Science should move past ‘top 2%’ label

1. At a Glance

• Core issue: Indian academia increasingly treats inclusion in the "top 2% scientists' database" as proof of excellence. Prof. Mamidala Jagadesh Kumar argues that research should be judged on its own merits, not by whether the researcher appears in the list. [1]
• The tension: India needs both fundamental research and practical solutions. Making inclusion in the "top 2%" list the main goal distorts research priorities. [1]
• Why it matters for UPSC: The issue links to research-evaluation reform, NEP 2020, metric-driven governance (Goodhart's Law), India's quantity-versus-quality gap in publications, and linguistic inclusion in science (GS-II Education; GS-III S&T; GS-IV Ethics). [1][3]

2. Why in the News

• An opinion article titled "Science should move past 'top 2%' label" appears in The Hindu (International edition, p. 9, print dated 5 November 2026). [1]
• Trigger: recent media reports said India's presence in the "top 2% scientists' database" is on the rise. [1]
• Author: Mamidala Jagadesh Kumar, Chairman of the Review Committee for NEP 2020 (Ministry of Education) and Chairman of the Board of Governors, IIM Calcutta. He was formerly Vice-Chancellor of JNU and Chairman of UGC. [1]

3. Background & Evolution

• Creators: John Ioannidis and collaborators built the database. They have themselves cautioned against treating inclusion as conclusive evidence of research excellence. [1]
• Data base: it ranks scientists using publication and citation records from Scopus. [1]
• Indian context: India has long scored higher on output volume than on citation impact. One Nature blog analysis ranked India 7th in total paper output but 10th in citation impact among ten countries surveyed. [3]
• Domestic monitoring: a DST-CPR (Centre for Policy Research) study, "Trends and Trajectories in India's STEM Research for the period 2014–2023: A Scientometric Study" (June 2026), tracks India's research output, citation impact and highly cited papers. [2]

4. Core Static Facts

Item	Fact
Database	"Top 2% scientists' database" [1]
Creators	John Ioannidis and collaborators [1]
Underlying data	Scopus publication and citation records [1]
Headline measure	"Main impact scores", which have become the measure of research quality [1]
Multi-author factor	An author's position in the author list is one factor in the score [1]
Article author	M. Jagadesh Kumar (Chair, NEP 2020 Review Committee, MoE) [1]
Indian official scientometric study	DST-CPR report covering 2014–2023 (June 2026) [2]

• Limitations of the database highlighted in the article: [1] 1. It does not check the originality of the publications and citations it draws from Scopus. 2. Errors in author profiles and in the classification of disciplines affect the results. 3. It excludes contributions published in regional languages. 4. The way its composite impact scores are calculated is flawed, including the use of author position on multi-author papers.

• Key terms (general terminology, not taken from the sources):

• Citation impact: how often a researcher's or country's papers are cited, which is used as a proxy for influence.
• h-index: a researcher has an h-index of h if h of their papers have each been cited at least h times.
• Goodhart's Law: "when a measure becomes a target, it ceases to be a good measure."

5. Multi-Dimensional Analysis

Scientific / Technological

• Citation-based lists reward volume and visibility, not originality. The database does not verify originality. [1]
• India's gap between output and impact (high output, lower citation impact) shows that counts alone mislead. [3]
• Chasing citations may push researchers away from long-gestation fundamental research and locally relevant applied work. [1]

Ethical / Governance

• The article's concern is about inclusion becoming a target, and about exclusion being mistaken for lack of worth. [1]
• If institutions use the list in hiring, promotion or rankings, they create incentives for citation gaming, such as self-citation, citation cartels and honorary authorship. (Analytical inference.)
• Even the database's creators warn against using it as conclusive proof of excellence. Policymakers should respect that limit. [1]

Social / Linguistic equity

• Because regional-language contributions are excluded, scholarship in Indian languages is undervalued. This conflicts with NEP 2020's emphasis on mother-tongue and Indian-language education. [1]
• Disciplines with lower citation cultures (humanities, field sciences, agriculture) and researchers at under-resourced state universities are structurally disadvantaged. (Inference.)

Administrative

• Regulators and institutions such as UGC, NIRF and university promotion committees need qualitative peer review alongside metrics. The author's past roles at UGC and JNU give the critique weight. [1]
• Official scientometric studies such as DST-CPR's give policy-grade, India-specific evidence that commercial vanity lists cannot replace. [2]

Economic

• Research funding should be tied to national problem-solving (health, agriculture, energy), not to citation leaderboards. The article calls for both fundamental research and practical solutions. [1]

6. Recent Developments (last 12–18 months)

• June 2026: DST-CPR published its scientometric study of India's STEM research for 2014–2023. [2]
• Late 2026: media reports said Indian representation in the "top 2% scientists' database" was rising. [1]
• 5 Nov 2026: The Hindu published Prof. M. Jagadesh Kumar's op-ed urging that Indian science move past the "top 2%" label. [1]

7. Prelims Hooks

• The "top 2% scientists' database" was created by John Ioannidis and collaborators. [1]
• It relies on Scopus publication and citation data. [1]
• The database does not verify the originality of the publications and citations it draws on. [1]
• Regional-language publications are not counted in the database. [1]
• In multi-author papers, an author's position in the author list feeds into the impact score. [1]
• The database's creators themselves caution that inclusion is not conclusive evidence of excellence. [1]
• M. Jagadesh Kumar heads the Review Committee for NEP 2020 under the Ministry of Education. [1]
• M. Jagadesh Kumar is a former Chairman of UGC and a former VC of JNU. [1]
• DST-CPR released "Trends and Trajectories in India's STEM Research 2014–2023" in June 2026. [2]
• In one Nature comparison, India ranked 7th in paper output but 10th in citation impact among ten countries. [3]

8. How a Useful List Turns Into a Target

• The list counts citations, so citations become the thing people chase
• A citation is when another paper refers to your work. The list is built from Scopus citation counts [1].

• Once a job or a promotion depends on getting into the list, the easiest way to rise is to collect more citations, not to do better science. This is Goodhart's Law at work.

• The gaming is real and has been measured

• A self-citation is when you cite your own earlier papers. Ioannidis and colleagues studied about 100,000 highly cited scientists. The typical (median) self-citation rate was 12.7% [4].
• At least 250 of them cited their own work in more than half of all their citations [4].

• Ioannidis himself says much of the excess is indirect. It comes through "citation farms": groups of scientists who keep citing each other to push up each other's careers [4].

• The government's own push for more papers adds to the pressure

• In June 2026 ANRF launched a portal to train researchers in writing research papers and filing patents, to strengthen the "scientific publishing ecosystem" [9].

• Training is useful. But if institutions then measure success by how many papers come out and how often they are cited, the list becomes the scoreboard. India already has the gap the note describes: many papers, lower impact [3].

• The creators warn against exactly this use

• The 2019 study's authors said nobody should set a fixed self-citation cut-off to label researchers "honest" or "unethical" [4].
• So a hiring committee that uses the list as a simple yes/no test is using it in a way its makers did not intend [1][4].

9. The Strongest Case for Keeping the List, and Where It Fails

• The opposing argument: numbers are harder to bend than people
• Peer review means senior experts judge a candidate's work. In any system it can be shaped by personal links, loyalty to a school of thought or seniority.
• An outside list built on public Scopus data is the same for everyone. A young researcher at a state university can point to it without needing a patron.

• The list also tries to correct for some gaming. It gives weight to an author's position on multi-author papers [1], and the 2019 database made extreme self-citers visible for the first time [4].

• What is right about this argument

• Simply dropping the metrics and going back to committee judgement alone could bring back favouritism. A check from outside has real value.

• Measuring citations at a national level, as the DST-CPR study does [2], is still needed to track where India stands.

• Why it still does not justify using the list as proof of merit

• Being open to everyone is not the same as being fair. The list leaves out regional-language work and does not check originality [1]. Scholars in those areas lose out however open the data is.
• A citation farm can beat the self-citation correction, because the citations come from other people's papers, not your own [4].
• So the honest position is to use metrics as one input that a human judge reads. They should not be the final verdict. Even the creators say inclusion is not conclusive proof of excellence [1].

10. Indian-Language Science: Translated for Outreach, Still Not Counted

• The government translates research for the public
• ANRF is building SARAL AI. It turns research papers and patents into podcasts, short videos and social media posts in 18 Indian languages [8].

• This helps ordinary people understand science. But it only moves knowledge from English into Indian languages.

• Research first written in an Indian language still earns no credit

• The top 2% list leaves out regional-language publications [1].
• So a scientist who publishes original work in Tamil or Marathi gets no citation credit. If hiring follows the list, that scientist loses out on jobs and promotions.

• This clashes with NEP 2020's push for teaching and learning in Indian languages. The author of the article now chairs the committee reviewing that policy [1].

• Global bodies count this as an inclusion issue

• UNESCO's Recommendation on Open Science asks countries to make incentives support open science. UNESCO also supports DORA's call to change how research is evaluated [7].
• For India, this means the rules for evaluation must count Indian-language work. Translating papers is not enough.

11. What Other Systems Did Instead, and What India Can Copy

• Three international reform efforts India can borrow from
• DORA (2012): the San Francisco Declaration on Research Assessment. It asks funders and universities to judge research on its own content, not on journal-based numbers [6][7].
• Leiden Manifesto (2015): a set of principles that allow numbers to support expert judgement but never replace it [6].

• CoARA (2022): the Coalition for Advancing Research Assessment, a group of organisations working together to reform how research, researchers and institutions are assessed [6].

• Reform has its own trap, according to the OECD

• The OECD's 2026 report finds that organisations often monitor and reward even good practices, such as open science, through assessment rules. Used this way, those rules can reduce the freedom of researchers [5].

• So replacing one checklist with a new one does not solve the problem. Whatever replaces the list must leave room for expert judgement.

• ANRF should build DORA-style rules into its grant reviews

• As India's main research funder, ANRF sets the rewards that researchers chase.

• It could ask applicants to submit a few of their best papers with a short note on what each one contributed. That is the DORA approach of judging the work itself [6][7].

• UGC and university promotion committees should treat list membership as one piece of information, not a qualification

• The list's own creators say inclusion is not conclusive proof of excellence [1].

• So no hiring or promotion rule should treat "in the top 2%" as a pass/fail filter.

• DST should use its own data, not commercial lists, to track quality

• The DST-CPR study already tracks output, citation impact and highly cited papers for 2014–2023 [2].
• Policy should rely on India's own evidence of this kind, which can also be designed to include Indian-language work.

12. Anchors for Answers

• Data: In a study of about 100,000 highly cited scientists, the median self-citation rate was 12.7%. At least 250 of them cited their own work more than half the time [4].
• Data: In one Nature comparison of ten countries, India ranked 7th in paper output but 10th in citation impact [3].
• Report/Committee: OECD, Reforming Research Assessment for Better Science (2026) [5]; DST-CPR, Trends and Trajectories in India's STEM Research 2014–2023 (June 2026) [2].
• Comparison: DORA (2012), the Leiden Manifesto (2015) and CoARA (2022) all judge research on its content and let expert judgement lead, with numbers only as support [6]. UNESCO's Recommendation on Open Science asks countries to make incentives support open science [7].
• Scheme: ANRF's SARAL AI shares research in 18 Indian languages but does not change who gets credit in evaluation [8]. ANRF's paper-writing training portal (June 2026) shows the official push for more publications [9].

13. Mains Relevance

• GS-II: issues relating to the development and management of Education and Human Resources; government policies (NEP 2020).
• GS-III: Science & Technology: developments, indigenisation and the R&D ecosystem.
• GS-IV: ethics in public institutions, integrity in research, and the gap between means and ends (metrics versus merit).
• Essay: "Not everything that counts can be counted."
• Possible question stems: 1. Bibliometric rankings have become proxies for research excellence in Indian academia. Critically examine their limitations and suggest a balanced framework for research evaluation. 2. India ranks high in scientific output but lags in research impact. Analyse the causes and suggest measures to improve research quality. [3] 3. When a measure becomes a target, it ceases to be a good measure. Discuss with reference to research assessment in Indian higher education.

14. Related Topics to Study Next

• NEP 2020 research provisions: the author chairs its review committee, and the policy pushes for a research-intensive higher-education system.
• Anusandhan National Research Foundation (ANRF): the new apex funding body whose evaluation norms will shape research incentives.
• NIRF rankings: these use bibliometric parameters, so the same metric-gaming critique applies.
• One Nation One Subscription: this links journal access to publication and citation capacity.
• Research integrity and predatory journals (UGC-CARE): citation gaming and low-quality outlets inflate metrics.
• DORA and the Leiden Manifesto: the global movement for responsible research metrics.
• Indian-language science communication: this addresses the regional-language exclusion flagged in the article. [1]
• India's GERD (gross expenditure on R&D) and the S&T Indicators: these give the structural reasons for the gap between output and impact.

15. NCERT Concepts to Revise

• Goodhart's law (Economy): Core idea of the article: once top-2% inclusion becomes a target, it stops measuring research quality.
• National Education Policy 2020 (Economy): The author chairs its review committee; it stresses research and Indian-language education, which the database ignores.
• Anusandhan National Research Foundation (Polity): India's apex research funder; how it judges research will shape incentives beyond citation-based lists.
• University Grants Commission (Polity): Regulator whose hiring and promotion norms decide whether metrics like this list drive academic careers.
• Tertiary education (Economy): Universities are where research is evaluated and funded; the metric debate plays out in higher education.

16. Common Errors / Trap Areas

• "Stanford list" is an informal name: the database comes from Ioannidis and collaborators and is built on Scopus data. It is not an official ranking issued by any government. [1]
• Inclusion is not certification: the creators themselves disclaim it as conclusive proof of excellence. [1]
• Output rank is not impact rank: India's high publication volume does not mean equivalent citation impact. [3]
• The author's current role: M. Jagadesh Kumar is Chair of the NEP 2020 Review Committee and of the IIM Calcutta BoG. His UGC and JNU posts were earlier roles. [1]
• Agency confusion: the scientometric study cited here is a DST-CPR report under the Department of Science & Technology, not a UGC or Ministry of Education document. [2]