The turnitin similarity index is the percentage at the top of your similarity report — the share of your paper's text that matches text in Turnitin's database. If your paper is 2,000 words and 300 of those words match existing sources, your index is roughly 15%. That's all it is: a raw overlap number. It doesn't know whether the match is a cited quote, your bibliography, or a phrase so common it appears in a thousand papers. Reading it correctly starts with understanding how it's built — and what the breakdown underneath it is telling you.
What the Turnitin Similarity Index Actually Measures
The similarity index measures one thing: text overlap. Not originality, not effort, not honesty. Just how much of your document's wording also exists somewhere in Turnitin's collection of web pages, journal articles, and student papers — matched words divided by total words, multiplied by 100.
Two things make this number easy to misread. First, Turnitin matches text mechanically — a direct quote with a perfect citation counts the same as a copied paragraph with no attribution. Both are just matched text to the machine. Second, the single number hides all the detail. A 22% index could mean one source matching 20% of your paper (a real problem) or forty tiny sources matching half a percent each (usually common phrases). The percentage alone never tells you which — the breakdown does.
How Turnitin Calculates the Similarity Index
When you submit a paper, Turnitin converts your document into text and scans it against its databases in chunks. Every section that closely mirrors an existing source gets highlighted, and the highlighted words are added up. Divide that total by your paper's word count and you get the index.
What counts as a match
- Verbatim or near-verbatim blocks: passages that mirror a source word-for-word (or nearly) get flagged — copied sentences, patch-written paragraphs where you've swapped a few words, or text lifted from web pages.
- Matches across the whole document: Turnitin scans body text, headings, footnotes, figure captions, and appendices. A matched methods section or quoted definition counts the same as a matched paragraph of analysis.
- Multiple small matches: common phrases, standard definitions, and field-specific terminology can each match different sources. Individually they're tiny; together they can add up to several percentage points.
What may not count — depending on settings
Your instructor controls filters that can remove whole categories of matches from the calculation:
- Bibliography exclusion: removes your references list, since journal titles and book entries exist online verbatim. Without this filter, a long references page can inflate your index on its own.
- Quoted material exclusion: removes text inside quotation marks, on the assumption that properly quoted material is already attributed.
- Small-match exclusion: drops matches below a set word count, filtering out coincidental common phrases.
Reading Your Turnitin Similarity Index Score Breakdown
The real report isn't the number at the top — it's the Match Overview panel: the per-source breakdown listing every source your paper matched, each with its own percentage, ordered largest to smallest. Understanding this panel is the whole game.
The per-source list
Each entry in the Match Overview represents one source — a web page, journal article, or student paper — and its percentage shows how much of your total word count matches that source. So if "Source 3" shows 4%, it means 4% of your paper's words mirror Source 3. Add up all the source percentages and you get (approximately) the total index. That's the relationship: the index is just the sum of the breakdown.
This changes how you read the report. If your index is 18% and the top source is 12%, one source is doing most of the work — go look at it first. If the top source is 1.5% with thirty sources below it, your matches are scattered and almost certainly harmless: common phrases, titles, standard terminology.
Clicking through a match
Clicking any source highlights exactly which passages of your paper matched it. For each top source, ask: Did I cite it? A match that's a quote or paraphrase with a proper citation is fine — the index flagged mechanics, not misconduct. Is it boilerplate? Method descriptions and standard definitions match constantly; nobody expects you to rewrite them. Is it mine? If the source is a paper you submitted to a previous class, that's a self-match from the student repository — your own earlier work stored against you.
Why the Same Paper Scores Differently on Different Checkers
Students are often confused when a free online checker says 6% and Turnitin says 19% — or vice versa. The reason is simple: every checker searches a different library. Turnitin's database includes the student-paper repository, published journal content, and archived web pages. Free tools mostly search the live web. A source that one tool knows about is invisible to the other.
The index is always relative to the database behind it, so the same paper genuinely has different similarity indexes on different tools — and neither is "wrong." They're measuring overlap against different collections. The number that matters is the one from the tool your instructor or institution actually uses, because that's the comparison your paper will be judged against.
Turnitin's database also grows daily, so the same paper can index slightly differently weeks apart — the index is a snapshot, not a fixed property of your paper.
How to Read the Number Without Panicking
Here's the calm, practical routine:
- Ignore the headline number first. Go straight to the Match Overview.
- Work the top three sources. Check each for citation, boilerplate, or self-match. Most anxiety comes from a source you've already cited properly.
- Confirm the filters. If your bibliography or quoted material is included, the number is inflated — ask whether your instructor excludes them, or get a check where they are.
- Don't chase zero. A low single-digit index built from scattered tiny matches is normal academic writing; rewriting standard terminology to dodge it usually makes your paper worse, not safer.
For related questions, see our guides to the report's color bands (what the blue-to-red indicator next to your number signals) and score benchmarks (what ranges instructors generally consider acceptable) — both in the TurnPro AI blog guides.
Frequently Asked Questions
Is the Turnitin similarity index proof of plagiarism?
No. The index measures text overlap mechanically — it can't tell a cited quote from copied text. Two papers can share a 20% index where one is perfectly cited and the other isn't. Plagiarism is an academic judgment your instructor makes after reading the matched passages.
Why did my similarity index change when I resubmitted the same paper?
Because the index isn't frozen. Turnitin's database grows daily as new papers and web pages are added, and your instructor's filter settings may have changed between submissions. A quote excluded on your first draft can count on the second if filters were turned off.
Does the similarity index count my references and bibliography?
It depends on your instructor's settings. Turnitin can exclude the bibliography and quoted material — and instructors usually turn these filters on — but not always. If your references page is full of journal titles that exist online, an unfiltered report can add several percentage points unrelated to your writing.
Why is my index different from the one my friend got on another checker?
Every checker searches a different database. Turnitin's strength is its student-paper repository and publisher content; free tools mostly search the open web. A source one tool knows about is invisible to the other — so the same paper can score 8% on one checker and 18% on another. Neither is "wrong."
Can I check my Turnitin similarity index without my instructor?
Yes — a non-repository third-party check gives you the official Turnitin similarity index without storing your paper. Never upload coursework to a standard assignment inbox just to see the number; those submissions are saved into the student repository and will match your own future work.
Final Thoughts
The Turnitin similarity index isn't a grade on your honesty — it's a measurement of text overlap, built by adding up the per-source matches in the Match Overview. Read the breakdown instead of the headline number and the report stops being scary: it shows you which passages to cite better, rewrite, or leave alone. Get the full picture with an official Turnitin similarity report from TurnPro AI — non-repository, never stored, delivered in 5–30 minutes via WhatsApp.