Direct answer
A 100% similarity score is almost never proof that you plagiarised; it means every word of your document matched something in the comparison database, which most often happens with a resubmitted draft, a copied template, a fully quoted block, or a misconfigured exclusion setting. You remove it by identifying the source of each match in the report, then either excluding legitimate material (quotes, bibliography, your own earlier submission) or rewriting and citing the passages that genuinely came from someone else. The score falls only when the underlying text changes — it cannot be lowered by reformatting, adding filler, or hiding text.
What Actually Causes a 100% Plagiarism Score?
A similarity score is a measurement of matched text, not a verdict on intent, and Turnitin states this explicitly when it explains how the percentage is calculated [1]. The number represents the proportion of your document that overlaps with sources in the comparison repository, so a score of 100 means the entire document was found somewhere — a signal to investigate, not a conclusion about your integrity [1]. In practice, whole-document matches cluster around a small number of causes, and identifying which one applies to you determines whether there is anything to fix at all.
The most common innocent cause is resubmission. If you upload a draft, revise it, and upload again, the new version can match the earlier version almost entirely, because the earlier version is now stored and comparable [2]. A second cluster involves structural material: title pages, cover sheets, tables of contents, standard methodology boilerplate, and reference lists are frequently copied verbatim from templates or from the sources themselves, and together they can account for a large share of a short paper [2]. Turnitin's report lets you review each match individually and set exclusions, but which exclusions are available depends on how the instructor configured the assignment [1][2].
A third cause is a genuine one: a block of text — sometimes the whole essay, if it was recycled from a prior course or purchased — was copied. The report distinguishes these for you, because matches are listed source by source with the overlapping text highlighted [1]. Read the list before you panic: if the top match is your own earlier file or a template, you are dealing with a settings problem; if it is a journal article or a website you never cited, you are dealing with a writing problem [2]. Treating the first as the second wastes hours and can make the draft worse.
There is also a technical explanation worth ruling out early. Repository settings, small-match exclusions, and whether quotes and bibliography are included all shift the number, and a student-facing check run under different settings can report a very different percentage than the instructor's view [1][2]. That is why the first diagnostic step is always to open the match list and read the sources, not to start rewriting sentences at random.
How Do You Rewrite and Cite Flagged Text So the Score Drops Legitimately?
Legitimate score reduction comes from changing the relationship between your sentence and its source, and citation style guides are precise about what that means. Purdue's OWL paraphrase guidance defines an acceptable paraphrase as one that restates the source's meaning in your own words and your own sentence structure, with the source attributed — swapping synonyms into the original syntax while keeping the original order is patchwriting, which is still treated as plagiarism [3]. This is the single most useful distinction for anyone trying to lower a score, because patchwriting is exactly what produces stubborn mid-range matches that never quite disappear.
Work through flagged passages in this order. First, decide whether the passage needs to be there at all — deleting redundant background is the fastest legitimate way to reduce matched text, and it improves the argument. Second, for passages that must stay, close the source, write the idea from memory in your own structure, then reopen the source and check that you have not reproduced its phrasing; add an in-text citation for the idea itself, not just for direct quotes [3]. Third, for text that genuinely must be quoted word for word — definitions, statutory language, a distinctive formulation — use quotation marks and a full citation, and expect the match to remain, because a correctly quoted passage is supposed to match [3].
Two habits undermine this work. The first is paraphrase-then-cite, where a student rewrites but omits the citation because "the words are mine now" — the idea is still the source's, so the citation is still required [3]. The second is relying on a thesaurus pass; automated synonym substitution preserves the original sentence architecture, which is precisely the pattern that plagiarism detection is built to recognise, and it also tends to produce prose that reads badly [3]. Rewriting properly takes longer per paragraph but removes the match permanently instead of shifting it a few percentage points.
Finally, clean up the material that inflates scores without being anyone's fault. Standardise your reference list to your required style so entries are correctly formatted rather than copied wholesale, replace template cover pages with your own, and cut boilerplate methodology sentences down to what your specific study actually did [3]. These changes are legitimate, they are good academic practice, and they remove matched text rather than disguising it.
How Can You Verify the Score Dropped Before You Submit?
Verification requires a second measurement under comparable conditions, because a score only means something relative to the settings that produced it. A pre-submission check exists precisely so a student can see the same report type the instructor will see, including the match list and the highlighted overlaps, before the work is finalised [4]. Comparing two reports generated under different exclusion settings tells you nothing useful, so keep the configuration consistent between your first and second check [4].
Read the second report the way an instructor would: look at the top matches, confirm that remaining overlaps are quoted material, bibliography, or unavoidable standard phrasing, and check that no source appears that you failed to cite [4]. If a match survives, open it and read the highlighted text — a surviving match to a source you have now paraphrased and cited usually means the paraphrase still tracks the original too closely, which sends you back to the rewriting step rather than to another exclusion [3][4]. Documenting this loop, draft → check → revise → re-check, is also the clearest evidence of good-faith process if a tutor ever asks.
Timing matters as much as method. Institutional submission portals often allow a single submission, and repository settings may store your file permanently, so the check you run before the deadline is the one that counts [4]. Running it early enough to leave a full revision cycle — not twenty minutes before the deadline — is what makes the verification step useful rather than merely anxiety-inducing [4]. If your first score was 100%, expect the second to be dramatically lower once a resubmission or template cause is removed, and expect a smaller, more meaningful drop if the cause was genuine copying that you have now properly rewritten and cited [1][4].
Once your rewrite is done, the only way to know the number actually moved is to look at a real report rather than guess. turnitin0 produces the same Turnitin AI and similarity reports your institution sees — match list, highlighted overlaps, and score summary included — so you can run that verification pass on the revised draft before the version that counts is submitted.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
Get Real Turnitin AI & [Similarity Report](https://www.turnitin0.com/guides/us/ai-content-detector)
FAQ
Does a 100% similarity score mean I will be accused of plagiarism?
No. The score measures matched text against the repository, and Turnitin itself frames it as a starting point for review rather than a finding [1]. A resubmitted draft or a copied template can produce 100% with no misconduct at all, which is why the match list matters more than the headline number [2].
Can I lower my score by changing the formatting or adding my own sentences?
Neither works. Reformatting, changing fonts, or padding the document with new paragraphs does not alter the matched passages, and the report recalculates from text overlap [1]. Only rewriting the flagged text in your own structure, with citation, or removing it, actually reduces the percentage [3].
If I paraphrase properly, will the match disappear entirely?
Usually, but not always. A correct paraphrase with an in-text citation normally clears the match, whereas a correctly quoted passage will still show as a match because it is supposed to be identical [3]. What matters is that remaining matches are attributable to quotes, bibliography, or standard phrasing [4].
Why did my score go up after I revised my essay?
The most likely explanation is that you resubmitted into a repository that now holds your earlier version, so the two drafts match each other [2]. Repository and exclusion settings also differ between checks, which can shift the number without any change in your writing [1][4].
How do I know the revised draft is actually clean before the real submission?
Run a check that shows the match list and highlighted overlaps, then read every remaining match and confirm you can attribute it [4]. Re-checking after each revision cycle, with consistent settings, is the only reliable confirmation [1][4].