NIRF penalised bad research this year. The deeper fix is self-reported data

For a decade, the National Institutional Ranking Framework has been the closest thing India has to an official verdict on its colleges. Launched in 2015 by the education ministry, NIRF now sorts thousands of institutions into league tables that shape admissions, funding conversations and reputations. In 2025 it ranked 7,692 of them. It also, quietly, admitted one of its own flaws.

That admission is worth reading closely, because it points to a larger one the framework has not yet made.

What did NIRF actually fix in 2025?

The 2025 edition introduced negative marking for retracted research papers, and for citations of discredited work, for the first time in the framework’s history. It is a real change. It concedes, in the scoring itself, that counting publications rewarded quantity over quality, and that the incentive had consequences.

Those consequences are documented. Reporting in Nature India in 2025 placed India third in the world for life-science research retractions, and tied the surge to publish-or-perish pressure that ranking metrics helped create. When a scorecard pays for papers, papers arrive, and not all of them survive scrutiny. Penalising retractions is the framework trying to close a loophole it opened. Good. But it treats a symptom.

Why is self-reported data the real problem?

The deeper issue is where NIRF’s numbers come from. A large share of the data, on faculty, finances, student outcomes and more, is submitted by institutions about themselves, with limited independent auditing. That design choice was always the framework’s soft underbelly, and in 2025 it surfaced in court.

A public interest petition in the Madras High Court briefly stayed the 2025 rankings over allegations of data manipulation, before the list was ultimately released. Whatever the merits of that specific case, it dramatised the structural point: a ranking is only as trustworthy as its inputs, and inputs you do not verify can be dressed up. A retraction penalty polices one field on the form. It does nothing about the honesty of the rest of the form.

What should replace the current approach?

Not the league table itself. Ranked information helps students, and NIRF’s participation numbers show real demand for it. The fix is in the plumbing.

Three changes would matter more than another tweak to the weights. First, audit the high-stakes fields, financial and outcome data at least, the way accreditation bodies audit, with panels and document checks rather than trust. Second, publish the raw submissions so that rival institutions, journalists and researchers can catch discrepancies, turning transparency into a verification engine. Third, lean the research score toward citation quality and integrity signals over raw counts, extending the logic of the retraction penalty rather than bolting it on. None of this requires scrapping NIRF. It requires taking its own data as seriously as it asks students to.

Conclusion

NIRF’s 2025 retraction penalty is the framework catching up with a problem its metrics helped cause, and it deserves credit for that. But a scorecard that penalises bad research while trusting institutions to grade their own paperwork has fixed the visible fault and left the load-bearing one in place. The question for the next edition is not which parameter to reweight. It is whether the numbers can be believed at all, and that is answered by auditing and disclosure, not by another line in the methodology.