Turnitin Similarity Score Explained: What the Percentage Actually Means

What the similarity score measures

The similarity score of Turnitin is the percentage of the document text that resembles text from comparison sources. It does not in-and-of-itself amount to a plagiarism verdict. A properly quoted text, entries of bibliography, assignment format, phrases of common use or names of institutions or laws can also be a part of the score. It's not just the percentage that's useful, but the source list, highlighted excerpts and everything else that stands behind the number.

Most often students panic seeing a high percentage. However, the first question should be simple – what type of text is being matched? A methods section, reference list, or required cover page can add to the number, but that does not mean the paper is fraudulent. A serious issue could be hidden behind a low score where a critical paragraph was lifted from a primary source.

Why a score can look higher than expected

Similarity scores may increase when there are long quotes, repeated assignment questions, tables taken from public documents, or very standard definitions. They can also rise when a student submits multiple drafts to systems. That is why a no repository check may be useful before a final submission as it checks for matching text without adding your draft to a repository.

The institution or service uses specific settings for calibrating the score. Excluding the bibliography, quoted text, or small matches changes the visible percentage. As a result, two reports for a single file may look different even with unaltered writing.

How to read the report

Begin with the most significant matched sources, not the percentage. Check the major source cards and the highlighted text. In case the match is a cited quote, check for quotation marks and citation. If the text is a paraphrase with no citation then change it and source it. If the citation is a reference entry, nothing needs to be done.

The best review habit is to distinguish one’s safe matches from risky ones. Non-threatening matches are typically references, required prompts, standard labels or short phrases. Risky matches are full sentences, paragraphs, unique phrasing or source structure that appears without citation.

Practical ways to lower risky similarity

• Add citations where borrowed ideas are used.
• Rewrite source-dependent paragraphs in your own structure, not just with synonyms.
• Keep direct quotations short and clearly marked.
• Remove assignment prompt text if your instructor does not require it.
• Check whether bibliography and quoted text exclusions are appropriate.
• Use the paraphrase tool for drafting help, then manually review the result.

What is a good score?

No safe percentage that works for all. Certain universities consider 10 to 15% as normal for essays of limited length, while essays of longer length with larger number of references may sit higher. A more pertinent question is whether the matching text is explainable. A report with a 22% similarity percentage that consists mainly of references and quoted material is less worrying than a 6% report with one identical paragraph.

If you'd like to obtain a full report instead of just a percentage, simply run the document checker and check the highlight source instead.

FAQ

Is a high Turnitin similarity score always plagiarism?

A high score indicates the presence of numerous text matches. Before deciding on what to change, you still have to check the matching passages, citations, quotes and types of sources.

Can a low score still be a problem?

A low score may still contain one significant uncited match. Be sure to look at the biggest and most specific highlights.

Does excluding the bibliography change the score?

Usually yes. By excluding bibliography, one can lower the score and make the report interpretation easier as reference lists can create many legitimate matches.