Direct answer
A plagiarism check compares your document against a large database of student papers, web pages and academic publications, then reports the percentage of your text that matches those sources as a similarity score. That score is not a verdict: it tells you how much of your writing overlaps with existing material, and it must be read alongside the highlighted matches to see whether the overlap is properly quoted, cited or genuinely unoriginal [1]. Turnitin is the system most universities run, and its report shows the similarity percentage, the matched sources and a separate AI writing indicator in the same document [1][4].
What Does a Plagiarism Check Actually Measure?
A plagiarism check measures textual overlap, not intent and not originality of thought. Turnitin compares your submission against its repository of previously submitted student papers, archived and live internet content, and scholarly publications, then flags every passage that resembles something in that index [2]. The headline similarity percentage is an aggregate: it is calculated from all matched text across all sources combined, so a paper that borrows three sentences from five different articles can show the same number as one that copies a single long paragraph [2].
The number is also sensitive to how the report is configured. Instructors can exclude quoted material, exclude the bibliography, and exclude matches below a word-count threshold, so the identical document can display a materially different percentage depending on whose settings are applied [2]. This is why two students comparing screenshots of "the same" score are often comparing two different measurements.
What the report body adds is provenance. Each match is listed with its source, its location in your document and the number of matched words, which lets you distinguish a correctly attributed quotation from an unattributed block [1]. Reading the source list is the actual check; the percentage is only the summary line.
Why Can a Plagiarism Checker Flag Text That I Wrote Myself?
Genuine writers get flagged constantly, and most of the time the cause is structural rather than dishonest. Standard methodology phrasing, discipline-specific terminology, statutory definitions, quoted legislation and reference lists all repeat across thousands of papers, so they register as matches even when you composed every word yourself [3]. A reference list alone can push a similarity score into double digits on a paper with no copying at all.
Self-matching is the second common cause. If you submitted an earlier draft, a group project or a previous assignment through the same institutional system, that text now sits in the student paper repository, and your new submission can match your own prior work [3]. This is why a final-year student sometimes sees a high score on a chapter they wrote from scratch two semesters earlier.
The third cause is settings, not text. Because quote, bibliography and small-match exclusions are controlled by the instructor or administrator, a paper that looks clean under one configuration can look compromised under another [3]. Understanding which exclusions were active is often enough to explain an alarming number without changing a single sentence.
How Can I See My Real Turnitin Similarity and AI Report Before I Submit?
Similarity and AI detection are two separate signals produced by the same system, and students routinely conflate them. The similarity report answers "how much of this text matches existing sources," while the AI writing indicator estimates how much of the document appears machine-generated; a paper can be 0% similar and still carry an AI flag, or 40% similar with no AI signal at all [4]. Checking only one of the two leaves half the risk unmeasured.
The practical route is a non-repository draft check. A non-repository check runs your file through the same detection logic without adding it to the student paper database, so the act of checking does not itself create a future self-match, and the reports are not shared with third-party databases [4]. That distinction matters most for students who plan to check several drafts of the same assignment.
Reading the returned report properly means opening both PDFs: the similarity report for source-by-source matches, and the AI report for the flagged passages and confidence band. Turnitin displays an asterisk (*%) rather than an exact figure when AI detection falls below its 20% confidence threshold, so a "*%" result is a low-confidence signal, not a clean bill of health [4]. Treat the two documents as a pair, because that is exactly how your instructor will read them.
If you would rather see those two PDFs before your deadline than after it, turnitin0 runs the same check on your draft and returns both reports in one order, usually in under 15 minutes.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Is a high similarity score the same as plagiarism?
No. The similarity score only reports the percentage of text that matches indexed sources, and correctly quoted, cited or bibliographically standard material counts toward it [1]. Instructors are expected to review the matched sources before drawing any conclusion [2].
Why did my score change between two checks of the same file?
Because the displayed percentage depends on exclusion settings such as quotes, bibliography and small-match thresholds, which are configured by the instructor or administrator [2][3]. A different configuration produces a different number from identical text.
Does checking my draft add it to the Turnitin student paper database?
Not if the check is run as non-repository. A non-repository check processes the file without storing it in the student paper database, which avoids creating a self-match on your later submission [4].
Can a paper be 0% similar and still be flagged for AI writing?
Yes. Similarity measures overlap with existing sources, while the AI writing indicator estimates how much of the text appears machine-generated; the two are independent signals [4].
What does an asterisk instead of a number mean on the AI report?
It means the AI detection result fell below Turnitin's 20% confidence threshold, so the system reports *% rather than an exact percentage [4]. Read it as a low-confidence signal rather than proof of anything.