There is no single percentage that Turnitin or universities officially label as “bad.” Turnitin itself states there is no set percentage that indicates an acceptable or unacceptable level of similarity. What matters is what’s behind the number. A 30% score built entirely from properly cited quotations and a bibliography is fine. A 15% score where every highlighted passage is lifted without attribution is a serious problem. That said, most students and instructors treat scores above 25% as a signal that something needs a closer look.
How Turnitin’s Color Coding Works
Turnitin assigns a color to every submission based on how much of the text matches content in its database of academic papers, websites, and student work. The exact color scheme can vary depending on how your school’s system is set up, but the most common version breaks down like this:
- Blue: No matching text (0%)
- Green: 1% to 24% matching text
- Yellow: 25% to 49% matching text
- Orange: 50% to 74% matching text
- Red: 75% to 100% matching text
Some integrations swap blue and green or use gray for low-range scores. The colors are a quick visual shorthand, not a verdict. Landing in the green range doesn’t automatically mean your paper is clean, and hitting yellow doesn’t automatically mean you plagiarized.
What Most Schools Consider Acceptable
Because Turnitin doesn’t set a universal cutoff, each university or even each instructor decides what triggers further review. In practice, most schools treat scores under 15% to 20% as generally unremarkable, assuming the matches are common phrases, properly cited material, or standard disciplinary terminology. Scores between 20% and 25% often land in a gray area where an instructor might open the full similarity report to check the details. Scores above 25% almost always prompt a manual review.
Some programs naturally run higher. Law, nursing, and other fields that require students to quote statutes, clinical guidelines, or standardized procedures tend to produce higher similarity scores even when students have done original work. Your instructor should understand this context, but you can help yourself by making sure the high-matching sections are clearly attributed.
Why Your Score Might Be High Without Plagiarism
Turnitin compares your text against everything in its database, and it flags matches regardless of whether you cited them. Several things can inflate your score without any wrongdoing:
- Direct quotations: Properly quoted and cited passages still count toward the similarity percentage. A paper heavy on primary source analysis will naturally score higher.
- Bibliography and reference list: Titles, author names, journal names, and URLs all match content in the database. A long reference list can add several percentage points.
- Common phrases and terminology: Phrases like “according to the findings” or field-specific terms like “standard deviation” appear in thousands of papers. Turnitin may flag them even though no one owns those phrases.
- Assignment prompts and headers: If you paste the assignment question into your document, or if your school uses a standard cover page template, that text may match submissions from other students in the same course.
- Your own previous submissions: Turnitin stores past submissions. If you reuse material from your own earlier paper, or if you submitted a draft to the same assignment before uploading your final version, the system will flag the overlap.
Many instructors use Turnitin’s filter settings to exclude bibliographies, quoted material, or matches below a certain word count. If your instructor hasn’t turned on these filters, your raw score will look higher than the “real” similarity in your original writing.
How Instructors Actually Review a Report
The overall percentage is just the starting point. When an instructor opens your similarity report, they see every matched passage highlighted and linked to its source. They’re looking at several things: whether the matched text is inside quotation marks with a citation, whether multiple matches come from a single source (suggesting you leaned too heavily on one article), and whether the matched passages represent your argument or just your evidence.
A paper with a 40% score might be perfectly fine if 30% comes from block quotes in a literary analysis and the remaining 10% is scattered common phrases. A paper with a 12% score could be a problem if one long paragraph is copied verbatim from a website with no citation. The percentage tells your instructor where to look, not what to conclude.
AI Detection Is a Separate Score
Turnitin now includes an AI writing detection indicator alongside the similarity score. These are completely independent measurements. The similarity score checks whether your text matches existing sources. The AI indicator estimates whether your text was generated by an AI tool. They appear in separate reports and don’t affect each other.
Turnitin’s own documentation acknowledges that AI detection scores between 0% and 20% have a higher rate of false positives, meaning the system sometimes flags human-written text as AI-generated in that range. Scores in that zone are displayed with an asterisk rather than a specific number, signaling that the result is less reliable. A high AI indicator is not the same thing as a high similarity score, and your school likely has separate policies for each.
How to Lower Your Score Before Submitting
If your school lets you view your similarity report before the final deadline, you can take steps to bring the number down without changing the substance of your work. Start by opening the full report and checking which passages are flagged. Remove any pasted assignment prompts or template text from your document. Make sure your quotations are formatted correctly with quotation marks and citations, so your instructor can immediately see they’re intentional. Paraphrase any passages where you borrowed too closely from a source, keeping the citation but rewriting the idea in your own words.
If your bibliography is inflating the score, don’t remove it. Your instructor can filter it out on their end, and a missing reference list is a much bigger problem than a slightly higher similarity percentage. Focus your revision energy on the body paragraphs where unattributed matches appear.
One important note: resubmitting to Turnitin after revisions can take up to 24 hours to generate a new report. If your deadline is tight, factor that delay into your timeline so you’re not stuck waiting for an updated score at midnight.