A software report came back at 34 percent and Nithya Raghavan spent a week convinced her doctorate was over. It was not. At that number the rule does not cancel anything. It requires her to submit a revised script within a fixed period. The panic came from not knowing where the lines are drawn, and from a supervisor who quoted a number that is nowhere in the regulation.
Your similarity percentage is not your university's opinion. Since 2018 it maps onto four named levels set by a central regulation, with the penalty for each level written down. Knowing which level you are in, and what the check was supposed to exclude before it produced that number, is most of the battle.
Where the lines fall. Up to 10 percent is Level 0 and carries no penalty. Above 10 and up to 40 is Level 1. Above 40 and up to 60 is Level 2. Above 60 is Level 3. Those four bands come from the University Grants Commission regulations on academic integrity notified in 2018.
These are the penalties the Institutional Academic Integrity Panel imposes for plagiarism in a submitted thesis or dissertation.
Two riders sit underneath the ladder and both matter.
Penalties for plagiarism in academic and research publications are not the same as for a thesis, and staff are treated more severely at the upper levels.
| Level | Similarity | Consequence for a publication |
|---|---|---|
| Level 0 | up to 10 percent | Minor similarities, no penalty |
| Level 1 | above 10 to 40 percent | Asked to withdraw the manuscript |
| Level 2 | above 40 to 60 percent | Withdraw the manuscript, lose one annual increment, and cannot supervise any new Master's, MPhil or PhD scholar for two years |
| Level 3 | above 60 percent | Withdraw the manuscript, lose two successive annual increments, and cannot supervise any new scholar for three years |
A repeated Level 3 offence takes it out of the regulation's own ladder and into service rules, where disciplinary action including suspension or termination becomes available to the institution.
This is the part most students never read, and it is where inflated scores usually come from. The regulation says the similarity check shall exclude the following.
There is one more exclusion in the note that follows, and it is the most useful sentence in the document for anyone arguing about a number. The check excludes common knowledge or coincidental terms up to fourteen consecutive words.
So before you accept a percentage, ask a plain question in writing: was the report generated with the bibliography, the table of contents and properly attributed quotations excluded? A raw score on a whole document, including the reference list, is not the figure the regulation is describing. Running the check again with the correct exclusions frequently moves a thesis down a level, and sometimes down to Level 0.
The same note sets the positive standard, and it is narrower than people assume. The research work has to be based on original ideas, and the parts that must carry no similarity are specifically the abstract, summary, hypothesis, observations, results, conclusions and recommendations.
That is a targeted list. A literature review that necessarily restates the field is not being judged the same way as your results and conclusions. If your similarity is concentrated in the review and in standard method descriptions, say so explicitly and show where it sits, because the regulation itself distinguishes between the two.
That last point is your strongest procedural protection. A penalty imposed without a hearing, or before your appeal is decided, is not what the regulation contemplates.
If the institution will not share its report, its panel constitution or the rules it is applying, and it is a public authority, that information can be sought under the Right to Information Act. The AI RTI Drafter will prepare the application, and The RTI Playbook explains how to phrase it so you get the record rather than a summary.
It falls in Level 1, which is above 10 and up to 40 percent. The prescribed consequence for a thesis is that you submit a revised script within a stipulated period not exceeding six months. Your registration is not cancelled and you are not debarred. Before accepting the figure, confirm the report excluded quoted matter, references, the bibliography, the table of contents, the preface and acknowledgements.
Up to 10 percent is Level 0, described as minor similarities with no penalty. That is the only band the regulation treats as carrying no consequence. It is a ceiling for penalties, not a target to aim at, and the parts of your work expected to carry no similarity at all are the abstract, summary, hypothesis, observations, results, conclusions and recommendations.
It should not. The regulation says the similarity check shall exclude all references, the bibliography, the table of contents, the preface and the acknowledgements, as well as properly attributed quotations and generic terms, laws, standard symbols and standard equations. A report that counted your reference list has not been generated the way the regulation describes, and that is a fair point to raise in writing.
The regulation deals with this directly. If plagiarism is proved on a date later than the award of the degree or credit, the degree or credit is put in abeyance for a period recommended by the Institutional Academic Integrity Panel and approved by the head of the institution. Note the language is abeyance for a recommended period, decided through the panel process rather than automatically.
You are punished at one level higher than the level you previously committed. If the highest level has already been reached, the punishment for that level applies again. For faculty publications a repeated Level 3 offence opens the door to disciplinary action including suspension or termination under the institution's service rules.
The regulation says penalties apply when plagiarism has been established without doubt, when all avenues of appeal have been exhausted, and when the individual has been given enough opportunity to defend themselves in a fair and transparent manner. If you have not been heard, or your appeal has not been decided, that sequence has not been followed and you should say so in writing.
No. For a thesis the student ladder runs from no penalty, to a revised script within six months, to a one year bar on resubmission, to cancellation of registration. For publications the consequences are withdrawal of the manuscript, loss of one or two annual increments, and a two or three year bar on supervising new research scholars, depending on the level.