Direct answer
Teachers tell whether a paper is plagiarized by combining a software similarity report with their own reading of the text. The similarity report flags stretches of wording that match other sources, and the teacher then opens each match to judge whether the overlap is quoted, cited, badly paraphrased, or copied outright [1]. No single percentage decides the outcome: markers look at the size, number, and pattern of matches, and at whether the student's voice stays consistent across the document [1]. Turnitin itself describes the score as an index of matched text rather than a measure of plagiarism, so the human judgement step is where the actual decision is made [3].
What do teachers actually look for when they suspect a paper is plagiarized?
The first thing an experienced marker notices is usually a change in voice. Teachers read dozens of scripts from the same cohort and build a feel for how a student writes — sentence length, favourite connectives, how they handle hedging, whether they use British or American spelling consistently. A sudden jump into denser, more uniform prose, or a conclusion that sounds like a different author wrote it, is enough to send a script for a closer look [2].
The second signal is structural rather than stylistic. Copied work often arrives with a mismatch between the sophistication of the argument and the quality of the surrounding scaffolding: a polished paragraph sitting inside a clumsy transition, or a citation list that does not quite match the sources actually used. Markers also notice paraphrasing that follows the original's sentence order too closely, swapping synonyms while keeping the same clause structure [2].
Once a script is flagged, the teacher's workflow is fairly mechanical. They open the similarity report, sort the matches by size, and open the largest source first to compare the two texts side by side. From there the judgement is about intent and attribution: is the overlapping passage in quotation marks with a citation, is it cited but poorly paraphrased, or is it uncited and unmarked? Referencing errors are handled differently from deliberate copying in most academic misconduct procedures [2].
Finally, teachers cross-check against the student's own history. Earlier submissions, drafts, and formative work give a baseline, and a marker who has seen a student's real writing for a semester has far more evidence than the report alone provides. That is why the same similarity score can produce very different conversations in two different modules [2].
How does Turnitin's similarity report work, and what does a high similarity score really mean?
The similarity score is the percentage of a submission's text that matches other sources in the databases Turnitin compares against — internet pages, academic publications, and previously submitted student papers. It is not a plagiarism percentage, and Turnitin explicitly states there is no "safe" threshold below which a paper is automatically fine [3]. A paper at 18% can be a serious problem and a paper at 34% can be entirely legitimate, depending on what the matches actually are.
The score is also configurable, which surprises many students. Instructors can exclude quoted material, the bibliography, and matches shorter than a set word count, so the identical document can produce different percentages in two different courses [3]. The report separates matches by source type, so a marker can see at a glance whether the overlap came from a journal article, a website, or another student's essay submitted in a previous year [3].
What matters is the shape of the report, not the headline figure. One match covering most of a paragraph in a single source reads very differently from thirty short fragments spread across properly cited references. A single long, unbroken block of matching text with no quotation marks is the pattern that most reliably triggers a misconduct referral, while a scatter of short, cited overlaps usually just reflects standard academic phrasing [3].
Can students see the same similarity and AI signals before they submit?
Not always through the institution. Whether a student can run a draft through Turnitin before the deadline depends entirely on how the university has configured its account, and many students only see a report after the assignment has been graded. The AI writing indicator is likewise an instructor-facing feature that reports the share of qualifying prose flagged as AI-generated, and it shows an asterisk rather than a number when the signal falls below its confidence threshold [4].
The AI indicator works by classifying text sentence by sentence and then aggregating those results into a document-level figure. That design detail matters practically: lightly editing a handful of sentences rarely clears a flag, because the surrounding sentences still carry the same statistical signature [4]. Turnitin also frames the indicator as advisory evidence for instructors rather than an automatic misconduct finding, which is consistent with the human-judgement step described above [4].
Because pre-submission access is inconsistent, the practical gap for students is visibility. Seeing the similarity percentage, the individual matched sources, and the AI indicator on your own draft before the deadline turns an invisible risk into something you can actually fix — and it is the same evidence a marker will be looking at [4].
Reading about detection only gets you so far; the useful next step is seeing your own draft through the same lens your marker will use. turnitin0 produces a Turnitin AI detection report and a similarity report together in one checkout, so you can open the matched sources, check the flagged passages, and decide what needs rewriting while there is still time to rewrite it.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Does a high similarity score automatically mean I plagiarized?
No. The score is the percentage of text matching other sources, and properly quoted, correctly cited material still counts toward it [1]. Markers open the individual matches and judge each one on attribution and intent.
Can teachers tell if I paraphrased something from a website?
Often yes. Paraphrasing that keeps the original's sentence order and clause structure is still detectable as a near-match, and it is one of the patterns markers look for specifically [2].
Will my university see that I used an outside checker?
No. A non-repository check does not add your file to Turnitin's student paper database and is not shared with third-party databases, so it does not appear in your institution's records.
How long does it take to get a Turnitin report?
With turnitin0, most orders finish in 5–15 minutes and delivery is guaranteed within 30 minutes even during queue spikes.
What if my paper is already flagged?
Focus on the specific matches rather than the overall number. Rewriting the largest unbroken block of matching text and fixing missing citations addresses the patterns that actually trigger referrals [3].