What Does an Originality Score Actually Mean?

You run a plagiarism check and receive an originality score of 78%. Is that good? Bad? Grounds for revision? The number feels definitive — but interpreting it correctly requires understanding what the score actually measures.

An originality score represents the percentage of your text that does not match detected sources in the checker's database. It is a similarity inverse — not a grade on writing quality, argument strength, or citation accuracy.

What the Percentage Means

• 90–100%: Predominantly unique text with minimal or no detected matches
• 80–89%: Generally strong originality; review flagged passages individually
• 70–79%: Moderate overlap; likely needs citation review or paraphrasing revision
• Below 70%: Significant similarity detected; thorough revision recommended
• Below 50%: Major overlap with external sources; high plagiarism risk

These ranges are guidelines, not rules. Institutional thresholds vary — some universities flag anything below 85%, while others focus on the nature of matches rather than the aggregate score.

Common Misinterpretations

Properly cited quotes lower your score

Correctly formatted quotations will appear as matches in most checkers. A paper with many legitimate citations may score lower than one with no sources at all — which would actually be worse scholarship.

Bibliography entries inflate match counts

Reference lists contain text that appears across thousands of papers. Most professional checkers exclude bibliographies from scoring, but not all free tools do. Always check your tool's settings.

A 100% score does not guarantee integrity

Checkers compare against indexed web content and proprietary databases. Unpublished sources, print-only books, and paywalled journals may not appear — meaning plagiarized content from those sources would not reduce your score.

Learn about semantic plagiarism that checkers often miss

Conclusion

Treat originality scores as diagnostic starting points, not verdicts. Read the match report, evaluate each flagged passage in context, and revise where attribution is missing or paraphrasing is too close.