Direct answer
Turnitin does not define any "allowed" plagiarism percentage. It reports a similarity score — the share of a document's text that matches sources in its databases — and leaves the judgement to your institution's academic integrity policy [1]. A 12% score can be a problem if the matched text is uncited copying, while a 30% score can be entirely legitimate if it is quotations, references, and standard terminology. There is no universal pass mark, and no number that guarantees safety.
Does Turnitin Have an Official Plagiarism Percentage Limit?
No. Turnitin is a text-matching and reporting tool, not a rule-setting body, and it publishes no threshold above which a submission counts as plagiarised [2]. The similarity score is explicitly described as a starting point for interpretation: it tells a reader how much of the text overlaps with sources in the comparison databases, not whether that overlap is misconduct [2]. This distinction matters because students often search for a magic number and then treat it as a pass/fail gate, which the tool was never designed to provide.
The score is also structurally incapable of being a plagiarism verdict. Turnitin matches strings of words, so it cannot tell whether a matched sentence was quoted with a citation, paraphrased poorly, or copied outright [2]. A bibliography, a block of direct quotations, a list of statutory definitions, or a common phrase like "the results of this study suggest" all count towards the percentage. Two documents with identical scores can therefore have completely different integrity profiles.
What Turnitin does provide is context: the match overview, the list of matched sources, and the highlighted passages in the document. Reviewing those elements is what turns a number into a meaningful judgement [2]. If you want to know whether your own draft is at risk, the useful question is not "what is my percentage" but "which sources are matching, and are they cited".
What Similarity Percentage Is Considered Acceptable At Universities?
There is no single acceptable band, because thresholds are set locally by each university, department, or individual instructor [3]. Some institutions publish guidance such as "matches above 20% are reviewed", others flag any submission where a single source accounts for more than a set share of the text, and many publish nothing at all and rely on academic judgement. Turnitin's own guidance to instructors recommends reading the report rather than applying a fixed cut-off, precisely because a number alone cannot distinguish citation from copying [3].
Discipline and genre shift the picture considerably. A literature review, a legal case analysis, or a lab report with a standard methods section will naturally produce higher similarity than an original reflective essay, because the conventions of those forms require reproducing standard wording and heavy citation [3]. Instructors are advised to weigh where the matches sit in the document — reference list, methodology, or the argument itself — and whether the matched material is attributed.
The practical implication is that you should find your own programme's rule rather than borrowing a number from a forum. Check your module handbook, the academic integrity policy on your institution's website, or ask your tutor directly what they do with similarity reports [3]. If your department uses a review band, that band is the only "allowed percentage" that applies to you, and it can change between modules and between years.
How Can I Check My Own Similarity Score Before Submitting?
The most reliable way to reduce uncertainty is to see the report before your instructor does. Reviewing matches ahead of submission lets you paraphrase or properly cite the passages that are driving the score, and Turnitin's own guidance on using the report for improvement recommends working through the match overview from the largest blocks downwards [4]. Fixing a single 9% block from one source is usually more valuable than chasing a dozen 1% matches from common phrasing.
When you re-check, keep the conditions comparable. A pre-submission check is most useful when it uses the same comparison databases and the same file format your institution will use, because the score is a function of what the document is compared against [4]. Checking a trimmed or reformatted version of the file can produce a different result from the one your tutor will see, which defeats the purpose of the exercise.
It also helps to read the report the way a marker would. Open the match list, look at each highlighted passage, and ask whether a reader could tell that the material came from somewhere else [4]. Uncited copying in your own argument is the real risk; correctly attributed quotation and a tidy reference list are not. Once you can answer that question passage by passage, the overall percentage stops being a source of anxiety and becomes a diagnostic.
If you would rather see the actual report than guess at it, turnitin0 gives you that view before the deadline does — a Turnitin AI detection report and a similarity report delivered together, showing the same score, match overview, and flagged passages your instructor will open, typically in under 15 minutes and without adding your file to the student paper database.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
Get [Real Turnitin](https://www.turnitin0.com/) AI & Similarity Report
FAQ
Is 20% similarity bad in Turnitin?
Not inherently. Twenty percent means a fifth of the text matched sources in the databases, which is normal in a cited literature review and concerning in a short original essay [1]. What decides it is whether the matched passages are attributed and whether they sit in your argument or in your references.
Does Turnitin flag plagiarism automatically?
No. Turnitin produces a similarity report and, separately, an AI writing indicator; neither issues a misconduct finding [2]. A human — usually your instructor or an academic integrity officer — decides whether a match represents a problem.
Can a high similarity score be legitimate?
Yes. Direct quotations with citations, reference lists, standard methodology wording, and discipline-specific terminology all raise the score without any integrity issue [3]. This is why institutions review the match list rather than the headline number.
What happens if my similarity score is above my university's limit?
Typically the submission is reviewed rather than automatically penalised. The reviewer examines which sources matched and whether attribution was present [3]. If the matches are properly cited, many policies treat the case as a formatting or paraphrasing issue rather than misconduct.
Should I check my similarity score before submitting?
It is a reasonable precaution, provided you act on what you find by paraphrasing or adding citations [4]. Note that a pre-submission check reflects the databases it was run against, so treat it as a diagnostic rather than a guarantee of the final result.