The data are bad (translation from El País)

Another break in programming–here is an article from yesterday’s El País newspaper, written by their science reporter, Javier Sampedro. I translated it quickly as usual, all mistakes are mine.

The data are bad

The number of coronavirus cases [in Spain] is not 56,188—Undoubtedly they are closer to a half a million

Javier Sampedro
26 mar 2020 – 19:30 EDT

The El Pais Ombudsman (1) titled his March 1 column “Journalists Against Mathematics” in an attempt to correct the blunders of arithmetic that we word-pilers (2), as they call me in my neighborhood, make. Read his article if you haven’t already, it is very instructive and entertaining. There are also journalists in favor of mathematics, and the newspaper’s data team (Andrino, Grasso, and Llaneras) is an impeccable example. Their live article—updated daily— about the curve of the infected, in their lineal and logarithmic versions, is my breakfast, especially now that I’ve run out of milk. But if it’s to be useful, mathematical analysis has to feed on reliable data. And we don’t have reliable data.

The statistics that Fernando Simón, the head of health emergencies, gave us yesterday (4,145 deaths and 56,188 infected) implies a mortality of 7%, which is an absurd rate. The deadliness of this coronavirus is less than 2%, probable closer to 1%. The reason that the mortality rate seems so high is that we are undervaluing on a monumental scale the total number of infected people. We have spent weeks without testing people who have mild symptoms or are asymptomatic, who are the lion’s share of the infected. Coronavirus doesn’t kill as much as we think, but it is far more contagious than we say. The official statistics are misleading. Why?

Antonio Durán Guardeño, professor of mathematical analysis at the Universidad de Sevilla, has spent the entire month working on this serious problem of estimation. According to the statistical methods that he has tried, the number of people infected with coronavirus in Spain are not the 56,188 of the official. They are undoubtedly closer to a half a million. The foundation of this estimation is strongly based on the segregation of data by the ages of the infected. Math fans can read the full story on the Instituto de Matemáticas’s blog at the Universidad de Sevilla. I hope that the politicians in charge read the blog. Although I think that some already have.

The mathematician published his blog entry last Monday. Maybe by coincidence, the government of Andalucía stopped publishing data about the infected and dead divided by age, as they had been doing before. Remember that it is exactly those groupings by age that have the greatest statistical value in order to calculate the total number of infected people. The Community of Madrid also does not divide their statistics by age. And the Ministry of Health, therefore, does not either. The authorities are hiding information from us that they think we are not prepared to hear. They are treating us like we are children or clueless. It is a terrible strategy. Scientists offer us the truth. Let’s listen to them.

Coronavirus Covid-19

(1) The literal translation of Ombudsman  (el Defensor del Lector) is the Reader’s Defender, which I love.

(2) “Juntaletras” which means literally something like, “word joiners” or “word gatherers”—such a great evocative word, but I can’t think of a good translation that captures the term properly.  It’s an old-fashioned term for “journalist” but I imagine it’s said by someone who is semi-literate. So if a person gathers wood into a wood pile, a journalist gathers words into a “word pile” hence, they are “word-pilers”…yeah, I overthought that one.


Original article:

Los datos están mal: Los contagiados por coronavirus no son 56.188. Seguramente están más cerca del medio millón

people waiting on a metro platform
people waiting on a metro platform in the BC (before coronavirus)