Direct answer
You detect plagiarism by running the text through a similarity checker that compares it against internet sources, publications and student papers, then reading the resulting similarity report match by match rather than trusting the headline percentage [1]. Turnitin's originality check works exactly this way: it returns a report listing every matched source with highlighted overlapping text, and a human decides whether each match is legitimate citation or copied material [1]. Detection finds text overlap, not intent — so the report is evidence to interpret, never a verdict on its own [1].
How Do Plagiarism Detectors Actually Find Copied Text?
Plagiarism detection is string matching at scale. A detector converts the submitted document into searchable text, compares it against an indexed corpus, and flags sequences that overlap with something already in that corpus [2]. Turnitin's corpus combines internet pages, academic publications and previously submitted student papers, and the report separates matches into those source categories so you can see where the overlap came from [2]. This is why the tool catches verbatim copying and close paraphrase reliably but cannot catch an idea you read once and rewrote from memory without ever matching a source [2].
The practical consequence is that detection is only as good as its index. If a source is not indexed — a paywalled book, a lecture handout, a translated article, a private class document — the matcher has nothing to compare against and returns nothing, even when the text was copied [2]. A clean report therefore means "no match found in the indexed corpus," which is a narrower claim than "original" [2].
The report also depends on settings. Most systems can exclude quotations and the bibliography from the calculation, and turning those exclusions on or off changes the reported figure without changing a single word of the paper [2]. When you read any similarity report, check which exclusions were applied before you react to the number, because the same draft can produce visibly different scores under different settings [2].
Finally, detectors do not distinguish between "copied" and "correctly quoted." A properly cited block quote will still appear as a match, because the matcher sees identical words [2]. That is a feature, not a bug: it forces the reader to look at the context of each match, which is exactly the judgment a detector cannot make for you [2].
What Does a Similarity Score and a Similarity Report Actually Tell Me?
The similarity score is the percentage of a paper's text that matches other sources — nothing more and nothing less [3]. It is an indicator, not a judgment of intent, and Turnitin is explicit that the number should be the start of a review rather than the end of one [3]. Two papers with identical scores can be completely different cases: one may be a paper of unattributed copying, the other a literature review that quotes heavily and cites correctly [3].
A low score does not clear a paper. If the source sits outside the index, or if the text was paraphrased far enough to break the string match, the score stays low while the borrowing is real [3]. Students sometimes treat a single-digit score as a guarantee and stop reading the report; that is the most common misreading of the tool [3].
A high score does not convict a paper either. Instructors are directed to open the report, inspect the highlighted passages and their surrounding context, and consider the assignment type before drawing any conclusion [3]. A methods chapter full of standard boilerplate, or a paper that quotes a primary source at length, can score high for entirely defensible reasons [3].
What the report genuinely gives you is a map. It shows which sources were matched, how much of the paper each source accounts for, and exactly which sentences overlap [3]. Read it that way — as a list of places to look — and the score becomes useful; read it as a grade, and it will mislead you in both directions [3].
How Can I Check My Own Draft for Plagiarism Before I Submit It?
The reliable method is to run your draft through a check that mirrors the one your institution uses, then fix the flagged passages before the real submission [4]. Mirroring matters because different checkers index different corpora; a clean result from a tool that does not include student papers tells you little about what your university's system will find [4]. Run the check with the same exclusion settings your department applies, so the report you see resembles the report your marker sees [4].
Before you check, make sure your citations and quotations are intact and correctly formatted [4]. Legitimate matches are far easier to dismiss when the quotation marks and references are already in place, and leaving them out turns a defensible match into an apparent one [4]. If you are checking a draft that has no citations yet, add them first — otherwise you will spend the review fixing formatting rather than substance [4].
When the report comes back, work through it source by source instead of reacting to the total [4]. For each highlighted passage, ask one question: is this attributed? If it is quoted and cited, leave it. If it is paraphrased but the idea is someone else's, add the citation. If it is copied prose you meant to rewrite, rewrite it in your own words and re-check [4].
One structural obstacle is worth knowing in advance. Institutional repositories generally do not allow students to self-submit into the same database their assignments are checked against, so you cannot simply preview your own paper inside the university system [4]. That is why students use a non-repository preview service: the file is compared against the corpus without being added to the student paper database, so the draft is not permanently indexed and the report is not shared onward [4]. Turnitin0 runs that preview in the format your marker will see — a Turnitin AI detection report and a similarity report delivered as two downloadable PDFs, usually in under 15 minutes, with 100,000+ reports delivered to 20,000+ students and a 4.9/5.0 satisfaction rating. A single check is $3.80, or $2.75 per check in a 10-check pack.
Reading a report you can actually see changes how you revise: instead of guessing whether a paragraph will be flagged, you open the same cover page, score, source flags and similarity summary your instructor would open, and fix the specific sentences that light up. That is the difference between hoping your draft is clean and knowing it is before the deadline closes.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Does a 0% similarity score mean my paper is plagiarism-free?
No. A zero means nothing in the indexed corpus matched, which can happen because the text is original or because the source is not indexed [3]. Paywalled books, lecture notes, translated material and non-indexed websites can all be copied without producing a match [2].
Can a plagiarism detector tell the difference between copying and correct citation?
Not by itself. The matcher flags identical word sequences regardless of attribution, so a properly quoted passage still appears as a match [2]. Attribution is what a human reader — or you, reviewing your own report — supplies when judging each highlight [3].
What is the fastest way to check a draft before submitting it?
Run it through a non-repository checker that mirrors your institution's corpus and exclusion settings, then review each flagged source in order [4]. Turnitin0 returns a Turnitin AI detection report and a similarity report as two PDFs, typically within 5–15 minutes, without adding the file to the student paper database.
Does checking my own draft get it added to Turnitin's student database?
It depends on the service. Institutional submissions are normally stored in the repository, which is also why students cannot freely self-submit there [4]. A non-repository preview such as Turnitin0 checks the file without indexing it and does not share reports with third-party databases.
What should I do with a passage that gets flagged?
Ask whether the idea is attributed. If it is quoted and cited, keep it; if it is someone else's idea in your words, add the citation; if it is copied prose, rewrite it and re-check [4]. Re-running the check after revision is the only way to confirm the fix worked.