HUSAIN ALKHALDY · ESSAYS  —  ESSAY

What Does the Man Behind the Stanford 2% List Say About It?

Husain AlkhaldyEnglish only

Every autumn, scientists named in the "Stanford Top 2% Scientists" list celebrate their inclusion with congratulatory posts and banners. Some universities record it on staff profiles as an honorary award granted by Stanford University. The man who built the database, John P. A. Ioannidis, has said something quite different. His own papers open with a warning, not a celebration: the first sentence of the 2019 paper that launched the database states that citation metrics are widely used and misused.

This essay sets out what Ioannidis and his co-authors actually wrote about the database: why they built it, what they said it should and should not be used for, and the warnings they attached to it. The irony is plain. A meta-researcher who has spent his career criticising crude metrics produced the dataset that is now most often used as one.

What he built, and why

Ioannidis built the database to fix the technical mess of citation counting, not to crown anyone. In the 2019 PLOS Biology paper, written with Jeroen Baas of Elsevier and Richard Klavans and Kevin Boyack of SciTech Strategies, he listed the problems. Different databases give different numbers. Scientists self-report figures on their curriculum vitae (CV) that are often inaccurate. Self-citations are handled erratically. Comparing fields with very different citation densities is shaky.

The answer was a standardized composite indicator (the "c-score") built from six metrics, drawn from Scopus:

Each is given with and without self-citations. Scientists are assigned to 22 fields and about 175 subfields, so a researcher can be compared with peers in the same discipline.

The "2%" itself came later. The 2020 update added scientists outside the overall top 100,000 who were in the top 2% of their own subfield. The stated reason was fairness to low-citation fields, which would otherwise rarely appear. Ioannidis framed the threshold as a correction for field bias, not as a badge of elite status.

What he says it is not

Ioannidis never called it a list of the world's best scientists. The papers describe a "database" of "standardized citation indicators", and the word "award" appears nowhere. He released it as downloadable spreadsheets on Mendeley Data, explicitly choosing not to build a formulaic website, so that users could analyse the data however they wished.

The caveats are written into the papers themselves:

His closing line in 2019 is the clearest statement of intent. He wrote that the team could not emphasise enough that these metrics must be used prudently. The hope was a more nuanced use of metrics, avoiding what he called the egregious errors of raw "bean-counting".

Self-citation and citation farms: his own warnings

Ioannidis built the means of detecting gaming into the list. He wrote that extreme self-citation and "citation farms" (small clusters of authors massively citing each other) make citation metrics spurious and meaningless. Every indicator is therefore reported with and without self-citations, and the database shows the ratio of citations to citing papers, a signal of citation stacking.

His own numbers were sobering. Among the 2019 top 100,000 career-long authors:

His advice for extreme cases was blunt: with very high self-citation, do not use any citation metric for that person, since it may signal other spurious features. Simply removing self-citations may not be enough. Each case needs individual examination.

Speaking to Nature in 2019, he went further, saying he believed self-citation farms are far more common than people assume. He also stressed that self-citation above 25% is not necessarily unethical but may warrant scrutiny (ScienceAlert).

The 2020 update quantified the effect on the 2% threshold itself. About 4.9% of scientists in their subfield's top 2% with self-citations fell out once self-citations were removed. The paper also noted that some authors had manipulated peer review to add citations to their own work.

The meta-researcher behind the metric

The database makes sense only against Ioannidis's wider work. He co-founded the Meta-Research Innovation Center at Stanford (METRICS), and his 2005 essay "Why Most Published Research Findings Are False" made him the best-known critic of how science rewards itself. The citation database is a tool from that same project: measuring research in order to expose its distortions.

The 2019 paper names the misuses he had in mind in hiring, promotion, tenure, funding and awards:

It cites the Leiden Manifesto for research metrics and a 2018 paper, co-authored by Ioannidis, on assessing scientists for hiring, promotion and tenure. That paper argued that institutions should reward rigour, transparency and reproducibility, not counts. The database was meant to make metric use more careful where metrics are used at all. It was never meant to replace judgement of the work itself.

Reading the list the way its author intended

Taken at his word, Ioannidis offers a diagnostic instrument, not a hall of fame. A reader who follows his guidance would:

The gap between the creator's caveats and the public celebration is the real story. Ioannidis published a tool for measuring how science is cited, with the means to catch its abuse built in. Institutions turned its threshold into a prize. On his own terms, being in the top 2% says something about how often a body of work is cited, adjusted for field. It says nothing on its own about whether that work is right, original or useful.

Sources

NEUTRAL FACT-CHECK

The factual claims in this essay (the dates, the figures, the attributions and the quotations) have been checked separately, without touching the argument. The essay was not edited to agree with the check. Read the fact-check →

© 2026 Husain Alkhaldy — published in the site's essays.