Direct answer
A plag report is the similarity report Turnitin generates after a paper is checked: it shows the percentage of your text that matches other sources, colour-codes every matched passage, and lists the source behind each match [1]. That percentage is not a plagiarism verdict — it is a measurement of textual overlap that a marker then interprets in context [1]. The report also carries a separate AI writing score, which is calculated independently of the similarity figure [4].
What Does a Plag Report Actually Show, and What Does the Similarity Percentage Mean?
A plag report has three layers, and students usually look only at the first. The headline number is the similarity index — the share of your document that overlaps with sources in Turnitin's comparison database [1]. Beneath it sits the match list: every source that contributed to that number, with the percentage each one accounts for [1]. The third layer is the document itself, where matched text is highlighted in the colour assigned to each source, so you can see exactly which sentence triggered which match [1].
The percentage is descriptive, not judicial. Turnitin's own guidance describes score bands qualitatively rather than as pass/fail thresholds, and states plainly that a low score does not prove originality and a high score does not prove misconduct [2]. A paper can score 0% because it is genuinely original, or because it was written in a language or style the database does not index. A paper can score 40% because it was copied, or because it is a literature review built from heavily quoted sources [2].
What changes the number matters as much as the number. Turnitin reports can be recalculated with quotes and bibliography excluded, and that single toggle frequently moves a score by ten or twenty points [3]. Markers typically read the report with those exclusions in mind, because a reference list is supposed to match other reference lists [3]. If you are quoting a report percentage to a friend or a tutor, state which view you are reading — the default "All Sources" figure or the filtered one.
Finally, the similarity score and the AI writing score are two different instruments measuring two different things [4]. The similarity report compares your words to existing text; the AI indicator estimates how likely the prose is to have been machine-generated [4]. A plag report with 8% similarity can still carry a high AI score, and a plag report with 30% similarity can carry none. Reading only one of the two numbers is the most common way students misjudge their own risk [2][4].
Why Is My Similarity Score High When I Did Not Copy Anyone — and Which Parts of the Report Actually Matter?
The most frequent cause of an alarming similarity score is not copying; it is scaffolding. Reference lists, in-text citations, direct quotations, module handbooks, and standard method descriptions all live in the same databases Turnitin searches, so they match by design [3]. Turnitin provides exclude-quotes and exclude-bibliography options precisely because these elements inflate the raw figure without indicating any misconduct [3]. Recalculating with those exclusions on is the first thing to do before you panic.
The second cause is legitimate reuse of your own or your cohort's material. If your institution allows resubmission, a draft you uploaded earlier can match your final version; if a previous student's paper on the same brief is in the repository, shared assignment wording can match too [3]. Turnitin also lets markers filter matches below a small word count, which removes incidental phrase collisions like "the results of this study suggest" [3]. A high score built from many tiny matches reads very differently from a high score built from one large block.
Which parts matter, then? Look at the largest single-source contribution first, then at the longest continuous highlighted block, then at whether the highlighted text is quoted and cited [1]. A 22% score composed of forty 1% matches against journal articles is usually citation noise; a 22% score composed of one 20% match against a single essay is a conversation with your tutor [1][2]. The distribution of matches tells the story, not the total [2].
Also check the settings printed on the report itself. A report generated with quotes and bibliography included, or without the small-match filter, will read higher than the same paper marked under a stricter filter [3]. If your tutor's report and your own report show different numbers for the same draft, compare the exclusion settings before comparing the scores — that discrepancy is almost always configuration, not a second submission [3].
How Can I See My Own Turnitin Similarity and AI Report Before I Submit?
Whether you can generate a plag report yourself depends entirely on your institution. Some universities enable a draft submission box or a self-check assignment that lets students run a paper through Turnitin and view the report before the graded deadline [4]. Others restrict report visibility to markers, or allow only one submission with no preview, which leaves students guessing at a number that will decide how their work is read [4]. Ask your module lead which configuration applies to you before the final week, not during it.
Where institutional preview is unavailable, students commonly use an independent pre-submission checking service that returns the same two documents a marker sees: a similarity report with match highlights and an AI writing report [4]. That is the practical value of a preview — not the number itself, but the ability to open the highlighted passages, notice an uncited paraphrase you genuinely missed, and fix the citation before it becomes a misconduct question [4]. Reviewing the report is a proofreading step, and it works best when there is still time to act on it.
Timing is the constraint that decides whether a preview helps or merely hurts. A report that arrives two hours before a deadline tells you your paraphrase on page four is uncited but gives you no room to rewrite it [4]. A report that arrives with a day or two in hand lets you re-cite, re-quote, or re-word the flagged passages and re-check the result [4]. Students who check early treat the plag report as a diagnostic; students who check late treat it as bad news.
Keep the two scores in view when you plan that check. If your similarity is clean but your AI score is high — a common outcome for AI-polished prose — the fix is rewriting the flagged passages in your own voice, not adding citations [4]. If your similarity is high but your AI score is low, the fix is citation and quotation hygiene [3]. Diagnosing which of the two is actually the problem is the whole point of seeing the report before submission [2][4].
Reading the report is one thing; getting your hands on the real one before the deadline is another. turnitin0 exists for exactly that gap — students upload a draft and receive the actual Turnitin AI report and similarity report, the same documents a marker opens, so the highlighted passages and both scores are visible while there is still time to edit rather than hope.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
Get [Real Turnitin](https://www.turnitin0.com/) AI & Similarity Report
FAQ
Is a 20% similarity score bad on a plag report?
Not on its own. A 20% score made of short matches against journal articles, quotations, and references is routine, while the same 20% concentrated in one uncited paragraph is worth a conversation with your tutor [2]. Read the match distribution before judging the total.
Does a plag report prove I used AI?
No. The similarity report measures overlap with existing text, and the AI writing score is generated by a separate model [4]. A plag report can show high similarity from references and still show no AI signal, or the reverse.
Why did my score change between two reports of the same draft?
Almost always because of report settings. Excluding quotes and bibliography, or applying a small-match filter, recalculates the percentage without altering a word of your paper [3]. Compare the settings printed on each report before assuming a second submission occurred.
Can I get a plag report before I submit my assignment?
Sometimes through your institution, if it enables draft or self-check submissions [4]. Otherwise students use a pre-submission checking service that returns the similarity and AI reports a marker would see, which is most useful when ordered early enough to act on the flags.
What should I fix first if my report is flagged?
Start with the longest continuous highlighted block that is not quoted and cited [1]. Fixing one uncited paraphrase usually removes more of the score than editing a dozen incidental phrase matches.