Why Does My Turnitin Similarity Score Change Between Submissions?

You submitted your thesis draft to Turnitin and got a 12% similarity score. Made a few corrections and resubmitted — now it shows 19%. You did not add new content. Nothing changed. So what exactly is going on?

Score changes between resubmissions catch Indian PhD students off guard every year. This article explains why Turnitin similarity percentages shift between runs, what a change actually signals under India’s UGC compliance thresholds, and what you should do when your score has climbed without obvious reason.

Table of Contents

• Why Your Turnitin Score Changes Between Submissions
• How Serious Is a Score Change? The UGC 2018 Perspective
• How to Fix a Higher-Than-Expected Score: Step-by-Step
• How to Prevent Unexpected Score Changes Next Time
• Conclusion

Why Your Turnitin Score Changes Between Submissions

Turnitin’s similarity score is not a fixed, objective measurement of your document. Think of it less like a test result and more like a search result — a comparison against whatever happens to be in Turnitin’s live database at the exact moment your file is processed. That database grows every day: new theses, journal articles, student papers, and web pages are indexed continuously. Two identical documents submitted a week apart will often return different scores, sometimes by several percentage points.

Five things typically cause scores to shift between resubmissions:

• Repository timing: Turnitin indexes tens of thousands of new documents daily. A source that was not in the database when you first submitted — a newly published paper on a similar topic, a recently uploaded thesis from another Indian university — may now be indexed and matching against your text.
• Self-matching from a previous draft: If Turnitin stores your earlier submission in its institutional repository, your resubmission will match against your own previous work. This is probably the most common cause of an unexpected score increase in Indian university contexts. You did not add plagiarism — you added yourself. (This is where most PhD students panic unnecessarily, by the way.)
• The 24-hour resubmission delay: From your 4th resubmission onwards, Turnitin enforces a mandatory 24-hour processing window. The database continues updating during that time. So the score you eventually see can differ from what you expected — and many students discover this only after the window has closed.
• Changed exclusion settings: Turnitin lets instructors and supervisors exclude quoted material, bibliography entries, and small word-count blocks from the comparison. If those settings were adjusted between your first and second submission, the system is now comparing different content — even on the same document.
• Updated internet index: Turnitin crawls publicly available web content. A new blog post, conference abstract, or open-access article published between your two submissions — if it contains language similar to yours — will now show up as a flagged match.

The core point: Turnitin compares your document against a moving target. The score is a snapshot of one specific moment in time, not a permanent verdict about your thesis. Understanding this is the first step to interpreting score changes calmly and acting on the right information.

How Serious Is a Score Change? The Indian University and UGC 2018 Perspective

A score change only matters if it pushes you across a compliance threshold. The UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018 define four levels of similarity that Indian universities must apply when evaluating PhD theses and research papers.

The four levels work like this:

• Level 0 (below 10%): No plagiarism detected. Submission accepted without further action.
• Level A (10–40%): Minor plagiarism. Revise and resubmit; no punitive action on the first instance.
• Level B (40–60%): Moderate plagiarism. Thesis suspended; student required to resubmit after six months to one year.
• Level C (above 60%): Major plagiarism. Registration may be cancelled; student barred from resubmission for two years.

A score jump from 8% to 12% feels alarming. But both numbers sit within the same Level 0–A zone — no compliance consequence has changed. A shift from 38% to 42% is an entirely different matter: that crosses from Level A into Level B, and it warrants urgent attention before your final submission date.

There is one more thing most students miss: only the score on your final official submission counts. Practice runs, iterative drafts, supervisor-review submissions — those exist for your benefit. Indian universities assess the similarity percentage on the document submitted through the official institutional channel, at the time of final thesis upload, under your guide’s oversight. A higher score on a practice run is information, not a verdict. Most supervisors across DU, Mumbai University, and affiliated institutions will tell you the same.

How to Fix a Higher-Than-Expected Score: Step-by-Step

If your similarity score has climbed unexpectedly, work through these steps in order. Do not panic and resubmit immediately — rapid resubmissions compound the self-matching problem and push you through the 24-hour delay trigger faster.

1. Download and read the full Similarity Report — not just the percentage. Open the report and examine which sources are flagged and what percentage each one contributes. A large portion of the increase may be your own previous draft appearing as a self-match. Self-match entries are not external plagiarism, and your supervisor can often request that the institution’s Turnitin administrator exclude them from the comparison.
2. Identify the sources contributing more than 2–3% each. Sort flagged sources by their percentage contribution. The long tail of 0.1–0.5% matches from common academic phrases is not worth rewriting — those phrases appear in thousands of papers. Focus your energy on sources with meaningful shares.
3. Rewrite flagged passages in your own words — properly. Close the source document. Write a fresh explanation of the idea in your own voice, then cite the original. Do not rearrange words or use synonym-swap tools — Turnitin’s semantic analysis and AI writing detection flag these patterns. Genuine paraphrasing means understanding an idea and restating it from scratch.
4. Ask your supervisor to enable bibliography and quote exclusions. If your institution’s Turnitin setup does not already exclude bibliographies and direct quotations, ask your guide to enable these options before your next submission. Unexcluded bibliographies alone can add 3–8% to your similarity score without representing any academic misconduct.
5. Wait at least 24 hours before resubmitting. This lets the database state settle and reduces the chance of adding another self-match layer. If you are on your 4th or a later resubmission, the 24-hour delay is mandatory regardless — plan your revision timeline accordingly.

For a detailed walkthrough of reducing your similarity percentage across each chapter, see our complete guide on how to reduce plagiarism in a PhD thesis.

How to Prevent Unexpected Score Changes Next Time

Once you understand why scores shift, the prevention habits become fairly obvious. These are what separate students who face a deadline-day surprise from those who submit with confidence.

• Run your final pre-submission check at least one week before your deadline. This gives you room to act on any unexpected increase rather than scrambling in the last 24 hours. It also means the database state you test against will closely reflect what you will face on actual submission day.
• Ask your supervisor to configure repository exclusions from your first submission. Most Indian university Turnitin setups allow supervisors to prevent student drafts from being added to the institutional repository during the review phase. Set this up before your first practice run and your drafts will not stack up as self-match sources against later submissions.
• Keep a submission log with dates and scores. Record the date, approximate word count, and similarity score for every run. When your score jumps between versions, this log helps you isolate what changed — a new external source, a self-match from a specific earlier draft, or an exclusion setting that was removed.
• Build citation discipline into your writing from chapter one. Students who write entire chapters and then try to reduce similarity afterwards are in a far harder position than those who paraphrase and cite correctly as they go. Write first, then check. Writing to a low score produces garbled, over-paraphrased text that reads poorly and may still get flagged.
• Never paste raw source text into your working notes file. A surprisingly common accidental similarity source: pasting source text into a personal notes document for reference, then unknowingly incorporating that language into the thesis draft. Write your own paraphrases in your notes — never the original source text.

If you are starting a new chapter or literature review section, consider running a short test submission on just that section before appending it to your full document. Catching a new match at section level — when revision is still manageable — is far better than discovering it at the full-thesis stage with days to go.

For a full explanation of how to read your report and what each colour-coded section means, see our guide on how to read a Turnitin Similarity Report.

Conclusion

A changing Turnitin similarity score is almost never a sign that your thesis has a fundamental plagiarism problem. In most cases it reflects something more mundane: how Turnitin’s live comparison system actually works. New documents get indexed. Previous drafts create self-matches. Exclusion settings shift what the system measures. None of that is your fault, and none of it is permanent.

What matters is not whether your score changed, but where your final official submission lands relative to the UGC 2018 thresholds — particularly the Level A (10–40%) and Level B (40–60%) boundaries. Use the steps in this article to understand the actual source of any increase, address the matches that genuinely need attention, and build the pre-submission habits that protect you from last-minute surprises.

Key Takeaways

• Turnitin compares against a live database — the same document can return different scores days or weeks apart.
• The five main causes of score changes are: new sources indexed, self-matching from previous drafts, the 24-hour resubmission delay, changed exclusion settings, and updated internet content.
• Under UGC Regulations 2018, compliance thresholds are 10%, 40%, and 60% — only a change that crosses one of these boundaries is a genuine compliance concern.
• Download the full Similarity Report and read the source breakdown before drawing conclusions from the percentage alone.
• Run your final check one week early and keep a submission log to avoid deadline-day surprises.