A good Turnitin similarity score is generally below 15%, although the acceptable percentage depends on your university guidelines and the type of assignment you submit. A low similarity score usually shows that your work is original and properly referenced, while a higher score may include correctly cited quotations, references, or common academic phrases. This guide explains what similarity percentages mean, what happens between submission and result, how to interpret the similarity report itself, and how you can reduce unnecessary matches before submitting your final draft with confidence.
What is a good Turnitin score?
As a general rule, a score under 15% is usually comfortable, a score between 15% and 25% is worth reviewing, and a score above 25% deserves careful attention before you submit. These are guidelines, not rules, because Turnitin does not set a pass mark and neither do most universities in any public way. Your department may have its own expectation, and if it does, that always beats a general figure you read online.
It also depends on what you are writing. A literature review that quotes and cites heavily will naturally score higher than a reflective essay written entirely in your own voice. A scientific report with a standard methods section will match other reports that describe the same standard procedure. None of that is plagiarism. So before you judge your score against a band, ask what kind of assignment it is and how much matching you would expect from honest work.
Is a 20% similarity score too high?
Not on its own. A 20% score can be perfectly healthy or a genuine problem, and the only way to tell is to open the breakdown. If those twenty points are spread across many sources in small slices, with your reference list and a few quotes making up a chunk of it, you are almost certainly fine. If a single source is responsible for most of the twenty per cent, that is the signal to look closer, because it suggests one passage leaned too heavily on one place.
This is why chasing a lower headline number is the wrong instinct. A student who panics at 20% and starts deleting citations to drop the figure is making their work worse, not better, and creating real plagiarism in the process. Read the similarity report, find the largest single match, and judge that. The total takes care of itself once the individual matches are clean.
Check your Turnitin score for UK universities US Universities, Australian universities, European universities, and Canadian universities.
What actually happens when your work is submitted to Turnitin?
When student submissions go through the system, Turnitin does not read your work for meaning. It compares your text against a database of billions of web pages, journals, and previously submitted student papers, looking for matching text, string by string. Every sentence, and often every few words within a sentence, gets checked against that database.
Once the check finishes, your paper receives a similarity score, shown as a percentage from 0 to 100. A 0% means no matching text was found anywhere. A paper that receives a 100 similarity score, meaning the entire document matches existing sources word for word, almost never happens from a genuine piece of writing. In practice, a 100 similarity score usually points to a file upload error, an empty or corrupted document, or a paper submitted with no original content at all, rather than a real result you would ever see from honest work.
The score itself is colour coded, so you can read it at a glance before you open the detail:
- Blue: no matching text found
- Green: low similarity, typically under 24%
- Yellow: matches in the 25% to 49% range
- Orange: 50% to 74%, considered high similarity
- Red: 75% and above
Colour is a starting signal, not a verdict. A yellow band from a similarity score of 25 sitting entirely inside a properly quoted and cited reference list is nothing to worry about. The same 25% concentrated in one unquoted paragraph is a real problem. This is exactly why you open the report rather than stopping at the colour, and why Turnitin similarity checking was designed to show sources, not just a number.
Why the percentage alone can mislead you
Turnitin matches text, and it cannot tell why two passages are similar. Your bibliography matches because reference lists repeat the same titles and authors that exist everywhere. Quotations match because they are, by design, someone else’s exact words. Common phrases in your subject match because everyone in the field writes them. A standard method description matches because the method is standard. All of these push your similarity scores up without any wrongdoing.
The result is that a clean, well referenced piece of work can score higher than a lazy one that cites nothing. The percentage rewards the wrong behaviour if you read it naively. That is the whole reason your tutor reads the report rather than just glancing at the number, and it is why you should do the same. To understand how the colours map onto these bands, see Turnitin report colours explained.
How do you interpret the similarity report?
The similarity report breaks the total score down source by source, and this is where the real reading happens. For each match, it shows you which source it came from, how much of your document that single source accounts for, and lets you click through to see your text sitting next to the original side by side.
To interpret the similarity report properly, work through it in this order:
- Sort matches by size, largest first.
- Open the largest match and check what it is: a citation, a quote, or unoriginal writing.
- Note whether that one source is doing most of the work, or whether the total is spread thinly across many small matches.
- Repeat for the next few largest matches until you’ve accounted for most of the score.
- Ignore matches under 1% individually. They rarely change the picture.
A report where twenty sources each contribute one or two per cent reads very differently from a report where one source alone explains eighteen of your twenty per cent, even though the headline similarity scores could be identical.
What raises your score without being plagiarism?
- Your reference list and in text citations, which repeat text found in many other documents.
- Direct quotations, even when you have marked and cited them correctly.
- Common phrases and set expressions that are normal in your discipline.
- Standard method or materials sections that describe widely used procedures.
- Tables, headings and template wording that your course provided.
Where your institution allows it, you can often exclude quotes and the bibliography from the calculation, which strips out a lot of this harmless matching and shows you a cleaner picture of your own writing. Whether that option is available depends on the settings your tutor uses.
How do I bring a high score down the right way?
- Open the source breakdown and sort the matches by size.
- Start with the largest match and decide whether it is a quote, a citation or text to rewrite.
- Rewrite any passage that stayed too close to its source, fully in your own words, and keep the citation.
- Turn long copied stretches into proper short quotations, or cut them.
- Exclude quotes and references from the score if your settings allow.
- Run the check again to confirm the score has dropped for the right reasons, before you treat the document as your final draft.
The full method, including what not to do, is in how to lower your Turnitin score. The goal is never to trick the tool. It is to make sure the work is genuinely yours, which brings the number down honestly.
Need plagiarism checks at scale? Get bulk Turnitin reports for your business in the UK, Australia, Europe and Canada.
What is student collusion, and how does Turnitin catch it?
Collusion is a separate problem from source plagiarism, and Turnitin flags it differently. Source plagiarism means your work matches a book, a website, or a published paper. Student collusion means your work matches another student’s submission on the same assignment. Turnitin similarity checks include a comparison against its student paper repository specifically for this reason, so copying a classmate’s work is checked for exactly as rigorously as copying a published article.
A simple example makes this concrete. Say Jane, who originally wrote a full assignment herself, shares her finished file with a coursemate so he can “check the structure.” Eric submits Jane’s paper with only light edits, changing a few words here and there, as his own final draft. Both papers go through Turnitin. Eric’s report comes back with a very high similarity score, and the top match is not a website or a journal. It’s Jane’s submission.
Collusion is typically identified when a student’s work matches another student’s submission on the same assignment, especially when the overlap is large, concentrated in one source, and involves two students on the same course or cohort. Small overlaps between classmates can happen innocently, particularly on templated assignments everyone was given, such as a shared lab report structure or an identical case study brief. What separates that from collusion is scale and pattern: near-identical paragraphs, matching errors, or whole sections lifted wholesale, rather than a shared heading or a common reference list.
This is also why checking your own work before the deadline matters more than most students realise. A private check on your draft shows you your similarity scores in isolation. Once your coursemate’s paper and yours both sit in Turnitin’s repository, any overlap between the two of you becomes visible to both markers, permanently, the moment either paper is submitted. Running your own check earlier catches accidental overlap from shared notes or group study sessions before it ever reaches that stage, and confirms your final draft is genuinely your own before you submit it for real.
What score do universities actually accept?
Most universities do not publish a fixed threshold, and the ones that mention a figure usually frame it as a trigger for review rather than an automatic fail. Some courses flag anything above 15%, others are relaxed up to 25% as long as the matches are clean. The safest move is to find your department guidance and follow it, then read your own report carefully regardless of what the number says.
If you want to see your real result before you submit, you can run a private check with our plagiarism checker while you draft, then confirm with the exact Turnitin report for $5 once your work is final.
Does a good score change by subject or assignment type?
Yes, and this is one of the most useful things to understand, because it stops you comparing your score against the wrong benchmark. Different kinds of work produce different natural levels of matching, all of it honest, so a score that would be high for one assignment is normal for another.
A reflective essay or an opinion piece written almost entirely in your own voice will usually score very low, because there is little to match beyond the odd common phrase. A literature review sits at the opposite end, because its whole job is to engage with many sources, quote some of them and cite all of them, so it naturally pulls in more matches. A scientific report with a standard methods section will match other reports describing the same procedure, since there are only so many ways to write a standard protocol accurately. A law essay quoting statutes and cases will match because the wording of a statute is fixed and everyone quotes it the same way.
So when you read your score, ask what you would expect from honest work of this type before you judge it. A 22% literature review with clean, spread out matches is in good shape. A 22% reflective essay would be worth a closer look, because that genre should produce less overlap. The band is the same, the meaning is different, and the assignment type is the missing piece of context. If your department has given subject specific guidance, that always beats a general rule, and for the mechanics behind the matching that drives all of this, see how Turnitin works.
Frequently asked questions
Is 10% a good Turnitin score?
Usually yes. A 10% score with small, scattered matches is a healthy result for most assignments. Still open the breakdown to confirm no single source is doing most of the work.
Does a 0% score mean my work is perfect?
Usually yes. A 10% score with small, scattered matches is a healthy result for most assignments. Still open the breakdown to confirm no single source is doing most of the work.
Does a 0% score mean my work is perfect?
No. A very low score can simply mean your topic has little matching text online, or that quotes and references were excluded. It shows low overlap, not quality, and it cannot catch an uncredited idea you reworded.
Can I check my Turnitin score before I submit?
Often you cannot, because Turnitin sits behind your university login. Some tutors open a draft folder for self checking. Otherwise, run a similarity check on your own first so you can fix issues before the real submission.
Is a high score on a literature review normal?
Often, yes. A literature review engages closely with many sources, quotes some of them and cites all of them, so it naturally produces more matching text than an essay written in your own voice. As long as the matches are spread across many cited sources rather than concentrated in one, a higher score on a review is usually nothing to worry about.
Does a higher score always mean a worse mark?
No. The score is not a grade and does not feed into one. A well referenced piece can score higher than a thin one, so your tutor reads the matches in context rather than treating the percentage as a mark in itself.
What is the difference between a high similarity score and student collusion?
A high similarity score simply means a lot of matching text was found somewhere, which could be citations, quotes, or common phrasing from any source. Student collusion is a specific type of match, one where the source Turnitin flags is another student’s submission on the same assignment rather than a published source. The score can look identical on the surface; only the source breakdown tells you which situation you’re looking at.
Want to see your similarity result before you submit? Run a plag check, and review every match in plain English.
