Your report says 27%, the sidebar is full of colored numbers, and your supervisor asks what source 4 is. Most students read a Turnitin similarity report by its big number and stop. The number alone says little: a 14% report can contain a copied paragraph, and a 35% report can consist entirely of correctly referenced quotations and a bibliography that nobody excluded.
This article is part of our guide How to Write a Master’s Thesis by the Mastermind PhD team and relates to our Paraphrasing & Plagiarism Check service. Sources last checked: November 2026.
This guide explains each part of the Turnitin Similarity Report as Turnitin documents it: the score and its colors, the Match Overview and All Sources panels, the four match groups of the new report, the filters and exclusions that change the figure, and the flags that expose manipulated text. It ends with a worked example of a fictional report and the action each type of match calls for.
What is a Turnitin Similarity Report? It is a comparison of a submitted file against Turnitin’s databases of current and archived web pages, previously submitted student papers, and periodicals, journals and publications. Every string of text that matches a source is highlighted, numbered and linked to that source, and the similarity score gives the percentage of the submission’s text that matches. Turnitin states that it does not check for plagiarism; it highlights similarity for a human reader to judge.
Step 1: Read the score and its color correctly
In the assignment inbox the report icon is colored by the overall percentage. For Feedback Studio and Originality Check, Turnitin lists blue for no matching text, green for one word to 24%, yellow for 25% to 49%, orange for 50% to 74%, and red for 75% and above. In the Similarity and SimCheck products green means 0% and blue means 1% to 24%, and Turnitin warns that color schemes can differ between integrations.
None of these bands is a verdict: Turnitin describes the score as “simply the percentage of text in a submission that matches other sources” and says it is natural for a submission to match some of the database. Treat the color as a sign of how much reading lies ahead, not as a grade.
Step 2: Open the Match Overview
Clicking the red percentage in the similarity toolbar opens the Match Overview: a list of sources ordered from the highest to the lowest percentage of your paper that each one matches. Each source has a number and a color that correspond to the highlights in the text. The arrow beside a percentage opens the Match Breakdown, which lists every source identified for that match, including duplicates.
Clicking a source places a box over the highlighted text showing where in the source the match sits; the book icon opens the full source text in a side panel with matching passages in red. This panel answers the question “where does the 27% come from?”: in most theses the first three sources account for most of it.
📱 Check your similarity with a Turnitin report — Every file comes with a Turnitin report and an AI check; similar passages are reworked without changing the meaning.
Step 3: Switch to All Sources when matches overlap
The All Sources panel lists every source found, including overlapping sources that match the same text, sorted by percentage. Selecting one shows how many times that single source matches in the submission and lets you step through each instance. The practical difference: Match Overview shows the best source for each passage and gives the big picture; All Sources shows whether the same sentence exists in ten places on the web, which usually means a common definition or standard wording rather than copying from one author.
Step 4: Use the four match groups in the new report
In the enhanced Similarity Report, the Match Groups tab divides the overall percentage into four categories based on citation and quotation marks:
| Match group | Turnitin’s criterion | What it means for you |
|---|---|---|
| Not Cited or Quoted | Text matches that are not written as a quotation and have no citation to the original source | The category that needs action: rewrite from understanding and cite, or quote and cite |
| Missing Quotations | The text is cited, but the match is so exact that it may also require quotation marks | Add quotation marks or paraphrase properly; the citation alone is not enough |
| Missing Citation | The text is written as a quote but lacks a citation | Add the citation and the page number |
| Cited and Quoted | The text is quoted and cited to a source | Check formatting only |
Each highlight carries a source number that appears on its source card, which shows the top source, the percentage matched, the number of matching text blocks and their total word count. The Sources tab lists the top source for each highlight, and “Show overlapping sources” reveals every source matched to the same words; where several sources match identical text, Turnitin ranks internet sources first, then publications, then student papers.
Step 5: Check which filters and exclusions were applied
Filters decide what counts. Turnitin’s exclusion guide describes them as follows:
| Filter | What it does | Note |
|---|---|---|
| Exclude bibliography | Detects and ignores the reference list; for non-English texts it looks for beginning phrases such as “references” and terminating phrases such as “appendix” | If references still show, Turnitin suggests resubmitting later when the advanced algorithm is available |
| Exclude quoted text (Exclude Quotes in the classic report) | Ignores text inside recognized quotation marks, including “…”, ‘…’, «…» and „…”; indented block quotes in .doc and .docx files are also excluded | Quotation marks must be real characters; hidden or fake marks are flagged |
| Exclude cited text (new report only) | Removes properly cited sentences together with their in-text citations | Falls back to header-based exclusion if detection is unavailable |
| Exclude small matches (new report) | Hides matches shorter than a set number of words; the default and minimum is 8 words | Removes noise from generic phrases |
| Exclude sources that are less than (classic report) | Drops whole sources below a word count or percentage; for a 5,000-word file set to 10%, sources matching fewer than 500 words disappear | Works at source level, not match level |
| Exclude a specific source | Removes one source and all its matches, for example your own earlier draft | An instructor action; a resubmitted draft is excluded automatically |
| Exclude assignment template | Ignores template text supplied by the department | Set at assignment level |
Turnitin cautions that applying filters “may or may not change the overall similarity percentage”, because one passage can match several overlapping sources. Two reports with different filter settings are therefore not comparable, and a self-check with generous filters can mislead you about the official check. Students can use exclusions in their own view only if the assignment settings allow it.
Step 6: Look at the flags
Turnitin examines any document of at least 500 characters for text manipulation and shows a red flag icon with a count in the Flags for Review panel. Two patterns are documented: replaced characters, where letters from other alphabets that look like Latin letters are inserted to break matches (Turnitin swaps them back before checking and flags the file), and hidden text, such as white or tiny characters or hidden quotation marks placed to trick the quote filter.
Turnitin says a flag “is not necessarily an indicator of a problem”, but it tells the instructor where to look. For a student the lesson is simple: the tricks sold online as ways of “cheating Turnitin” are detected and displayed in their own panel.
Step 7: Understand “Submitted to” sources
A source labeled with an institution name is a student paper in Turnitin’s repository. Its full text is not shown; viewing it requires a paper view request to the institution that holds it. If it is your own earlier draft, ask your instructor to exclude that submission. If it is another student’s work, read the passage carefully: a match with an unseen paper is exactly the kind of evidence examiners follow up.
Worked example: read a Turnitin similarity report from a fictional thesis
The following breakdown is invented for this guide and is purely illustrative. A 15,000-word master’s thesis returns 27% with no filters applied.
| Match type | Share of text | What the report shows | Action |
|---|---|---|---|
| Reference list | 9% | 140 entries matching other papers’ bibliographies | None; ask for the bibliography filter |
| Cited and quoted definitions | 5% | Six definitions in quotation marks with author, year and page | Check formatting only |
| Instrument and method wording | 4% | Items of a published questionnaire; standard statistical phrases | Keep, but cite the instrument and add quotation marks where wording is verbatim |
| One source, one paragraph | 4% | 600 words in the literature review matching a single article, cited once at the end | Rewrite from your reading, cite each borrowed idea, quote anything kept verbatim |
| Small scattered matches | 3% | Phrases of 4 to 7 words across all chapters | None; the small-match filter removes them |
| Your own earlier draft | 2% | “Submitted to” your university | Ask the instructor to exclude the earlier submission |
After the bibliography, quote and small-match filters, the same thesis reads about 10%, and only one of the six categories needed real work. The 600-word paragraph would have been a problem in a 9% report as much as in a 27% one.
What to fix and what to leave
Once you read a Turnitin similarity report this way, the decisions are short. Fix: uncited copying, close paraphrase cited only at the end of a paragraph, verbatim sentences without quotation marks, and any match in your own voice that you cannot explain. Leave: correctly quoted and cited text, reference lists, technical terms, instrument names and standard statistical wording, and matches shorter than a sentence.
Never fix a number by swapping synonyms or running the text through a spinner; Turnitin notes that spun text can score low while remaining unoriginal, and its AI detector treats text modified by AI paraphrasing tools as AI-generated. Rewriting from understanding, with citations, is the only method that survives a careful examiner.
📱 Ask on WhatsApp now — 5 years of experience across all disciplines and 14 academic services. Two-stage payment (80/20); revisions are free until supervisor approval, within 30% of the chapter.
Frequently asked questions
What does a 0% similarity score mean?
That no text matched Turnitin’s databases. For a thesis with quotations and a bibliography this is unusual, since quoted and referenced text normally produces matches. Check that the file was processed fully; a very low score can also come from spun or purchased text, which is why low scores are not proof of originality.
Why does my report show a match with “Submitted to” another university?
The source is a student paper in Turnitin’s repository, so its text is not displayed. If it is your own earlier draft, ask your instructor to exclude it; if it is someone else’s paper, review the highlighted passage carefully and make sure your wording and citation are your own.
Do filters always lower the percentage?
No. Turnitin states that applying filters and exclusions may or may not change the overall percentage, because the same text can match several overlapping sources; excluding one source may leave the match standing through another.
What is the difference between the similarity score and the AI writing indicator?
They are independent. The similarity score measures matching text against the database; the AI writing indicator estimates the share of prose sentences likely generated by AI, is shown only to instructors, and does not appear among the similarity highlights.
Can I apply filters myself before my supervisor sees the report?
Only if the assignment is set to let students view the report and use exclusions. The official check uses the instructor’s settings, so note which filters you applied in any self-check and do not rely on a figure produced with filters that will not apply on the day.
Sources
- Turnitin. Understanding the similarity score. https://guides.turnitin.com/hc/en-us/articles/23435833938701 (accessed October 2026)
- Turnitin. Navigating the Similarity Report. https://guides.turnitin.com/hc/en-us/articles/23438306524173 (accessed October 2026)
- Turnitin. Navigating the new, enhanced Similarity Report. https://guides.turnitin.com/hc/en-us/articles/27395528629517 (accessed October 2026)
- Turnitin. How exclusion filters refine the Similarity Report. https://guides.turnitin.com/hc/en-us/articles/23539146689549 (accessed October 2026)
- Turnitin. Managing filters and exclusions in the Similarity Report. https://guides.turnitin.com/hc/en-us/articles/23476785544973 (accessed October 2026)
- Turnitin. Flags in the Similarity Report. https://guides.turnitin.com/hc/en-us/articles/23351259442829 (accessed October 2026)
- Turnitin. Turnitin and plagiarism. https://guides.turnitin.com/hc/en-us/articles/34400565079053 (accessed October 2026)
Sources last checked: November 2026. This article is general educational guidance and does not replace your university regulations or your supervisor’s instructions. How we prepare our content.