Your similarity report came back above the permitted limit and your submission date is weeks away. Do not start rewriting yet. The first step often removes several percentage points without changing a single sentence, because a large share of inflated reports are generated without the exclusions the UGC Regulations actually specify.
Work through these seven steps in order. Skipping to step four is the mistake that wastes the most time.
Step 1: Confirm the report was generated with the correct exclusions
Clause 7 of the UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018 states that the similarity check shall exclude:
- All quoted work reproduced with all necessary permission and attribution.
- All references, bibliography, table of contents, preface and acknowledgements.
- All generic terms, laws, standard symbols and standard equations.
A note to the same clause also excludes common knowledge or coincidental terms up to fourteen consecutive words.
In a thesis with a substantial bibliography, the reference list alone can account for a large block of matched text. If your report included it, the number in front of you is not the figure the Regulations describe. Ask whoever generated the report to rerun it with bibliography and quoted matter excluded before you touch the manuscript.
Expected output: a corrected report, and often a materially lower figure.
Step 2: Read where the overlap sits, not just the number
Two theses at the same percentage can be in completely different amounts of trouble.
Overlap concentrated in your literature review, in standard method descriptions and in definitions is the ordinary kind and is fixable by rewriting. Overlap in your results, discussion or conclusions is a different and far more serious matter, because the Regulations state that the research work must be based on original ideas covering the abstract, summary, hypothesis, observations, results, conclusions and recommendations, and that these shall not have any similarities.
Open the report and write down which chapters the flagged passages fall in. That distribution determines your whole plan.
Step 3: Separate the sources you are matching against
Look at what each flagged passage matches. Three categories need different handling:
- Matches to sources you cited. Usually paraphrasing that stayed too close to the original. Rewrite properly and keep the citation.
- Matches to sources you did not cite. Either you have absorbed phrasing from reading without realising, or a citation is missing. Add the citation and rewrite.
- Matches to your own earlier work. Your published papers, your synopsis, or another submission. This is self-overlap and must be handled by citing your earlier work, not by concealment.
That third category surprises scholars constantly, particularly where a synopsis was reused verbatim into the thesis. Ask your department how it treats overlap between your own synopsis and your thesis, because practice differs.
Step 4: Rewrite by reconstructing, not by substituting words
This is where most scholars go wrong under time pressure. Replacing words with synonyms leaves the sentence architecture intact. Detection tools identify that pattern, and examiners recognise it immediately as an attempt to disguise rather than rewrite.
The method that works:
- Read the source passage, then close it.
- Write down, in your own words, what it established and why it matters to your argument.
- Write your sentence from that note, not from the source.
- Reopen the source only to check accuracy and to place the citation.
This produces genuinely different prose because you are writing from your understanding rather than editing someone else’s sentence. It is slower per paragraph and much faster overall, because you do it once.
Step 5: Convert close paraphrase into synthesis
Where several flagged passages summarise studies one at a time, the underlying problem is structural. A literature review organised study by study will always sit close to its sources, because each paragraph is essentially a restatement of one paper.
Reorganise those sections by theme instead. Write what several studies collectively establish, where they disagree, and what that means for your question. Synthesis across sources is structurally difficult to match against any single source, and it also produces a stronger chapter, which is why this step is worth doing even when similarity is not a problem.
Step 6: Handle quotations and standard material correctly
- Direct quotations must be marked properly and attributed. Correctly quoted material is excluded under Clause 7, but a thesis leaning heavily on quotation raises a separate question about original contribution. Where a quotation is not doing distinctive work, convert it to synthesis instead.
- Standard definitions, equations and legal provisions are excluded as generic terms, laws, standard symbols and standard equations. Do not distort a statutory provision or a standard equation to reduce a percentage; that introduces an error far worse than the match.
- Method descriptions for established procedures legitimately resemble other descriptions of the same procedure. Write them in your own words with your specific parameters, which differentiates them naturally.
Step 7: Recheck, then document what you did
Run the check again with correct exclusions once revision is complete. If you are still above the limit, repeat steps two to five on the remaining passages rather than making broad edits across the whole document.
Keep a short record of what you changed and why. If your institutional panel reviews the case, a scholar who can show a documented, principled revision is in a substantially better position than one who cannot explain what happened.
The thresholds and the consequences at each level are set out in full in the UGC plagiarism limit for a PhD thesis, including the separate schedules for students and for faculty.
What not to do
- Do not use text-spinning or paraphrasing tools to force the number down. They characteristically produce awkward or subtly wrong phrasing, and you still have to defend every sentence in the viva.
- Do not manipulate the file with hidden characters, white text, or image substitution. This is treated as deliberate misconduct rather than a similarity problem, and the consequences are far more serious.
- Do not delete your citations to reduce matches. Removing attribution converts a formatting problem into actual plagiarism.
- Do not quietly cut your literature review to shrink the denominator. It weakens the thesis and examiners notice a review that no longer supports the gap statement.
How much time does this take?
Realistically, steps one to three take a day. Rewriting depends on how many passages are flagged and where. A literature review requiring reorganisation is several days of focused work, not an evening.
If the honest assessment is that you cannot complete the revision properly before your date, speak to your supervisor now rather than submitting something rushed. Requesting a short extension is a normal academic conversation. Submitting a manuscript that fails the check is not, and the Regulations attach real consequences to it.
Preventing this on the next chapter
Almost every high similarity report traces back to the same origin: notes taken by copying sentences from sources, which later flowed into the draft. The fix is upstream. Take notes in your own words from the start, record the citation alongside, and never paste source sentences into your working document.
Tesify helps you draft chapters from your own notes and framework, so the prose is constructed rather than assembled from source text, and keeps citations attached as the chapter develops. It does not rewrite flagged passages to beat a detector, and it does not replace your judgement about what to cite. You remain the author and are responsible for the integrity of the work.
Draft your next chapter in Tesify
Related requirements at submission
Similarity is one of several things that converge at submission. What happens to your thesis afterwards is explained in what Shodhganga is, and if you are still at the document stage, how to write a PhD synopsis covers the consistency checks that also apply to the full thesis.
Frequently asked questions
Does the bibliography count towards my similarity percentage?
It should not. Clause 7 excludes references, bibliography, table of contents, preface and acknowledgements. If your report included them, request a corrected report before revising anything.
My report is 12%. Is that a problem?
Under the Regulations, above 10% falls into Level 1, which carries a requirement to submit a revised script. Your university may apply a stricter internal limit. Check the distribution too, since 12% concentrated in your results is more serious than 12% spread across your review.
Can I be penalised even if the percentage is low?
Yes. The Regulations require the abstract, summary, hypothesis, observations, results, conclusions and recommendations to be based on original ideas without similarities. A small match in exactly those sections is treated seriously regardless of the overall figure.
How different does a rewritten sentence need to be?
Different enough that it reflects your own construction of the idea rather than an edited version of the original. If you can map your sentence word by word onto the source, it is not rewritten. Write from a note, not from the source text.
Is it acceptable to reuse text from my own published paper?
Cite your own work as you would cite anyone else’s. Uncited reuse of your own text is treated as a form of plagiarism and detection tools flag it, since your paper is in the comparison database.
Will different tools give me different percentages?
Yes. Comparison databases and settings differ, so a free online checker will not predict your department’s official figure. Use a self-check to locate overlap, not to estimate the number.
Should I tell my supervisor the report was high?
Yes, promptly. Supervisors deal with this routinely and can advise on your institution’s process. Concealing it until submission removes every option that might have helped you.
Can I be failed for a single high report?
A single report normally triggers examination rather than an automatic outcome. The Regulations provide for institutional panels to assess cases, and consequences follow the level and the circumstances. That is precisely why a documented, principled revision matters.
Does translated text get detected?
Detection of translated material is improving and cannot be relied upon to fail. More importantly, translating someone else’s work without attribution is plagiarism regardless of whether a tool catches it.
How early should I run my first similarity check?
Chapter by chapter as you draft, not once at the end. Fixing one flagged chapter is manageable; discovering the problem across an entire thesis weeks before submission is what turns it into a crisis.