You upload your assignment. The screen loads. One number then appears to be more significant than the 3,000 words you spent days writing: 27 percent.

The first question you may ask is, “Is 27 percent plagiarism?” Your second question may be, “How do I keep bringing it down to less than 10 percent? Then the internet makes things worse. According to one forum, anything above 15 percent is bad. 25 % is considered “normal” according to a video. One of the classmates asserts that they were submitted at 32% and they got a distinction.

That is why Turnitin Similarity Scores are so misinterpreted. This might seem like a simple percentage, but the report is not simple. A score may include the use of material that is correctly quoted, referenced, common phrases, source wording, previous submission, and, if appropriate, genuine problematic copying. This number doesn’t tell you which of those situations it is.

When you can understand how Turnitin Similarity Scores really function, the report is not as terrifying and is more useful. Rather than worrying about a percentage, you can look at past matches that are relevant, work on your weaker academic writing, and submit with a better idea of what your university will be looking for.

The percentage is a signal to investigate, not a verdict to fear.

The 30 Second Answer

1. What Turnitin Similarity Scores Actually Measure

Turnitin Similarity Scores show how much text in a submitted document matches material in the sources available to the similarity checking system. The report highlights those matching passages and connects them with possible sources so that a reader can examine the context.

That distinction matters. Similarity is about textual overlap. Plagiarism is about using someone else’s words, ideas, or work without appropriate acknowledgement. Those concepts can overlap, but they are not interchangeable.

The University of Oxford’s Turnitin guidance makes the same point: Turnitin is a text matching tool, and a similarity report needs academic judgement. UCL likewise explains that the percentage should not be used on its own as proof that plagiarism occurred.

This is the foundation for understanding Turnitin Similarity Scores. The software can show you where text resembles other material. It cannot understand your intention, your assignment brief, your lecturer’s instructions, or whether a particular match is academically acceptable in your discipline.

2. Why 20 Percent Can Be Fine in One Paper and a Problem in Another

Suppose Student A receives 20 percent. Most of the matches come from a reference list, several properly marked quotations, a standard research method description, and short phrases that appear across many academic papers.

Student B also receives 20 percent, but one large block closely matches a journal article, and the wording has been reproduced without quotation marks. The percentages are identical. The academic risk is not.

This is why searching “Is 20 percent Turnitin similarity bad?” rarely gives a useful universal answer. Turnitin Similarity Scores have to be read at match level, not only at percentage level.

A better question is: what is creating the 20 percent? If the score is made up of small, understandable matches that are correctly acknowledged, the report tells a very different story from a score dominated by a few long, unexplained matches.

3. Is There a Safe Turnitin Percentage?

No universal Turnitin percentage is automatically safe. Universities, faculties, lecturers, and assessment types can apply different expectations. A dissertation literature review, a laboratory report, a legal problem answer, and a reflective essay do not use sources in exactly the same way.

Some institutions may give students local guidance about similarity, but that should never be confused with a global rule. Oxford’s guidance specifically warns against treating one number as a magic threshold. Scribbr’s similarity guidance also stresses that every meaningful similarity should be reviewed rather than assuming a low total guarantees originality.

The most useful approach is to check your module handbook, assessment instructions, university academic integrity policy, and any directions given by your lecturer. Those sources matter more than a percentage recommended on social media.

When reviewing Turnitin Similarity Scores, aim for accurate source use rather than an artificially low number. A low score achieved by removing necessary evidence, avoiding legitimate quotations, or rewriting clear technical language into awkward sentences is not a better academic result.

4. Can Low Turnitin Similarity Scores Still Be a Problem?

Yes. This is one of the most important facts students miss. A similarity score of 4 percent can look reassuring, but that 4 percent could still include a serious passage if the match represents copied material without acknowledgement.

Imagine a 2,500-word assignment in which one key paragraph is taken directly from an online source. The total percentage may remain relatively low because the rest of the document is original. The low overall number does not make the copied paragraph acceptable.

This is why Turnitin Similarity Scores should not be treated like a traffic light where green always means safe and red always means guilty. The highlighted wording, the source, and the way the material has been incorporated into the assignment matter much more.

UNSW’s academic integrity guidance gives a useful practical example: simply adding a citation does not make copied sentence wording acceptable if quotation marks are required. That is exactly the kind of issue that a percentage by itself cannot resolve.

5. Why High Turnitin Similarity Scores Are Not Automatically Plagiarism

A high result can happen for legitimate reasons. An assignment may contain quotations that are correctly identified, a long bibliography, standard terminology, a template supplied by the university, statutory wording, commonly used definitions, or repeated phrases from the assignment question.

Certain disciplines naturally produce more fixed language than others. A methods section may describe a standard procedure. A law assignment may quote legislation. A research paper may need exact names of established instruments or measures. Similarity can therefore reflect necessary academic material as well as poor source use.

UCL’s guidance on Turnitin encourages staff to look beyond the overall number because a report can contain many small matches or a smaller number of substantial matches. That is a much better way to think about Turnitin Similarity Scores than treating every highlighted phrase as suspicious.

The correct response to a high score is not “rewrite everything.” The correct response is “inspect the important matches and understand why they are there.”

6. What Actually Gets Matched?

Turnitin compares submitted text with material available through its comparison databases and the sources enabled for the assignment. Matches can come from internet material, academic publications, and previously submitted papers, depending on institutional access and assignment settings.

When Turnitin Similarity Scores flag an unfamiliar source, look at the wording first. Ask whether you recognise the idea, whether the sentence came from another source you used, and whether your citation points to the source you actually consulted.

Scribbr’s guide to similarity reports makes a similar recommendation: examine each relevant match and compare it with the source instead of assuming that every unfamiliar source proves copying.

7. Why References, Quotations, and Common Phrases Can Raise Turnitin Similarity Scores

A reference list is a list of information that is extremely standardized. Author names, article titles, journal names, publication year, and DOI information should all match the actual source. The details would be odd if they were changed just to change them.

It is also clear that similar quotations are found. The wording is deliberately the same if you are quoting an author properly. Academically, the question is whether the quotation is clearly cited, marked, is relevant to the text, and is used appropriately.

Common academic phrases can match as well. Statements like “The results of this study indicate,” or general descriptions of a commonly used method are likely to be repeated in numerous papers. Text that is not indicative of misconduct may, therefore, be reported in Turnitin Similarity Scores.

8. Why Paraphrasing Is Where Many Students Get Into Trouble

Weak paraphrasing is one of the biggest reasons a similarity report becomes concerning. Changing a few synonyms while keeping the original sentence structure is not strong paraphrasing. It may still be too close to the source, and it may also fail to demonstrate genuine understanding.

Good paraphrasing begins with meaning. Read the source, identify the idea you actually need, move away from the original wording, then explain the point in a structure that fits your own argument. You still cite the source because the idea came from someone else.

UNSW describes paraphrasing as putting a passage into your own words while preserving the original meaning and attributing it to the source. Purdue’s writing resources likewise emphasise that paraphrasing is more than swapping individual words.

If Turnitin Similarity Scores repeatedly highlight your paraphrases, the solution is not to use an online “humanizer” or replace words randomly. Return to the source, understand the idea, and write it again from your own understanding.

9. Why Turnitin Similarity Scores Can Change After Resubmission

Students sometimes upload a revised file and become alarmed because Turnitin Similarity Scores change. That can happen for several reasons. The document itself may have changed, exclusions may differ, the report may be processed under different settings, or a previous submission may now be available for comparison depending on how the institution manages repositories.

The important thing is not to assume that a changed percentage means something went wrong. Compare the actual matches. Did a new source appear? Did you add a quotation? Did you change your reference list? Did a long match disappear after improving a paraphrase?

Turnitin Similarity Scores are outputs from a comparison process. When the text or comparison conditions change, the result can change too.

10. Does Turnitin Similarity Detect AI Writing?

Not in the way many students assume. Turnitin Similarity Scores and AI writing indicators are different things. A similarity score reports matching text. An AI writing feature, where available to an institution, is a separate analysis.

Universities are also developing their own rules about generative AI. UNSW, for example, advises students to check course instructions and acknowledge AI use where required. The safest approach is to follow the specific policy for your course rather than relying on rumours about what software can or cannot detect.

For students trying to understand Turnitin Similarity Scores, keep the two questions separate: “Does this text match other sources?” and “Does my institution permit the way I created or edited this work?”

11. The Biggest Mistake: Writing for Turnitin Instead of Writing for Your Lecturer

Once students become obsessed with the percentage, their writing often gets worse. They replace precise academic terms with strange alternatives, remove useful quotations, weaken explanations, and spend hours trying to force the number down by one or two points.

Turnitin Similarity Scores can support that process by revealing areas worth checking. They should not become the purpose of the assignment.

The University of Reading’s academic integrity resources encourage students to develop skills such as paraphrasing, referencing, and avoiding unintentional plagiarism. Those skills improve the paper itself. Chasing a cosmetic score does not.

12. A Better Way to Read Turnitin Similarity Scores

Start with the largest matches rather than staring at the overall percentage. A single substantial match is usually more informative than twenty tiny matches to generic phrases.

Open the source and compare the wording. Is the passage a quotation? Is it cited? Is it a reference entry? Is it standard terminology? Is it a paraphrase that remains too close to the original? Is the match coming from an earlier version of your own work?

Then look for patterns. If multiple sections repeatedly follow source wording, you may need deeper rewriting. If most matches are bibliographic details and clearly marked quotations, the overall score may be less concerning than it initially appears.

Finally, review your university’s academic integrity guidance before making unnecessary changes. Turnitin Similarity Scores are most useful when they lead you back to good academic practice.

13. What Turnitin Similarity Scores Tell You to Fix Before Submission

Fix missing citations first. If an idea, fact, argument, or distinctive wording came from a source and your referencing style requires acknowledgement, make sure the source is credited correctly.

Next, check direct quotations. Exact wording should be clearly identified according to the required referencing style. A citation alone may not be enough when you have reproduced the original sentence.

Then review paraphrases that remain very close to source language. Rebuild those passages around your own understanding rather than changing isolated vocabulary. UNSW’s guidance on quoting and paraphrasing is especially useful for seeing the difference between borrowed wording and genuine paraphrase.

Finally, check whether the assignment itself has special rules about templates, appendices, declarations, group work, previous submissions, or AI use. Those instructions can affect how Turnitin Similarity Scores should be interpreted.

14. What Not to Do Just to Lower Turnitin Similarity Scores

Avoid deliberately misspelling words, using invisible characters, substituting letters with symbols, converting words to images, etc., to disrupt the ability of the text to match. Those actions will not strengthen academic integrity and can make a much deeper problem.

Don’t delete matches since they are citations. If the Turnitin Similarity Score is low but the paper does not cite the sources, it will not be considered worthy of submission.

Please do not rewrite each of the highlighted sentences in some rewriting service until they are unnatural. Your goal is to convey information in a clear manner through your own academic writing style, not to write something that is just a different version of everything in a database.

The best way to get Turnitin Similarity Scores to go down is to do the most academically beneficial: research broadly, take good notes, comprehend what you read, paraphrase sincerely, quote judiciously, and reference properly.

15. Turnitin Similarity Scores in Practice: The 28 Percent Report

Suppose Sara turns in a 3200-word literature-based assignment and gets a grade of 28%. She assumes that she has to rewrite almost one-third of the paper.

Once she accesses the report, the picture changes. Seven percent is due to “reference list” and “bibliographic wording. The remaining 6% is based on quotations that she has marked correctly and that the question had requested her to analyse. There are some small matches from the Assignment Title and the standard terms. The other 5 percent is from a journal article in which her paraphrase is too similar in wording and sentence order to the original.

The nifty trick isn’t to rewrite everything until the score is 10 percent. Sara must concentrate on that troublesome 5 per cent passage; check that the 5 per cent is the correct passage; verify that the quotations are needed and properly cited; and adhere to any local policy as to withholding references or quotations.

In real-world situations, this should not be the only metric to judge whether the assignment is good or bad; it is merely a resource that guides you toward where to search for similarities and differences.

16. Turnitin Similarity Scores Myths Students Still Believe

Myth one: anything above 20 percent is plagiarism. False. The source and pattern of matches matter.

Myth two: anything below 10 percent is automatically safe. False. A small score can still contain an important unattributed passage.

Myth three: 0 percent is the ideal academic score. Not necessarily. Correct quotations, references, and standard wording can create legitimate similarity.

Myth four: adding a citation makes copied wording acceptable. Not always. Exact wording generally needs to be presented as a quotation according to the required style.

Myth five: similarity and AI detection are the same report. They are different analyses.

Myth six: the goal is to beat Turnitin. The real goal is to submit work that is genuinely yours, accurately sourced, and compliant with your institution’s rules.

17. Five Questions to Ask Before You Submit

• First, can I explain where every major idea in this assignment came from? If not, review your notes and citations.
• Secondly, have I indicated specifically every passage in which words are not my own? Otherwise, review the rules for quotes in your style of referencing.
• Thirdly, do my paraphrases sound like my own explanation or do they still sound too much like the source sentence?
• Fourth, have I read the instructions for the assessments rather than assuming a percentage on the internet?

Fifth, if the Turnitin Similarity Score indicates a significant match, can I provide an explanation for this match, and does it require changes? If you can answer those 5 questions with conviction, you are thinking of the issues that matter.