Before an MD, MS, DNB or MSc Nursing thesis is accepted, most Indian universities now ask for a plagiarism (similarity) report from the library. For many residents this is the most stressful step of submission: a number appears on the screen, and nobody explains what it means or what to do about it.
This guide explains what the similarity index actually measures, which tools are used, what the UGC 2018 regulations say about similarity levels, and how to bring your number down honestly, without the shortcuts that can turn a similarity problem into a misconduct case.
1What the similarity index actually means
A similarity index is the share of your text that the software found matching text in its database: journal articles, websites, books, and earlier submitted theses. If a report shows 18%, it means roughly that proportion of your document overlaps with something the tool has seen before.
Three points are often misunderstood:
- Similarity is not the same as plagiarism. A correctly quoted and cited definition, a standard drug name or a reference list will all "match". The software finds overlap; a human decides whether it is acceptable.
- Zero is not the goal. Any thesis that cites literature and uses standard medical terms will show some similarity. Chasing 0% usually leads to unnatural writing.
- The pattern matters as much as the number. Many tiny matches scattered across dozens of sources are usually standard phrases. A single source contributing a large block means copied paragraphs, and examiners notice it even when the total looks low.
2Tools used in India: Turnitin, iThenticate and DrillBit
| Tool | Where you will meet it |
|---|---|
| Turnitin | Widely used by universities and colleges to check student work, including theses and dissertations. |
| iThenticate | From the same company as Turnitin and aimed at researchers and publishers; many journals use it to screen submitted manuscripts. |
| DrillBit | An Indian plagiarism detection software made available to many Indian universities under the Government of India's ShodhShuddhi initiative, run through INFLIBNET. If your university library issues the official report, there is a good chance it uses this tool. |
Each tool compares your text with a different database and uses different matching settings, so the same chapter can produce different percentages on different tools. Do not compare a free online checker's number with your university's official report. The number that counts is the one from the tool and settings your university uses.
Uploading an unpublished thesis to unknown websites can expose your work and your patients' context to third parties. Some institutional tools may also store submitted files, so a draft checked in a stored mode can later match your own final thesis. Ask your library how draft checks are handled before you submit chapters repeatedly.
3What the UGC 2018 regulations say
The UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018 set out levels of similarity for academic and research work in higher educational institutions:
| Level | Similarity | Broad meaning |
|---|---|---|
| Level 0 | Up to 10% | Minor similarities |
| Level 1 | Above 10% to 40% | Consequences apply |
| Level 2 | Above 40% to 60% | Stricter consequences |
| Level 3 | Above 60% | Strictest consequences |
For students, the regulations attach progressively heavier consequences to the higher levels, ranging from no penalty for minor similarity to revision and resubmission and, at the highest levels, more severe action. Read the actual text of the regulations, or your university's version of them, for the exact consequences rather than relying on summaries.
Material that may be excluded
The regulations also list material that is excluded from similarity, including, in summary, properly quoted and attributed work, references and bibliography, the table of contents and acknowledgements, generic terms, standard symbols and equations, and common knowledge. In practice, whether your library applies filters such as "exclude quotes" or "exclude bibliography" depends on its own policy.
Health-science universities, and bodies such as the NMC and NBE for their programmes, may set their own thresholds, procedures and certificate formats for theses. Many institutions expect a lower similarity than the outer limits of Level 1. Ask your library or PG cell for the written rule, the tool used, and which exclusions are applied, before you finalise your chapters.
4How to read your similarity report
Open the full report, not just the summary percentage, and go through the source list from the top.
- Large blocks from one source: copied or lightly edited paragraphs. These are your first priority.
- Your own earlier work: your approved synopsis or a paper you published from the same data. Reusing your own published text without acknowledgement can also count as a problem (often called self-plagiarism or text recycling); cite your published paper and rewrite.
- Methods text: descriptions of standard procedures copied from other studies or kit inserts.
- Standard phrases and definitions: diagnostic criteria, scale names, drug names, institutional title-page wording. These are usually acceptable if quoted or cited.
- References: if the bibliography is not excluded, references will match. Ask whether exclusion is permitted rather than changing reference formats.
5How to reduce similarity ethically
The honest way to lower similarity is also the way that makes your thesis better: write in your own words, from understanding, and cite properly.
Paraphrase from understanding, not word by word
Read the source, close it, write the idea from your notes in your own sentence structure, then check that the meaning is accurate and add the citation. Changing a few words while keeping the original sentence structure is still close copying.
An illustrative example (sentences written for this post):
| Version | Text |
|---|---|
| Source sentence | Early identification of high-risk patients allows clinicians to start preventive measures before complications develop. |
| Poor paraphrase | Early detection of high-risk patients lets doctors begin preventive steps before complications appear. |
| Better paraphrase | Preventive treatment works best when it begins before complications occur, which depends on recognising at-risk patients early [citation]. |
Quote sparingly, and quote correctly
Use quotation marks and a citation for exact definitions, criteria or a statement whose wording matters. Do not quote whole paragraphs to escape paraphrasing.
Synthesise in the review of literature
Group studies by theme and compare them, rather than writing one paragraph per paper that retells its abstract. Synthesis naturally produces original text. See our guide to writing the review of literature.
Write your own methods
Describe what you did: your setting, your sample, your timings, your instruments. If you follow a published technique, summarise it briefly in your own words and cite it.
Make the discussion your interpretation
The discussion should compare your findings with published studies and explain agreements and differences in your own reasoning. Copying other authors' discussion paragraphs is a common source of large single-source matches. State your result, then the comparison, then your explanation, citing each study you mention; our guide to writing the discussion chapter shows a structure.
Cite consistently
Uncited ideas are plagiarism even if paraphrased. Use a reference manager such as Zotero, Mendeley or EndNote and follow one style throughout; our Vancouver referencing guide shows how.
If your library allows draft checks, check the review of literature and methods as soon as they are written. Fixing one chapter at a time is far easier than rewriting the whole thesis two weeks before the deadline.
6What not to do
Some "tricks" circulate among PG students every year. They are not shortcuts; they are misconduct, and they can make your situation much worse.
| Don't | Why |
|---|---|
| Run text through synonym spinners or paraphrasing tools | Produces awkward, often wrong medical language ("cardiac arrest" becoming nonsense); examiners notice, and the ideas are still uncredited. |
| Use AI rewriters to disguise copied text | The source is still not credited, AI output can introduce factual errors and invented references, and some tools now flag likely AI-generated text. University policies on AI use vary, so check and disclose as required. |
| Character tricks: look-alike letters from other alphabets, hidden or white text, inserted invisible characters | Detection tools and reviewers can flag them, and deliberate manipulation of the check is treated as dishonesty in itself. |
| Paste text as images to hide it | Reviewers can see it, and it is still copying. |
| Copy a senior's thesis and change the data | Earlier theses may be in the tool's database, and copying them is plagiarism regardless. |
Integrity in writing is part of research ethics, just like consent and confidentiality; see our post on ethical standards in medical research.
7Checklist before the final check
- Confirm your university's threshold, tool, exclusions and certificate format in writing.
- Every borrowed idea has a citation; every exact phrase has quotation marks.
- Methods describe your own study in your own words.
- Your synopsis or published paper is cited where you reuse its content.
- No spun, machine-rewritten or manipulated text anywhere.
- Keep a copy of the final report, and have your guide review it before you sign any declaration.
8Need help with a high similarity report?
A high number is fixable when you deal with it early and honestly. If you would like help understanding your report, restructuring your review of literature, or rewriting sections in your own voice with proper citations, the PubMedico team can guide you; message us on WhatsApp +91 78787 33004.
❓ Frequently Asked Questions
Quick answers to common questions on this topic
There is no single number for every university. The UGC 2018 regulations treat similarity up to 10% as minor (Level 0), but health-science universities and bodies such as the NMC or NBE may set their own limits and procedures. Ask your university library or PG cell for the written threshold, the software used, and which exclusions are applied.
They will match in the software, but the UGC 2018 regulations list properly quoted and attributed material, references and bibliography among the items that are excluded. Whether your library actually switches on filters such as exclude quotes or exclude bibliography depends on its own policy, so confirm before you worry about matches in your reference list.
Each tool compares your text against a different database of sources and uses different matching settings. A free checker may report a very different figure from Turnitin or DrillBit. Only the report generated by your university's tool, with its settings, is used for submission, so treat other numbers only as rough guides.
It is risky and usually counts as misconduct when used to disguise copied text, because the original ideas remain uncredited. AI rewriters can also distort medical meaning or add invented references, and some detection tools now flag likely AI-generated text. Write in your own words, cite your sources, and follow your university's policy on AI use.
It can, if the paper is in the tool's database and your thesis reuses its text. Cite your own paper where you draw on it, rewrite overlapping passages, and tell your guide and library in advance. Some universities have specific rules about publishing before thesis submission, so check them.
📍 Medical Thesis Help in Your City
PubMedico works with MD, MS, DNB, MDS, PhD and MSc Nursing scholars across India. Find thesis writing, statistical analysis and publication support near your college:
…and 38 more cities across India. See all locations on the home page →