EditverseReading a similarity report: what the percentage does and does not mean
How to read a thesis similarity report by source and by section, what the UGC 2018 regulations say about levels and exclusions, and how to resolve each kind of match properly.
Key points
- A similarity percentage measures matching text. It is not a measure of plagiarism, and the software provider says plagiarism cannot be determined from the score alone.
- Read the report by source and by section before looking at the total. A 9% match to your own published paper and a 9% match to an uncited source are different findings.
- Whether a number is acceptable depends on the regulation your university applies. In India, the UGC regulations of 2018 treat similarity up to 10% as Level 0, minor similarities.
- The same regulations exclude quoted work with attribution, references, bibliography, table of contents, preface and acknowledgements, and generic terms, laws, standard symbols and standard equations.
- A match is resolved by attribution: quotation marks and a citation, a cited paraphrase written from understanding, or disclosure of your own earlier work.
What the percentage is
Similarity software compares a document with the databases its operator selects: web pages, published literature and earlier student submissions. The report highlights matching text and names the source of each match. The percentage is the share of the document that matched.
The provider's own guidance for instructors sets out the limits of the figure. A similarity report's function is to highlight matching content. Plagiarism cannot be determined by the score alone. A high score does not automatically mean plagiarism, since quotations and reference lists raise it, and a low score does not guarantee originality. Longer assignments, research papers and literature reviews with many quotations and citations tend to score higher. The judgement belongs to a person who reads the matches in context.3
That last point is the working rule for a thesis candidate. The report is a list of places to check. The check is whether each matched passage is attributed correctly.
Reading the report in four passes
| Pass | What you see | What it tends to indicate | What to do |
|---|---|---|---|
| 1. By source | One source with a large share | Your own article, an earlier draft stored in the database, or one paper relied on heavily | Identify it. If it is yours, follow the section on own work below. If it is another author's, check every matched passage for quotation or citation. |
| Many sources at 1% or less each | Standard phrases, terminology, names of instruments and tests | Scan them. Most need no change. | |
| A student paper as the source | An earlier upload of your own draft, or shared departmental templates | Ask the administrator how earlier uploads were stored and whether the match can be excluded under university rules. | |
| 2. By section | Matches concentrated in the references | Reference lists match by nature | Confirm whether the university's settings exclude the bibliography. |
| Matches in methods | Standard procedures described in standard words | Cite the source of the method. Rewrite where the wording follows one paper closely. | |
| Matches in results, discussion or conclusions | These sections should be your own account of your own data | Read each one closely. This is where attribution matters most. | |
| 3. By match type | Long continuous blocks versus scattered short strings | A block is a copied passage. Short strings are often coincidence | Blocks need quotation marks and a citation, or a real rewrite with a citation. |
| 4. By settings | The filters applied: quotes, bibliography, small matches, databases searched | Two reports on the same file can differ if settings differ | Record the settings with the report. Use the settings the university specifies. |
A worked reading
The breakdown below is illustrative and uses invented figures for a fictional MD thesis. It shows how a single total hides five different situations.
| Source group | Share | Where in the thesis | Reading | Resolution |
|---|---|---|---|---|
| Candidate's own published article from the thesis data | 9% | Methods, results | Own earlier work, not yet declared in the thesis | Cite the article, state in the thesis that the chapter draws on it, and follow the university's rule on including published work. |
| Reference list | 6% | Bibliography | Expected | Excluded where the regulation excludes references and bibliography. |
| Guideline definitions and diagnostic criteria | 4% | Introduction, methods | Reproduced wording, cited but not marked as quotation | Add quotation marks or set as a quoted table with the source and any permission needed. |
| Two papers describing the laboratory method | 3% | Methods | Method text following the source closely | Describe the procedure as performed in this study and cite the papers for the method. |
| Fourteen sources below 1% each | 5% | Throughout | Common phrases and terminology | No change after scanning. |
The five rows add to 27%. Only two of them, 7 percentage points in all, call for rewriting or quotation. One calls for disclosure. Two call for nothing beyond confirming the settings. A candidate who saw only "27%" could not have known where to start.
What the UGC 2018 regulations say
For Indian higher education institutions the governing text is the UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018, notified on 23 July 2018.1 They apply to theses and dissertations at master's and research level and to publications. Five provisions matter when reading a report.
Exclusions (clause 7). Similarity checks exclude all quoted work reproduced with the necessary permission or attribution; all references, bibliography, table of contents, preface and acknowledgements; and all generic terms, laws, standard symbols and standard equations.
Core content (note to clause 7). The research work is to be based on original ideas, which the note lists as the abstract, summary, hypothesis, observations, results, conclusions and recommendations, and these are not to have similarities. The note excludes common knowledge or coincidental terms up to fourteen consecutive words.
Levels (clause 8).
| Level | Similarity | Provision for a thesis or dissertation |
|---|---|---|
| Level 0 | Up to 10% | Minor similarities, no penalty |
| Level 1 | Above 10% to 40% | A revised script is to be submitted within a stipulated period not exceeding six months |
| Level 2 | Above 40% to 60% | The student is debarred from submitting a revised script for one year |
| Level 3 | Above 60% | Registration for the programme is cancelled |
Undertaking and certificate (clause 6). Every student submitting a thesis or dissertation gives an undertaking that the document is their original work and has been checked through a detection tool approved by the institution. Each supervisor submits a certificate that the work is plagiarism free. The UGC PhD Regulations, 2022, repeat the requirement for doctoral theses: an undertaking from the scholar and a certificate from the supervisor attesting to originality.2
Institutional policy (clause 6). Each institution is to adopt its own plagiarism policy, approved by its statutory bodies, and place it on its website. Read that policy as well. It names the software, the settings and the procedure, and it is what your departmental panel applies.
Outside India, or in an institution governed by another regulator, the number to use is the one in your own university's regulation. The software provider sets no threshold. If your university publishes none, the report is read qualitatively, source by source, as described above.
Resolving each kind of match
| Match | Resolution |
|---|---|
| Exact wording from a source | Keep it only if the exact words matter, such as a definition or criterion. Put it in quotation marks or a clearly marked quoted block and cite it. Otherwise restate the point in your own words and cite. |
| Close paraphrase | Close the source, write the point from your understanding of it in the context of your study, then cite. A cited paraphrase is your sentence about their finding. |
| Your own published article | Cite it and say plainly in the thesis which chapters draw on it. Publishers follow COPE's definition of text recycling as reuse of one's own previous writing without transparency or appropriate reference, so the remedy is transparency.4 Check the publisher's permission terms and your university's rule on including published papers. |
| Standard methods | Cite the original description and write what was done in your study: your equipment, settings, timings and deviations. Wiley's ethics guidelines note that reuse of text in a methods section may be more acceptable than in a discussion, provided the source is cited.4 |
| Figures and tables from other work | Cite the source in the legend and obtain written permission from the copyright holder unless the material is in the public domain.6 |
| References and common phrases | No change. Confirm that the report settings follow the regulation. |
Changing words without adding attribution does not resolve a match, because the report is pointing at a missing attribution, not at the wording. A passage that owes its content to a source needs the citation whatever its phrasing, and a passage that is quoted and cited is acceptable however exactly it matches. Working on attribution is also the faster route: it improves the thesis and the report together.
A sequence before submission
- Obtain the university's plagiarism policy and note the software, the exclusions and who runs the official check.
- Ask how drafts are stored, so that an early upload does not become a source for the final one.
- Run a check on the complete draft early enough to revise, using the university's settings.
- Work through the report source by source, largest first, recording what was done for each.
- Verify every reference against the original paper or a bibliographic database, as the ICMJE Recommendations, updated January 2026, advise for manuscripts.6
- Declare any published article, conference abstract or preprint that the thesis draws on. The ICMJE approach to overlapping work rests on disclosure.5
- Keep the final report and its settings with your submission papers, and sign the undertaking on that basis.
What editing contributes
An editor reads for attribution as part of reading for accuracy. In Editverse Medical's MD, MS and master's thesis editing, the candidate supplies the thesis file, the university guidelines and the supervisor's comments, and may add the similarity report. The editor returns tracked and clean files formatted to the university template, with a reference check, and marks passages that need quotation marks, a citation or the candidate's own restatement. The restating is the candidate's work, because it depends on what the candidate understands the source to say. For doctoral work the equivalent service is doctoral dissertation editing. All thesis work follows the candidate's university rules and the reference style the university names. See medical dissertations for the full list.
Frequently asked questions
Is there a universally acceptable similarity percentage for a thesis?
No. The software provider sets none. The figure comes from the regulation your university applies. Under the UGC Regulations, 2018, similarity up to 10% is Level 0, described as minor similarities with no penalty.
Is excluding the bibliography and quotations legitimate?
Yes, where the governing regulation provides for it. Clause 7 of the UGC Regulations, 2018, excludes references, bibliography, table of contents, preface and acknowledgements, and quoted work reproduced with attribution. Use the exclusions your university specifies and record them with the report.
My own published paper is the largest source. What should I do?
Cite the paper, state in the thesis which chapters are based on it, and check both the publisher's reuse terms and your university's rule on including published work. Tell your supervisor and the panel in advance so the match is expected.
Can a report with a low percentage still need work?
Yes. The provider's guidance notes that a low score does not guarantee originality. One uncited passage in the results or conclusions matters more than a larger share of matched references. Read the matches, not only the total.
References
- University Grants Commission. UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018. Gazette of India, Part III, Section 4; 23 July 2018. Official PDF
- University Grants Commission. UGC (Minimum Standards and Procedure for Award of Ph.D. Degree) Regulations, 2022. Gazette of India, Part III, Section 4; 7 November 2022. Official PDF
- Turnitin. Turnitin and plagiarism. Turnitin Guides. Read 10 November 2026. guides.turnitin.com
- Wiley. Best Practice Guidelines on Publishing Ethics: plagiarism, duplicate and redundant publication, text recycling. Read 10 November 2026. authorservices.wiley.com
- International Committee of Medical Journal Editors. Overlapping publications. In: Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals. Updated January 2026. icmje.org
- International Committee of Medical Journal Editors. Preparing a manuscript for submission to a medical journal. In: Recommendations. Updated January 2026. icmje.org