How to Read a Turnitin Similarity Report — Full 2026 Guide

Most students open their Turnitin report, see a percentage in red or orange, and panic. Understandable — but almost always unnecessary. The Turnitin Similarity Index is not a plagiarism verdict. It’s a matching percentage, and reading it correctly takes about five minutes once you know what to look at, what to exclude, and what threshold your institution actually cares about.

Below, we cover every part of the report interface — colour bands, the match sources panel, the AI writing detection score, and the thresholds that govern Indian PhD researchers under UGC regulations. What you need, without the filler.

Key Takeaways

• The Similarity Index shows what percentage of your text matches sources in Turnitin’s database — it does not label anything as plagiarised.
• Always apply standard exclusions (bibliography, quoted text, small matches) before reading your score.
• Colour bands range from blue (0%) to red (75–100%) — but colour alone doesn’t determine an outcome.
• Under UGC Plagiarism Regulations 2018, the acceptable threshold for PhD theses is ≤10% after exclusions (Level 0).
• A separate AI Writing Detection panel gives a 0–100% score for AI-generated content — this is independent of the similarity score.
• Matches to student papers in the Turnitin database are treated more seriously than matches to websites.

What Is the Similarity Index and What Does It Actually Mean?

The Similarity Index is a percentage representing how much of your submitted text matches content already in Turnitin’s database — websites, academic publications, and previously submitted student papers. The higher the percentage, the more your text overlaps with existing sources. That’s it. The report says nothing about why those matches exist.

Here’s the point most people miss: Turnitin does not detect plagiarism. It detects textual similarity. A passage could match a source because you copied it without attribution, because you cited it correctly as a quote, or simply because certain phrases are common in your field. The software cannot distinguish between these. Your supervisor or institution makes that call.

Think of the Similarity Index as a highlighting tool, not a grading tool. It shows which parts of your document share text with other sources — so a human reader can decide whether those matches are acceptable. A score of 18% could be perfectly fine. A score of 8% could still contain a problematic unattributed passage. The percentage is a starting point, not a conclusion.

How to Read the Colour Bands on Your Turnitin Report

When you open your Turnitin report, the first thing you’ll notice is a coloured percentage in the top-right corner of the document viewer. This colour is a quick visual indicator of the score range — nothing more. Here’s what each band means and how universities typically respond.

These colour bands are display conventions — not pass or fail thresholds. Most universities in India apply the UGC 2018 framework rather than relying solely on the colour shown. A green score is not automatically safe, and a yellow score is not automatically a problem.

One mistake we see constantly: students read the colour before applying exclusions. If you haven’t removed your reference list and quoted material yet, the displayed percentage can be 8 to 12 points higher than your actual working score. Configure exclusions first. Then read the colour.

How to Apply Standard Exclusions Before Reading Your Score

Before drawing any conclusions from your Similarity Index, apply the three standard exclusions that Turnitin allows. They remove content from the calculation that would otherwise inflate your score artificially — and in most Indian universities, examiners expect to see the post-exclusion figure, not the raw one.

The three exclusions are:

• Bibliography / reference list — Your reference section will always match sources in the database because those are titles and authors that exist elsewhere. Excluding it removes this automatic inflation.
• Quoted material — Text you have properly enclosed in quotation marks can be excluded. If you’re legitimately quoting a source, that match shouldn’t count against you.
• Small matches — Turnitin lets you exclude matches below a word threshold, typically fewer than 8–10 words. This removes common phrases, methodology boilerplate, and field-standard terminology that would otherwise register as matches.

To apply exclusions, look for the filter and settings icon — the funnel symbol — in the right-hand sidebar of the Similarity Report. Toggle each exclusion on or off and watch the score update in real time. The figure you see after applying all three is the one to compare against UGC thresholds.

What Do the Match Sources Mean — and Which Ones Matter?

On the right side of the Turnitin report viewer, you’ll find the Match Overview panel. This lists every source your document matched against, ranked by how much of your text overlaps with each one. Three source categories exist — and they carry very different weight.

• Internet sources — Websites, blogs, news articles, and publicly accessible pages. Matches here may indicate copy-pasting, but could equally reflect common phrasing or shared factual statements.
• Publications — Journal articles, conference papers, and academic databases. If your work matches a publication you cited correctly, that’s expected and usually fine.
• Student papers — Submissions made by other students to Turnitin globally, with identifying details removed. A match here is the most serious category you can encounter.

A match to a student paper immediately raises the question of collusion — whether you accessed that student’s work, or they accessed yours. Even when it’s genuinely coincidental, institutions treat student-paper matches with far more scrutiny than internet or publication matches. If you see a notable percentage matched to a student paper, don’t wait — identify which section triggered it and find out why. (This is the one scenario where a seemingly low overall score can still lead to a very uncomfortable conversation with your research supervisor.)

When you click any match in the panel, the corresponding text in your document gets highlighted and the source appears on the right. Review it: is this a legitimate citation, standard field language, or something that needs rewriting? That source-by-source review is the actual purpose of the Turnitin report — not the headline number.

What Is the AI Writing Detection Score?

Since 2023, Turnitin has included an AI Writing Detection feature — a distinct panel within the same report interface. It gives a percentage from 0–100%, representing the proportion of your submitted text the system estimates was likely written by an AI tool such as ChatGPT, Gemini, or similar.

The AI detection score and the similarity score are completely independent. Low similarity, high AI score. High similarity, low AI score. Both low, both high. They measure different things — similarity measures textual overlap with known sources, while AI detection flags statistical patterns in the language itself.

Know the limits of this feature. Turnitin’s own documentation acknowledges its AI detection can produce false positives, particularly for non-native English writers whose phrasing may inadvertently resemble AI output, or for highly technical or formulaic writing common in Indian engineering and management theses. A high AI score is a flag for review, not a finding of misconduct.

If your AI score is high and you did not use AI tools, keep a record of your writing process where possible — drafts, notes, supervisor emails — to support your position if questioned. Most Indian universities are still formalising their AI detection policies, so check your institution’s current guidance directly rather than assuming any particular threshold applies.

What Is an Acceptable Turnitin Score for Indian PhD Students?

For PhD researchers in India, the benchmark comes from the UGC Plagiarism Regulations 2018. These define four levels of similarity — applied to the score after standard exclusions. Universities affiliated with UGC are expected to enforce these thresholds for theses and dissertations.

• Level 0 — ≤10%: No plagiarism detected. The thesis proceeds normally.
• Level 1 — 10% to 40%: Minor similarity. Typically requires revision and resubmission within a specified period.
• Level 2 — 40% to 60%: Moderate similarity. Resubmission after a longer period; academic sanctions may apply.
• Level 3 — above 60%: Serious similarity. The thesis may be rejected and disciplinary proceedings may follow.

The critical phrase across all four levels: after applying standard exclusions. If your raw score before exclusions is 22% but drops to 9% after removing the bibliography, quoted text, and small matches, you’re at Level 0. Always confirm with your institution which exclusion settings they require before treating any figure as final.

Also worth knowing: individual universities often set stricter internal thresholds than the UGC minimum. Some institutions — particularly in the IIT and central university system — require ≤7% or even ≤5% for PhD theses. The UGC regulation sets a floor, not a ceiling. Ask your research supervisor or graduate school coordinator what your university’s internal policy actually says before concluding you’re in the clear.

Frequently Asked Questions

Does a high Turnitin score mean I’ve plagiarised?

No. A high Similarity Index means a high proportion of your text matches sources in Turnitin’s database. Whether any of those matches constitute plagiarism is a judgement your institution makes after reviewing the actual matched passages. Properly cited quotations, common academic phrases, and standard methodology descriptions can all contribute to the score without being plagiarism.

Can I submit my own previous work and get a high similarity score?

Yes — if you’ve submitted work to Turnitin before, at the same or a different institution, it may already be in the database. Reusing your own previously submitted material without disclosure is called self-plagiarism or text recycling, and most Indian universities treat it as a genuine academic integrity issue, not a technicality. Cite your own prior published or submitted work if you’re drawing from it.

What should I do if my score is much higher than I expected?

Start with the three standard exclusions — bibliography, quoted text, small matches — and see whether the score drops to an acceptable range. Then open the Match Overview panel and go through each source match. Find which sections are driving the score and decide whether those matches are properly cited, common phrasing, or passages that need rewriting. The headline number rarely tells you what you need to know; the individual matches do.

Does Turnitin check against all published journal articles?

No. Turnitin has licensing agreements with many major academic databases and publishers, but not all. Paywalled or regional journals — including a number of Indian journals — may not be in its database at all. A low similarity score means no overlap was found in the sources Turnitin can access. It doesn’t guarantee your work has no overlap with all existing literature.

Is the AI Writing Detection score shown to my supervisor automatically?

It depends on how your institution has configured its Turnitin account. In many setups, both the Similarity Report and the AI Writing Detection score are visible to instructors and supervisors when they open your submission. Some institutions have disabled the AI detection feature or restricted who can see it. Check your university’s Turnitin policy, or ask your supervisor directly — they’ll know exactly what appears on their end.

Conclusion

Reading a Turnitin Similarity Report correctly means going past the headline percentage. The colour band gives you a quick orientation — but the real analysis starts when you apply standard exclusions, work through the Match Overview panel source by source, and compare your adjusted score against your institution’s thresholds or the UGC Plagiarism Regulations 2018.

The report is a review tool, not a verdict. A match to a publication you cited correctly is not a problem. A match to a student paper needs immediate investigation. The AI Writing Detection score is a separate, independent measure with its own limitations — and for most Indian universities right now, no fixed formal threshold. Treat each component on its own terms.

If after applying exclusions and reviewing your matches you’re still seeing a score that concerns you, see our guide on what to do if your similarity score is high for practical next steps.