Direct answer
Checking an assignment for plagiarism means running your own draft through a similarity checker and reading the matched passages before your institution does, so you can fix uncited copying, weak paraphrasing, and over-quoted sections while you still have time. The number alone tells you very little: Turnitin's similarity score is the share of your text that matched a source, not a verdict on your integrity [1]. What matters is which passages matched and whether each one is properly quoted and cited. A pre-submission check is therefore a reading task first and a scoring task second.
How do you check an assignment for plagiarism before you submit it?
The practical sequence is upload, wait for processing, then review every highlighted match in context. You submit a supported file format — typically .docx, .pdf, or .txt — and the system compares its text against internet sources, academic publications, and previously submitted student papers [2]. When the report opens, you are not looking at a percentage to feel good about; you are looking at a list of passages that exist somewhere else in almost identical wording [2]. Each one needs a decision: is it quoted with quotation marks and a citation, is it paraphrased closely enough to still need a citation, or is it uncredited text that must be rewritten?
Work through the matches in descending order of length, because the longest overlaps are where the risk concentrates. A 40-word block that mirrors a journal abstract is a citation problem; a three-word overlap on "the results of this study" is noise. Turnitin's report tools let you exclude quotes, bibliography entries, and small matches so the remaining view shows only the material that genuinely needs your attention [1]. That filtered view is the one worth acting on.
Then revise and check again. Students routinely fix the flagged sentence and never re-run the draft, which means they never learn whether their rewrite introduced a new match or whether the fix actually worked [2]. A second pass on the revised file takes minutes and converts an assumption into evidence. If your institution offers a draft-submission space or an in-course checking tool, use it for that second pass — it is the cheapest confirmation available [2].
Finally, keep the report. If a tutor later questions a passage, a dated pre-submission report showing how you handled the match is far stronger evidence than your recollection of editing it [2].
What does a plagiarism score actually mean, and what counts as too high?
There is no universal number that means "too high." Turnitin states plainly that the similarity score is a proportion of matched text and that no fixed percentage indicates plagiarism [3]. A 28% report on a literature review dense with quoted definitions may be entirely clean, while a 9% report containing one unquoted paragraph lifted from a single article is a serious problem [3]. The score measures overlap, and overlap is caused by citation practice as much as by misconduct.
The composition of the score matters more than its size. Reference lists, direct quotations, and standard methodological phrasing all inflate the total, which is why excluding bibliography and quotes is standard practice when interpreting a report [3]. What remains after those exclusions is the part that reflects how you handled other people's ideas. If that residual is a handful of short, cited phrases, you are in normal territory; if it is long uninterrupted blocks, the score is telling you something real [3].
Institutions set their own thresholds and read the score next to the highlighted text, not in isolation [3]. That is why two students with identical percentages can receive very different outcomes: one cited properly and one did not. Treat the number as a pointer to passages worth reading, and treat the passages as the actual evidence.
One more distinction worth holding onto: similarity is not the same as AI writing detection. A fully human-written essay can score highly on similarity through heavy quotation, and an AI-drafted essay can score low on similarity while still being flagged by an AI detector. They are separate reports answering separate questions, and a complete pre-submission check looks at both.
Can a third-party checker show the same similarity report my university will see?
Only if it runs the same engine against the same databases. Your institution's report is generated inside its own Turnitin account and delivered through its learning management system, so a preview is only meaningful when it reproduces that report — the same similarity categories, the same source overlays, the same highlighted passages your tutor will open [4]. A checker that produces its own proprietary "plagiarism percentage" from a different index is answering a different question, and a clean result there proves nothing about the report that actually decides your grade.
Report fidelity is what makes a preview actionable. If the preview shows the internet/publications/student-paper breakdown and lets you click into each matched source, you can fix the exact passage your tutor will see [4]. If it only returns a number, you have learned nothing you can act on. Turnitin's own guidance on academic integrity stresses that these conversations work when the evidence is transparent and comparable between what the student saw and what the institution saw [4] — which is precisely the argument for previewing the real report rather than an approximation of it.
There is also a database side-effect to watch. Some checking services add your uploaded draft to a student paper repository, which means your own draft can later match your own final submission, or another student's, and appear in a future report [4]. For draft checking you want a non-repository check: the file is compared against existing sources but is not stored as a source itself. That single detail prevents a self-inflicted similarity match at the worst possible moment.
Used correctly, a faithful preview changes your editing behaviour. You stop guessing which paragraph is risky and start working from the same evidence your tutor will use [4].
Reading a real report is what turns a vague worry into a short, fixable list — and turnitin0 exists to put that report in your hands before the deadline rather than after it. Upload your draft to turnitin0 and you get the Turnitin AI detection report and the similarity report together, generated the way your institution sees them, without your file being added to a student paper database.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Is a 0% similarity score good?
Not automatically. A 0% score can mean original writing, but it can also mean the checker's database is thin or your file was not processed properly [3]. Read the matched-source breakdown before treating a low score as clearance.
Should I check my assignment before or after proofreading?
Check after proofreading but before final submission. Editing changes your wording, and a check run on an early draft can flag passages you have already rewritten [2].
Does checking my draft add it to Turnitin's student paper database?
It depends entirely on the service. Institutional submissions may be stored as sources, while a non-repository check compares your file without retaining it as a future match [4]. Confirm the setting before you upload.
Will my university know I ran a pre-submission check?
A private draft check does not appear in your institution's Turnitin account; only work you submit through your course does [1]. Your tutor sees the submission, not your preparation.
Can I check for plagiarism and AI writing at the same time?
Yes, and you should. Similarity and AI detection are separate reports, so a complete check returns both — a low similarity score says nothing about whether your prose will be flagged as AI-generated [1].