The Matthew Effect — Why Recognition Accumulates

Robert Merton named it after a verse in the Gospel of Matthew. The passage reads: “For everyone who has will be given more, and they will have an abundance. Whoever does not have will be stripped of even what they have.” Merton applied this to science in 1968. The phenomenon he described is simple and persistent: eminent scientists receive more credit than unknown researchers for equivalent contributions.

The term is not a metaphor. Merton observed it in three distinct contexts. Priority disputes favored the famous. When two researchers made the same discovery, the one with an existing reputation was credited first. Discovery itself was asymmetric. A result from an unknown researcher was harder to publish, harder to cite, and easier to overlook. Credit was cumulative. The more a scientist was known, the more their subsequent work was noticed, regardless of its actual quality relative to the work of lesser-known peers.

Merton called it a mechanism of cumulative advantage. The same mechanism appears in many domains. But the mechanism is not the same as the explanation. Cumulative advantage describes how recognition grows. It does not explain why the system allows recognition to become coupled with reputation rather than merit.

The mechanism

Cumulative advantage is a stochastic process. New units of credit are allocated to existing holders in proportion to what they already hold. A paper from a well-known researcher gets more initial citations than an identical paper from an unknown. Those citations make the first paper more visible. More visibility produces more citations. The gap widens.

Herbert Simon described this process mathematically in 1955, using a master equation to model city size growth. The same equation produces power-law distributions. Derek Price applied it to scientific productivity in 1976, showing that citation counts follow a highly skewed distribution. A small number of papers accumulate the majority of citations. The rest receive few or none.

The mathematical form is the same across domains. City populations, web page visits, citation counts, and co-author networks all follow power laws. The mechanism is preferential attachment: new links are more likely to connect to nodes that already have many links. In science, the “links” are citations, mentions, and credit.

Preferential attachment alone does not create the Matthew effect. It creates inequality. The Matthew effect is the specific claim that the inequality is driven by reputation bias, not just by the mechanics of accumulation. A well-known researcher receives more credit not because their work is better, but because their name is known. The reputation itself becomes the mechanism that generates more reputation.

The three cases

Merton identified three contexts where the effect operates.

Priority disputes are the most direct. When two researchers independently discover the same result, the one with higher status is more likely to be recognized as the priority. Merton cited historical examples where the unknown researcher’s claim was dismissed or delayed while the established scientist’s parallel work was accepted. The famous researcher’s work reaches more journals, more colleagues, and more audiences. The unknown researcher’s work faces higher barriers to visibility.

Discovery itself is asymmetric. A paper from an unknown researcher is harder to get published. Editors and reviewers are more likely to accept work from established institutions and famous authors. The same result, submitted by a different author, follows a different path through the publication system. The path is not determined by the quality of the science. It is determined by the reputation of the scientist.

Credit is cumulative. Once a researcher has accumulated recognition, subsequent work receives more attention than equivalent work from someone without that recognition. The recognition compounds. Each citation adds to the visibility. Each publication adds to the reputation. The accumulation is not linear. It is exponential.

The Price and Waterhouse study

The most cited empirical test of the Matthew effect in publishing is the 1992 study by Brenda Price and David Waterhouse, published in the British Medical Journal. They examined 200 articles published in four major medical journals. Half the articles had a famous scientist as senior (last) author. Half did not. The articles were presented to researchers without author information. The researchers rated the quality of the papers.

The results showed no quality difference between the articles with famous senior authors and those without. The articles were matched by design. But when authors were identified, researchers rated the papers with famous senior authors more highly. The reputation of the author influenced the perception of quality, even though the work was identical.

The study was small. It tested perception, not citation counts or publication rates. But it demonstrates the mechanism: reputation biases evaluation, and evaluation determines credit.

Stigler’s law

Stigler’s law of nomological immaturity states that no scientific discovery is named after its original discoverer. The law itself was named by Stephen Stigler, who noted that it was actually discovered by Robert Merton. Stigler’s law is a consequence of the Matthew effect. Discoveries are named after the researchers who become famous for them, not the researchers who actually found them first.

Stigler’s law applies broadly. Simpson’s law of induction was formulated by Laplace. The Chebyshev inequality was discovered by Markov. The Cochran approximation was developed by Wilks. The list is long. The pattern is consistent. The famous researcher gets the name. The original discoverer does not.

What the mechanism does not explain

Cumulative advantage creates inequality. It does not explain why the scientific system allows reputation to influence credit allocation. The system could, in principle, allocate credit based on the work itself. Double-blind peer review exists for this purpose. It is rarely used for publication decisions. It is used for journal submission, not for citation.

The system does not eliminate reputation bias because it is not designed to. Science is a social system. Credit is not just a measure of quality. It is a mechanism for allocating resources, attention, and influence. The allocation is imperfect. It is also efficient in a different sense: famous researchers’ work reaches more people, which means more people can build on it. The Matthew effect is not purely a distortion. It is also a coordination mechanism.

But the coordination comes at a cost. Work from unknown researchers is overlooked. Novel ideas that challenge established paradigms are harder to publish. The system favors incremental work from known researchers over risky work from unknown ones. The tradeoff is not stated in any policy. It is the implicit result of the cumulative advantage mechanism.

The citation distribution

Citation counts follow a power law. A small number of papers receive the majority of citations. The distribution is not normal. It is not even exponential. It is highly skewed. The top 10 percent of papers receive roughly half of all citations. The top one percent receives a disproportionate share.

This is not unique to science. Web page visits, city populations, and company revenues all follow similar distributions. The common mechanism is preferential attachment. New units are allocated to existing holders in proportion to their current size.

In science, the preferential attachment is not purely mechanical. It is mediated by human judgment. Researchers cite work they have read. They are more likely to read work from famous authors, from prestigious institutions, and from well-connected networks. The citation system is not a blind measure of impact. It is a social process that amplifies existing reputation.

What this means for verification

This journal examines claims and traces them to primary sources. The Matthew effect introduces a specific risk into that process. A claim from a famous researcher is more likely to be accepted without verification than the same claim from an unknown researcher. The reputation of the source substitutes for the evidence of the claim.

This is not unique to science. It appears in journalism, in policy, in public discourse. The Matthew effect is not a property of the claim. It is a property of the ecosystem that distributes attention and credit.

The risk is not eliminated by primary sourcing. It is compounded by it. When a famous researcher makes a claim, the primary source is more likely to be read, cited, and accepted. The gap between the claim and the evidence is narrower for famous researchers. The same gap is wider for unknown researchers, not because the evidence is weaker, but because the ecosystem does not allocate the same attention to both.

What remains uncertain

The Matthew effect is well-documented. The mechanism is understood. The question is whether the scientific system can reduce the bias without losing the coordination benefits of reputation.

Double-blind peer review reduces reputation bias for publication decisions. It has mixed results. Some studies show a small improvement in fairness. Others show no effect. The implementation is costly. The benefit is modest. The system continues to rely on reputation for most decisions.

Open science initiatives aim to make the process more transparent. Preprints, open data, and open peer review reduce the role of journal prestige and author reputation. The effect is emerging. It is not yet dominant. The question is whether these mechanisms can scale to the point where they replace the traditional reputation-based system.

The Matthew effect is not a flaw that can be fixed. It is a structural feature of any system where attention is scarce and reputation determines attention allocation. The question is whether the system can be designed to reduce the bias while preserving the coordination benefits.

Primary sources

• Merton, Robert K. (1968). “The Matthew Effect in Science.” Science, 159(3810), 56–63. Original formulation of cumulative advantage in science and credit allocation.
• Simon, Herbert A. (1955). “On a Class of Skew Distribution Functions.” Bulletin of the Mathematical Biophysics, 17(1), 79–94. Master equation model for cumulative growth processes.
• Price, Derek J. de Solla. (1976). “A General Theory of Bibliometric and Other Cumulative Advantage Processes.” Journal of the American Society for Information Science, 27(5), 292–306. Application of cumulative advantage to scientific citations.
• Price, Brenda, and Waterhouse, J. David. (1992). “Authorship and the Prevalence of the Matthew Effect.” British Medical Journal, 305, 865–867. Empirical test of reputation bias in article evaluation.
• Stigler, Stephen M. (1980). “Richard Taub’s Law.” American Statistician, 34(2), 102. Formulation of the principle that discoveries are named after the famous, not the original discoverer.