To read a Turnitin report, don’t stop at the headline percentage. Go through the source list, the highlighted text and the filter settings at the end. Mostly the report comes to you as a PDF from the library or the research cell. At the top there is one number in a circle. Everybody looks at that number first, but it says very little. The useful parts are the source list, the highlighted text and one small line of filter settings at the end, which most scholars never read.
We go through the report from top to bottom, part by part. Then we look at the source list, the filters, and what the report cannot say about your work. At the end there is a worked example of one draft that came back with 23%.
The UGC similarity levels and the rules on what can be excluded are not repeated on this page. You will find those in our plagiarism limits guide, and finally your own university’s policy is what applies to you. Turnitin changes its interface now and then, so your labels may differ a little. Last reviewed September 2026.
What the similarity index measures
Turnitin checks your document against three collections. One is the web pages it has crawled. Another is published content from the publishers who take part. The third one is documents that institutions all over the world have sent to Turnitin. When a run of your words matches a run of words in any of these, Turnitin highlights it. The similarity index is how much of your document these highlights cover.
From this, a few things follow.
- It counts matching text, not plagiarism. Even a definition you quoted and cited correctly will still match. So will the name of a statistical test, a standard methods sentence or the title of a scale. Somebody has to sit and decide if a particular match is really a problem, and the software can’t do that for you.
- It only finds what is in its databases. Suppose some text was copied from a book that Turnitin doesn’t have, or it was translated from a Hindi or Tamil source. Then Turnitin may show nothing for it.
- It is not an AI score. With some Turnitin licences you also get an AI writing indicator. That is a different figure, calculated in a different way, and the similarity index does not include it.
Parts of a Turnitin similarity report
Mostly a Turnitin PDF report starts with one summary page. After that you get your document with the highlights in it. The last pages have the list of sources and the filter settings. The labels change a bit from version to version.
| Part of the report | What it means | What to do with it |
|---|---|---|
| Similarity index (the headline %) | How much of your document’s words match text in Turnitin’s databases, after applying the filters given at the end of the report | Write it down and keep it aside till you read the remaining parts |
| Internet sources % | The part matching web pages, and that includes Shodhganga, ResearchGate and document-sharing sites | Find the one or two biggest web sources first |
| Publications % | The part matching journal articles, books and conference papers | See if your own papers are there, and any article you paraphrased too closely |
| Student papers % | The part matching documents that any institution has submitted to Turnitin | See whether one of them is your own old draft |
| Primary sources list | Every matching source with a number, and the percentage of your document it covers | Start from the top and go down. Most of the problem is in the big sources only |
| Highlighted text | Your document, with every match coloured and numbered as per its source | Open each match coming from sources above about 1% |
| Filter line at the end | Whether quotes and bibliography were excluded, and what small-match threshold was used | Look at this line before you trust the headline figure |
Why the three category figures don’t add up
Say your index is 18%, and the report shows Internet sources 15%, Publications 9% and Student papers 11%. There is no mistake here. One sentence may match a web page, a journal article and a student paper together, and it gets counted under each category. In the headline figure, every word of your document is counted one time only.
Why the primary sources do add up
In the source list, every highlighted passage is given to one source, and usually that is the source with the biggest match. That is why the percentages in the list come to roughly the headline figure. Any other sources having the same text get hidden under the top one. So don’t panic over a source name you have never read. The same text may be in the paper you actually used, which Turnitin just ranked lower.
How to read the Turnitin source list
Start at the top of the list. Many times the first five to ten sources make up a big part of the similarity. The usual entries you will see are explained below.
“Submitted to [a university name]”
Somebody sent this document to Turnitin through that institution. You cannot open it, because the text shows only as highlights inside your document. Check two things. Is it your own institution, and is the match big and continuous? If yes, most probably it is something you submitted earlier. Many times it is the synopsis itself. Sometimes it is a chapter you checked last year, or some draft your guide had uploaded. Ask the research cell if they can exclude it. Whoever holds the Turnitin account is able to exclude a source. That person can also request to see a paper submitted at another institution.
Your own publications
Suppose one chapter comes from a paper you have already published. Then that paper will show as a large source. Some universities count it and some don’t, and our plagiarism limits guide has more on that. Counted or not, cite the paper in that chapter and say in the thesis that the chapter came from it.
Shodhganga and document-sharing sites
You will see Shodhganga matches very commonly. It has huge number of Indian theses, and many of them use the same standard phrasing when the topic is the same. When a match is to a document-sharing site like Scribd, mostly the same article or thesis was uploaded there also. Trace the original source and treat the match as a match to that original.
The long tail of sources under 1%
Dozens of tiny sources are usually just normal academic phrasing. Take “the data were analysed using”. Thousands of theses have that line, so it will match. Full test names and institution names match the same way. Mostly this is harmless. Worry only when lot of small matches sit together in one paragraph, because maybe that passage was joined from several sources.
Where the matches are
Don’t look only at the list. Scroll through the highlighted document also. In the literature review, some similarity is expected. The results chapter is different. If highlights come in the results or discussion or conclusions, take them seriously even when the total is low. Those sections are meant to be your own writing.
How Turnitin filters change the similarity figure
Turnitin shows three filters at the end of the report. If you are seeing the report online, they will be in a settings panel.
- Exclude quotes. Text inside quotation marks is left out when this is on. Forgot the quotation marks around a cited definition? Then the filter can’t help you.
- Exclude bibliography. If this is on, Turnitin tries to locate your reference list and leave it out. It searches for a heading like “References” or “Bibliography”. When the heading is unusual, or the references are split chapter-wise, Turnitin may not recognise them, and then those entries get counted.
- Exclude small matches. Matches below a set number of words (or below a set percentage) are ignored. The person running the check decides this threshold.
One thesis can give very different numbers based on these three switches. So before comparing two reports, or your report with your university’s limit, first see that the filters are same. Which exclusions are allowed depends on your university’s policy. Above that there are the UGC regulations, and our limits guide gives a summary of what UGC says.
On a PDF you cannot change the filters. If you feel wrong settings were used, request the person who ran the check to run it again with the settings your policy permits. Take the new report in writing.
Your thesis stays in the repository
Based on the account settings, a checked document may get stored in Turnitin’s repository. After that, your next version gets compared with your own stored copy also. This is the reason why checking drafts again and again can increase the final figure. Ask the research cell whether drafts go with the “no repository” option, and whether your earlier submissions can be excluded.
Worked example: reading a 23% Turnitin report
One scholar’s draft thesis comes back with 23%. Quotes and bibliography are excluded, and small matches under 10 words are also excluded. As per the university, the limit is 10%. Opened source by source, the report reads like this.
| Source | % | What it turned out to be | Action |
|---|---|---|---|
| 1. Submitted to [your university] | 6% | The scholar’s own synopsis, which was submitted two years back | Request the research cell to exclude it. This is not plagiarism |
| 2. A journal article | 4% | In the review chapter, a passage of two pages paraphrased very close to one paper | Write it again from your own understanding, and give the citation |
| 3. The scholar’s own published paper | 3% | Chapter 5 is written based on this paper | Cite the paper, declare it, and follow what your university says about own publications |
| 4. shodhganga.inflibnet.ac.in | 2% | Some standard definitions and a questionnaire scale, which were cited but not put in quotes | Put quotation marks around the definitions and cite the scale correctly |
| 5–40. Many sources | <1% each, 8% in total | Common phrases, methods wording, names of tests and institutions | Mostly you can leave these. Check the ones coming in results or conclusions |
First exclude the synopsis (6%) and put quotation marks on the definitions (about 2%). Without rewriting even one word, the figure comes down to around 15%. The passage in the review chapter (4%) needs proper rewriting. For that the scholar should keep the paper closed, write from what he understood, and only then put the citation. For the match to the own paper (3%), it depends on the university’s rule. The long tail may also reduce a little along with these changes.
Turnitin did its job correctly in this case, and it showed what it is designed to show. The headline figure had a settings issue, a citation issue, a genuine writing issue and harmless phrasing all mixed. Only going through it source by source separated them. Fixing citations chapter after chapter by hand is slow, and Zotero type reference manager saves a lot of time there. Our Zotero vs Mendeley guide has more on this.
What a Turnitin report cannot tell you
- Whether you plagiarised. People decide this. It is an academic judgement about attribution, and the software cannot make it.
- Whether a paraphrase is acceptable. Sometimes a passage has no highlight at all, but the scholar has taken the structure and ideas of one paper and not cited it. Turnitin shows nothing there. Examiners still catch it.
- What another checker will say. In many Indian universities DrillBit is used through INFLIBNET’s national licence, and many journals are using iThenticate. These tools search different databases, so for the same text they can give different figures.
- Whether AI was used. The similarity index tells nothing about it. Wherever an AI indicator is shown, Turnitin itself says it should not be the only basis for taking action against a student.
Use the report as a map that shows you where to look. Then you and your guide have to do the actual looking, and discuss about each problem area together. If your report shows real problems and you want help with them, our similarity report review goes through them with you, passage by passage, and you do the rewording. Journal submissions need the same care. There the editor sees an iThenticate report even before reading your paper, as our desk rejection guide explains.
Want a specialist to help with this in your own work? Similarity Report Review and Citation Repair: a free first consultation, then a written scope and fixed quote.
Sources
FAQ
Questions scholars ask
Why did my score go up when I checked the same chapter again?
Mostly the first check got saved in the repository, and then the second check matched with it. There can be other reasons also. Maybe the filters were set differently the second time. Or Turnitin added new sources to its databases in between.
Can I see the student paper my thesis matched?
You can’t see it directly. Whoever holds the Turnitin account can ask the institution where it was submitted. When it is your own old work, usually the research cell can confirm it and exclude it.
My bibliography is highlighted even though “exclude bibliography” is on. Why?
Turnitin did not recognise that part as a bibliography. Check your heading first. If it says something unusual, or your references come at the end of every chapter, Turnitin mostly misses them. You can use a standard heading. Otherwise ask them to exclude those matches manually.
Is a 0% or 1% report a good sign?
Not every time. There may be copied lines from material that is not in Turnitin’s database, and no quotation marks on them. Or the text may have gone through a paraphrasing tool, and examiners can make out that from the writing. A normal thesis will usually show some similarity.
Can I buy a Turnitin check for my thesis online?
Turnitin gives licences to institutions, not to individual scholars. Any third party selling “Turnitin checks” is running them on some institution’s account. Your thesis may then sit in a repository you don’t control, and later match against your own final submission. Better to use the checker your university provides.
