Inequality in science: an outstanding debt

The most unequal country, according to the 2022 measurement, is South Africa, with a Gini index of 0.63, while the countries with the most balanced societies are Iceland, the Czech Republic, Norway, Finland and others, with values around 0.25.

We have heard this index mentioned many times, and a few considerations are worth explaining. The Gini index is a measure of the inequality of any distribution and can be applied to many areas of human activity. It is a mathematical measure and does not depend on the specific context, but on the distribution of a concrete variable within a population. It is best known for comparing income or wealth inequality between countries, but it can also be used to measure inequality in other domains such as education, health and access to resources — even science.

For example, in education the Gini index can be used to measure the distribution of educational attainment or exam grades across a population. In health, it can be used to measure the distribution of health outcomes or access to them.

Let us see how it is calculated. The index is obtained from the so-called Lorenz curve, which shows how a certain outcome is distributed in a population and compares it with perfect equality. By “outcome” we can understand various variables such as income, access to health, athletes’ salaries, citations of scientific articles, and so on.

To build the Lorenz curve, we plot the cumulative percentage of the population on the horizontal axis (X axis) and the cumulative percentage of the outcomes obtained on the vertical axis (Y axis). Perfect equity would appear on the graph as a line in which each percentage increase in population equals the same increase in the outcome: 20% of the population would obtain 20% of the outcomes, 50% would obtain 50%, and so on. The Lorenz curve would be a straight line with a 45° slope.

Curves that deviate further from the equality line have more apparent inequality and a higher Gini index. From the figure, the index is obtained by dividing area A by the sum of A+B: G = A/(A+B).

Many factors influence the Gini index, and each setting has its own peculiarities. For instance, to lower the Gini index of wealth, many countries take measures such as raising taxes or creating incentives for companies to raise salaries, and so on.

The index is applied to other activities. For example, if we compare the salaries earned by athletes, we see that Major League Baseball (MLB) has the highest index, with G=0.63. The English Premier League (EPL), however, has G=0.45. This could be explained because MLB does not have a salary regulation system, whereas the competitive EPL does.

Fortunately, in health the Gini values obtained come close to full equity, with values below 0.05 in studies carried out on various populations.

In an article published in the journal PNAS, two researchers from Denmark analyzed 4 million authors and 26 million scientific articles, and found that between 2000 and 2015 the Gini coefficient for citations of scientific articles rose from 0.65 to 0.70. The top 1% of most-cited scientists increased their cumulative share of citations from 14% to 21%.

A deeper analysis shows that the most-cited scientists reside at top universities in Western Europe and Australia, while in the United States there has been a slight decline.

The authors state that the distribution of scientific rewards is remarkably unequal, and a relatively small stratum of elite scientists enjoys exceptional privileges in terms of funding, research facilities, professional reputation and influence. This is known in the literature as the Matthew effect: the most prominent scientists receive more rewards than their research deserves. Recent data point to a growing gap between the “rich” and the “poor” of science in terms of salary levels, research funding and the accumulation of scientific awards.

Inequality can foster creative competition within the scientific system. However, it can also lead to a dense concentration of resources with diminishing returns on investment (intellectual and fiscal) and to monopolies in the marketplace of ideas.

This is no trivial matter, since many variables can influence the situation. The factor I consider most influential is the amount or percentage of GDP these countries devote to science, and specifically the funding received by the most-cited research groups. Generally, there is a high correlation between articles and citations and the funding received.

On the other hand, publishers have made a business out of science. Publishing in high-impact journals is a goal of many scientists to ensure greater visibility for their work, better evaluation of projects, and so on. These journals, in general, since they want to keep citations — and with them the journal’s “impact factor” — high, do not take risks and usually only pass established groups on to the editorial committee. In Nature, one of the journals with the highest impact index, only 10% of submitted articles are accepted for review; 5 countries have approval rates above 5% and 17 above 1%. The rest of the countries have approval rates below 1%.

Authors with more citations and articles are more likely to obtain funding, which allows them to expand their research labs and collaboration networks and, ultimately, to increase their publication and citation rates. In an evaluation system where funding and hiring decisions are largely based on bibliometric indices, an increase in the concentration of publications and citations is to be expected.

There are measures that can mitigate the Matthew effect. In Chile, for example, only the applicant’s track record over the last 5 years of scientific output is assessed through national projects.

The trend shows that science worldwide is moving toward knowledge-concentrating centers, a sort of research monopoly. Researchers from third-world countries try to collaborate with these elite centers to give visibility to their ideas, but this has only reinforced the effect.

There are many stories where scientific results obtained in the Southern Cone are cited less than others of equal or lower level produced in the Northern Hemisphere. In my opinion, there is a tendency toward “easy” citation: citing articles from established groups in the hope of indirectly boosting our own result. A deep and impartial literature review of the research topic is imperative, and citing accordingly.

Finally, by way of conclusion, there are many questions to answer.

Is this the best formula for quality science? Is a decentralized science preferable? Are measures needed to prevent inequality in science, just as is done in the economy? Can we do anything more in our countries, beyond increasing the GDP allocated to science?

In my opinion, we must at least promote internal spaces for collaboration — at the level of universities, country and region. Basic science requires substantial funding, so alliances would help offset the deficit. In addition, we must, institutionally, encourage collaborative work. Faced with a global variable, only local policies would help reverse this phenomenon.