Quick answer: If you are wondering how to reduce Turnitin similarity legitimately, fix the causes rather than the number.

Quote and cite every passage you use word for word, replace long or frequent quotations with genuine paraphrases that you also cite, rewrite any passage that follows a source’s wording too closely, reuse your own earlier work only with permission and acknowledgement, and ask whether your instructor’s report excludes the reference list and quoted material. The similarity percentage measures how much of your text matches Turnitin’s databases; it is not a plagiarism verdict, and no single percentage is “acceptable” everywhere. Disguising matches with synonym swaps, invisible characters or rewriting tools is misconduct and does not solve the underlying problem.

Students usually look up how to reduce Turnitin similarity in one of two situations: before submission, when a draft report shows a number that looks alarming, or after submission, when a tutor has raised a concern. In both cases the right response is the same. Work out which matches are legitimate, which reflect a citation or paraphrasing problem, and which are simply noise, then fix the second group properly. This guide explains how to read the report, the main causes of high similarity, and a step-by-step method for bringing it down honestly.

What the similarity score actually measures

Turnitin compares your submission with its databases of web pages, published works and previously submitted student papers, and the Similarity Report shows the proportion of your text that matches something in those databases. Turnitin’s own guidance stresses that the report is not a determination of plagiarism: it highlights matching text so that a person can judge whether each match is properly attributed. A paper full of correctly quoted and cited material can score high, and a paper with a short uncited passage can score low.

That is why there is no universal “good” percentage. Expectations depend on the assignment, the discipline and your institution. A literature review that quotes key definitions will match more than a reflective essay, and a methods section describing standard procedures will match more than a discussion. The site’s explainer on what counts as a good Turnitin similarity score covers the typical ranges people quote and why they should be treated with caution.

How to read the report before you change anything

1. Look at the sources list, not just the total. One source accounting for a large share of matches points to a specific passage to fix; many sources each at 1% usually indicate common phrases and properly quoted material.
2. Open each highlighted passage. For each one, ask: is this a quotation with quotation marks and a citation, a paraphrase with a citation, a common phrase or technical term, a reference list entry, or a passage that follows a source too closely?
3. Check which filters were applied. Turnitin offers options to exclude quoted material, exclude the bibliography and exclude small matches below a threshold. Whether these are switched on is usually set by your instructor or institution, and your view may differ from theirs.
4. Note matches to your own earlier work. If you submitted a draft or an earlier assignment to the same system, it may match itself. Ask your instructor how they want this handled.

How to reduce Turnitin similarity: six legitimate fixes

1. Quote correctly, and quote less

Every word-for-word passage needs quotation marks, or block format for long quotations, plus a citation with a page or paragraph number. Correctly quoted text can still count towards the score if your instructor has not excluded quotations, so the second half of the fix is to quote only when the exact wording matters, such as a definition, a legal provision or a striking phrase you analyse. Everything else should be paraphrased or summarised.

2. Paraphrase properly

A real paraphrase restates an idea in your own words and your own sentence structure, and it is still cited. The reliable method is to read the passage, close the source, write the idea from memory in the way you would explain it to a classmate, then check it against the original for accuracy and accidental borrowing. Swapping synonyms while keeping the source’s structure produces patchwriting, which Turnitin often still matches and which most policies treat as plagiarism. The site’s guide on how to paraphrase academically shows the technique with examples.

3. Synthesise instead of summarising source by source

Literature reviews often score high because each paragraph follows one source closely. Reorganise around themes: state a point in your own words, then show which studies support or complicate it. This lowers matches naturally, because your sentences are about the relationship between sources rather than restatements of any one of them.

4. Rewrite close paraphrases of your own notes

Many unintentional matches come from note-taking. If you copied sentences into your notes and later wrote from those notes, source wording slips into your draft. Find these passages in the report and rewrite them from the original idea, not from your notes.

5. Handle self-reuse openly

Reusing substantial parts of your own earlier submitted work without permission is treated as self-plagiarism by many institutions. If you need to build on an earlier assignment or a published paper, ask your supervisor what is allowed, cite the earlier work and rewrite rather than paste.

6. Ask about exclusions, do not assume them

If most of your score comes from the reference list, quotations or standard phrases, explain this to your instructor or supervisor and ask whether the report is being viewed with those exclusions. Do not delete your reference list or remove quotation marks to change the number.

What not to do

Search results for this topic are full of tricks. They are worth naming because each one creates a bigger problem than the score:

• Synonym swapping and rewriting tools. Running text through a paraphrasing or “humanizer” tool to break matches is disguising plagiarism, and the result often reads oddly enough to attract attention.
• Character substitution and hidden text. Replacing letters with look-alike characters, inserting invisible characters or using white text are deliberate attempts to defeat the system and are treated as misconduct.
• Submitting images of text. Pasting passages as images to avoid matching is easily noticed and treated the same way.
• Removing citations or quotation marks. This may change the score but turns a properly attributed passage into an unattributed one.

Integrity panels look at what you did and why, and evidence of an attempt to manipulate the report is often more serious than the original matching text.

A worked example

The scenario below is illustrative. A master’s student’s draft literature review returns a similarity of 31%. The sources list shows one review article at 9%, the student’s own submitted proposal at 7%, and dozens of sources at 1% or less.

• The 9% source: three paragraphs summarise that review article closely, sentence by sentence. The student rewrites them as a synthesis, combining the review’s points with two other studies, and cites all three.
• The 7% self-match: the background section was pasted from the student’s own proposal. The student asks the supervisor, who confirms reuse is acceptable with acknowledgement; the student notes this and revises the section to reflect the final study.
• The 1% sources: these are reference list entries, standard phrases such as “semi-structured interviews were conducted” and short quotations of definitions. The student leaves the correctly quoted definitions, shortens two long quotations into cited paraphrases, and leaves the standard phrases, which are not a problem.

The new report is lower, but more importantly every remaining match is either properly attributed or a common phrase that the student can explain.

Why some chapters and disciplines match more

Similarity is not spread evenly through a thesis, and knowing where matches cluster helps you judge your report fairly.

• Methods sections reuse standard wording for procedures, instruments and statistical tests. Phrases such as “data were analysed using a one-way ANOVA” will match thousands of documents. That is expected; what matters is that you describe your own procedure accurately and cite the instruments you used.
• Literature reviews match more than other chapters because they discuss published work. The fix is synthesis, as described above, not the removal of necessary citations.
• Law, history and literature theses quote primary sources such as statutes, archival documents and literary texts. Those quotations are evidence, and the right approach is precise quotation and citation, not paraphrasing a statute into looser language.
• Theses by publication include papers you have already published, which will match the published versions. Follow your institution’s rules for including them and state their publication status clearly.
• Technical and scientific fields share fixed terminology, chemical names and equation descriptions that cannot be reworded without losing precision. Keep the correct terms.

If your report is dominated by any of these, a short note to your supervisor explaining the pattern is far more useful than trying to force the number down.

If your submitted work has been flagged

If a tutor has raised a concern about similarity after submission, do not try to resubmit a disguised version. Ask which passages concern them, review those passages against your sources and notes, and be ready to show your drafts and reading notes. If the matches are properly quoted and cited, explain this. If you find a genuine attribution error, say so and ask how your institution handles it; many treat honest first errors as an academic practice issue rather than misconduct. Contact your students’ union or advice service if a formal process begins. If the concern is about AI rather than matching text, the guide on reducing AI detection false positives covers what evidence to gather.

Habits that keep similarity low from the start

The best answer to how to reduce Turnitin similarity is to prevent unattributed matches in the first place.

• Take notes in your own words, and mark any copied wording in your notes with quotation marks and a page number.
• Record the full reference for every source as you read it.
• Draft from your understanding of the argument, not with a source open beside you.
• Run a draft through your institution’s checker only if it offers a draft submission option, and treat the report as a guide to passages to review.
• Keep version history and notes, so you can show how your text developed if a question is ever raised.

Why similarity panic costs more time than it saves

A high number on a draft report can trigger hours of anxious rewording, and that is exactly the wrong response. Rewording sentence by sentence rarely lowers matches much, it often makes the writing worse, and it leaves the real cause untouched: a draft built around sources rather than around your own argument. Students who start from a clear structure, with each section stating a point of their own and using sources as evidence, produce far fewer close matches in the first place, because their sentences are about their argument rather than about one author’s wording. That is the problem Tesify is built around: it helps you structure and organize your thesis around your own argument from the first outline, so writing from sources becomes synthesis rather than paraphrase.

For related guides, see how to rewrite a paragraph yourself and in-text citations in APA 7.

Where Tesify fits

Similarity problems usually start with structure: when your argument is clear, your sentences are yours. Start structuring your thesis in Tesify. It helps you structure and organize your thesis around your own argument, and 9,000+ students have written 15,000+ chapters with Tesify. Every word stays 100% written by you.

Frequently asked questions (FAQ)

What is an acceptable Turnitin similarity percentage?

There is no universal acceptable percentage. Turnitin reports matching text, not plagiarism, and expectations depend on the assignment, discipline and institution, so ask your instructor what they look for.

Does a high similarity score mean I plagiarised?

No. A high score can come from correctly quoted and cited material, reference list entries and common phrases. The report highlights matches for a person to evaluate; it does not decide whether plagiarism occurred.

Do quotations count towards the Turnitin score?

They can, unless the report is viewed with quotations excluded, which is usually set by your instructor or institution. Correctly quoted and cited text is not plagiarism even when it is counted.

Is it allowed to use a paraphrasing tool to lower my similarity?

Using a tool to reword sources so they no longer match is disguising plagiarism and is treated as misconduct by most universities. Paraphrase yourself from your understanding of the source and cite it.

Why does my reference list show as matching?

Reference entries match other papers that cite the same sources. Many instructors view reports with the bibliography excluded; if yours does not, explain that the matches are references.

Can my own earlier work raise my similarity?

Yes. If you submitted a draft or an earlier assignment to the same system, it may match. Ask your instructor how to handle it, and cite and rewrite any earlier work you reuse with permission.

How do I lower similarity in a literature review?

Organise paragraphs around themes rather than individual sources, paraphrase and synthesise rather than summarise one study at a time, and quote only when the exact wording matters.

Why does my methods section match so many sources?

Methods sections use standard wording for common procedures, instruments and statistical tests, so short phrases match many documents. That is expected. Describe your own procedure accurately, cite the instruments you used and explain the pattern to your supervisor if it raises the score.

What should I do if my submitted paper has high similarity?

Ask which passages concern your tutor, review them against your sources, show your drafts and notes, and acknowledge any genuine attribution errors. Do not resubmit a disguised version.