is honest. A disguised one is the problem.
Check the passages you rewrote hardest, not the ones you wrote quickly. Struggle is a reliable signal that the source's structure is still there.

The part worth remembering

Patchwriting is a comprehension problem wearing a citation problem's clothes. When you find it in your own draft, the useful question is not which words to swap next, it is whether you actually understood the passage well enough to say it your own way.

Usually the answer is that you half understood it, which is fine and fixable. Go back and read the source properly, then write the sentence from what you know. Once you understand something, writing it in your own words stops being a task you have to perform, and a checker's yellow highlights become easy to sort: the ones left are common phrasing and your own citations.