Direct answer
There is no universal "allowed" plagiarism percentage. Turnitin itself publishes no pass/fail cutoff, and acceptable similarity ranges are set locally by each university, department, or instructor — many treat roughly 10–20% as a discussion point rather than a rule, while others publish no number at all [1]. What matters far more than the total figure is which text matched and whether it was correctly cited [1].
Is There a Universal Plagiarism Percentage That Is Allowed?
No. The single most common misconception about similarity scores is that somewhere a governing body has published a number — 10%, 15%, 20% — above which a submission fails. No such number exists, and Turnitin explicitly frames its similarity score as a signal for a human to interpret rather than a verdict on academic integrity [2]. Because the score is produced by pattern matching rather than judgement, it cannot know whether a student cited properly, so the platform leaves the decision to the institution [2].
What does exist is a wide spread of local policy. Some departments publish a soft ceiling in a student handbook; others deliberately refuse to publish one to prevent students from treating it as a target to write up to. A few institutions apply different expectations by assessment type — a literature review that quotes heavily may be reviewed at a higher percentage than a reflective portfolio [2]. The practical consequence is that the only authoritative answer to "what percent is allowed here" is your own module handbook or your instructor's guidance.
A further complication: the number is not stable. Turnitin lets instructors exclude bibliographies, quoted material, and small matches, and lets them filter out specific sources, so the same essay can legitimately display several different percentages depending on the settings applied [2]. Chasing a specific figure is therefore less useful than understanding what your institution does with the report.
Why Can a High Similarity Score Still Not Be Plagiarism?
Because similarity measures matched text, not misconduct. A report highlights every passage that overlaps with a source in the database, and that includes correctly quoted and properly referenced material, standard terminology, common phrasing, and reference lists [3]. A paper with an extensive block quotation and a long bibliography can easily show a percentage that looks alarming while containing no integrity problem at all.
The report is built to be read match by match, not as a single headline number. Each highlighted passage is linked to the source it matched, so a reader can see exactly what overlapped and judge whether attribution was correct [3]. This is the step students skip and instructors never do: the instructor opens the matches, checks the citation, and forms a view. Two essays with an identical 22% can produce completely different outcomes if one's matches are all cited quotations and the other's are unattributed paragraphs.
The inverse is equally important. A low score is not evidence of originality — a paper assembled from sources not indexed by the database, or paraphrased just far enough to break the string match, can score near zero while still failing academic-integrity standards [3]. A score is a diagnostic prompt, not a certificate.
How Can I Check My Own Similarity Score Before I Submit?
Run the draft through a check yourself, while you still have time to act on what it shows. Previewing a draft before the deadline turns the report from a post-mortem into an editing tool: you see the matched passages, fix the attribution gaps, and resubmit with confidence instead of discovering the problem after your instructor has already graded it [4]. Students who check early consistently spend that time on citation repair rather than on damage control.
Choose a check that does not add your paper to a student paper database. Non-repository checking means your draft is compared against existing sources without becoming a source itself, so a later submission does not flag as matching your own earlier upload [4]. It is also worth remembering that the point of the exercise is not to push the number down — it is to make sure every borrowed idea is attributed. Rewriting a cited quotation to dodge a match is the wrong fix; adding the missing citation is the right one [4].
Finally, read the report the way your instructor will. Open the matches, look at what each one actually is, and ask a simple question of every highlight: did I credit this? If the answer is yes, the percentage is not your problem. If it is no, you have found something worth fixing — and you found it before submission rather than after.
Reading the report match by match is exactly what your instructor will do — so it helps to see the same thing they will see. turnitin0 lets you upload your draft and receive a Turnitin AI detection report and a similarity report together, showing the score, the flagged passages, and the source of every match, so you can judge your own percentage against real evidence instead of a rumour about what's "allowed."
※ 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% plagiarism bad?
Not automatically. Similarity percentages count quoted, cited, and bibliographic material, so 20% can be entirely legitimate [3]. Whether it is a problem depends on what matched and whether it was attributed — the same figure means different things in different papers.
What is an acceptable similarity score for a university essay?
There is no standard figure. Institutions set their own ranges, and many publish none at all, precisely so students do not write to a target [2]. Check your module handbook first, then your instructor.
Does Turnitin fail you for a high percentage?
No. Turnitin produces a report; your institution makes the decision [1]. The score is a starting point for a human review of the matched passages, not an automatic penalty.
Can I check my similarity score before submitting?
Yes, and doing so is the most useful step you can take. A pre-submission, non-repository check shows your matches while there is still time to correct attribution [4].
Why did my score change between two checks?
Because settings and databases differ. Excluding quotes, bibliography, or small matches — or checking against a different index — will produce a different number for the same text [2]. Compare reports only when the settings behind them are comparable.