Every thesis or dissertation goes through an originality check before submission. The result is a report showing what percentage of your text matches other sources. That percentage is not automatically proof of plagiarism. This article explains how the process works, what the systems actually compare, why a higher similarity score does not always indicate a problem and how to write and cite correctly to stay on the right side of academic integrity rules.
What is plagiarism and what forms does it take
Plagiarism means using someone else's ideas, wording or data without properly crediting the source. It is important to understand that plagiarism comes in several forms, some of which are far less obvious than others.
Intentional plagiarism
This is the deliberate copying of text from a book, website or another paper without citation. The author knowingly presents someone else's work as their own. This is a serious breach of academic integrity and can result in disciplinary action or, in some cases, legal consequences.
Unintentional plagiarism
Unintentional plagiarism usually results from poor source management, inadequate paraphrasing or simply forgetting to add a citation. The student typically has no intention to cheat, but the outcome looks the same from an academic rules perspective. This is the most common type and the easiest to prevent.
Mosaic plagiarism (patchwriting)
Mosaic plagiarism happens when a writer rearranges sentences from multiple sources, changes a few words but keeps the original structure and ideas intact. The result looks original on the surface but is essentially hidden copying. Originality checking systems can partially detect this, but an experienced supervisor will often catch it more reliably.
Self-plagiarism
Self-plagiarism means reusing your own previously submitted text without disclosing it. If you copy sections of an earlier essay into your thesis without citing the original, most universities consider that a violation. A thesis must represent new, original work.
How originality checking works in UK, US and international universities
The most widely used system internationally is Turnitin, which compares submitted work against a database of billions of web pages, academic journals and previously submitted student papers. Many universities use Turnitin as their primary tool, while some institutions use alternatives such as Ouriginal or iThenticate for postgraduate research.
When you submit a paper, the system compares it against:
- previously submitted student work stored in the Turnitin repository,
- live and archived web pages,
- academic journals and publications,
- institutional repositories and open-access databases.
The output is a Similarity Report with an overall similarity score and a highlighted view of matched passages. Crucially, it is the supervisor or examiner who decides whether a match represents plagiarism, not the software. More on how AI text detection tools work alongside these systems can be found in the article How AI-generated text is detected.
What the similarity score means and why it is not the same as plagiarism
The similarity score is the percentage of your submission that matches text found elsewhere. This is the most misunderstood number in academic writing. There is no single universally accepted threshold that applies to every university, department or paper type.
Your similarity score can be elevated for entirely legitimate reasons:
- Properly cited direct quotations (quoted text with a source reference still registers as a match),
- Standard academic and technical terminology that appears in many papers,
- References and bibliography entries (titles, author names and journal names are matched),
- Names of laws, institutions and standardised concepts,
- Transitional phrases commonly used in academic writing.
A paper with many correctly cited sources may score higher than one where nothing is cited but ideas are still copied in spirit. The interpretation always sits with the human reader, not the algorithm. If you are unsure what score is acceptable at your institution, ask your supervisor directly.
How to cite and paraphrase correctly
Proper citation and paraphrasing are the most effective tools against plagiarism. A detailed guide to citation formats can be found in the article How to cite according to ISO 690, but the core principles are straightforward.
Direct quotation
Use a direct quotation when the exact wording is significant. Place the text in quotation marks and include a citation with the page number. Keep quotations selective: they should support your argument, not replace your analysis.
Paraphrasing
Paraphrasing means expressing someone else's idea in your own words and sentence structure. A citation is still required. The common mistake is changing only one or two words while keeping the original structure intact, which is the hallmark of mosaic plagiarism. A genuine paraphrase demonstrates that you understood the source, not just that you read it.
When no citation is needed
Widely known facts and information that comes directly from your own research, measurements or original analysis do not need a citation. The rule is simple: if the idea is not yours, say where it came from.
Common mistakes that lead to unintentional plagiarism
Most cases of unintentional plagiarism trace back to a handful of recurring errors:
- Pasting source text into a working document without immediately noting the source. Over time, students forget where the text came from.
- Translating text from another language without citing the original. A translated passage is still someone else's idea.
- Paraphrasing too closely. Changing one or two words is not enough.
- Omitting the source for a figure, table or graph taken from another work.
- Overlooking self-plagiarism when reusing content from earlier submissions.
AI text detection
Beyond classic plagiarism checking, a growing number of institutions use tools designed to detect AI-generated text. Turnitin has added an AI writing detection module, and other platforms offer similar features. These tools are not infallible and can produce false positives, which is why most universities treat the output as one signal among several rather than definitive proof. For a deeper look at how this works, see the article on AI text detection linked above.
What to do before submitting your thesis
A few practical steps will help you submit with both a clear conscience and a clean similarity report:
- Keep a running source log. Every time you read something you might use, record the author, title, year and page number immediately.
- Review every citation. Check that all direct quotes are in quotation marks and every paraphrase has a source reference.
- Practice proper paraphrasing. Read the original, set it aside and write the idea from memory in your own words.
- Run a preliminary check if possible. Some universities allow you to upload a draft before the official submission deadline. Use that opportunity.
- Talk to your supervisor. If you are concerned about your similarity score, raise it with your supervisor before submission rather than after.
FAQ
What similarity score is considered acceptable?
There is no universal number. Every university, department and supervisor applies their own criteria. What matters is what is behind the matched text, not the figure itself. Some institutions publish internal guidelines; others leave it to the supervisor's discretion. Check with your department directly.
Can a correctly written thesis still be flagged for plagiarism?
The similarity report does not label anything as plagiarism. It identifies matches. Whether those matches constitute plagiarism is determined by your supervisor or examination committee after reviewing the highlighted passages in context.
What is self-plagiarism and is it treated seriously?
Self-plagiarism means reusing your own previously submitted work without disclosure. Most universities treat it as a breach of academic integrity because a thesis must be original. If you want to build on earlier work you submitted, discuss it with your supervisor and cite the original piece.
Does Turnitin check text in foreign languages?
Yes. Turnitin and similar tools compare text across multiple languages and can match content in translation. Translating a passage from another language does not make it original.
What if the system flags my own research data?
Original data, charts and conclusions generated by your own research should not trigger matches unless you have published them elsewhere. If they do appear, your supervisor can exclude those sections when interpreting the report.
Does a high similarity score mean I will fail?
Not automatically. A high score triggers a closer review by your supervisor or institution. The outcome depends entirely on what the matches represent. Correctly cited quotations and standard terminology are not a problem. Uncited copied passages are.