Estimate a Similarity Score yourself.

A Similarity Score describes text overlap.

The value shows how much submitted text overlaps with content in the comparison collection used by the service. A match may be a correctly identified quotation, a reference entry, a standard phrase or problematic copying.

Only Turnitin can generate the actual Turnitin value when the paper is processed inside the school's configured system with its databases and settings. A self-calculated result remains a documented approximation.

There is no universal “permitted” percentage. Independent work, transparent acknowledgement, current school rules and human review of every match are what matter.

Approximation calculator: S = matched words ÷ total words × 100

Worked example

For 3,200 total words and 416 uniquely counted matched words:

416 ÷ 3,200 × 100 = 13%

“Unique” matters: if the same 20 words match three sources, they contribute 20 words—not 60—to the overall approximation.

A reproducible self-check method.

1. Fix the version: record filename, date, language and total word count.
2. Define scope: decide whether the title page, footnotes, appendices and bibliography are included.
3. Name the comparison set: list sources and online services used. A personal source folder cannot reproduce Turnitin's databases.
4. Mark matches: identify identical or highly similar word sequences and define a minimum length.
5. Merge overlaps: combine overlapping spans and count each word position once.
6. Calculate gross and adjusted values: report all matches and a second result excluding correctly formatted quotations and bibliography.
7. Review every match: decide whether it is a correct quotation, valid paraphrase, common knowledge or a passage needing revision and attribution.

Why the actual Turnitin result can differ.

• Depending on the school, Turnitin compares internet content, publications and previously submitted work.
• The school can configure databases and exclusions for bibliography, quotations or small matches.
• An earlier submission of the same draft may match if it was stored in a repository.
• Other providers use different collections, preprocessing and matching algorithms.
• The displayed score may change when legitimate matches are excluded.

Describe a self-calculated value as an “approximation under the documented comparison procedure”, never as an exact predicted Turnitin score.