WHEN DATA SAYS NOTHING

THE MEDIA OFFERS CONTEXT-FREE DATA THAT DO NOT INFORM THE CITIZEN.

A figure alone says nothing. It seems obvious, but I constantly repeat this to my students in the data visualisation courses I teach. And I do so because it is very common to see in the media how we are bombarded with facts and figures that say nothing.

One of the examples that bothers me the most is the number of ICU beds occupied in a given region or province. At the height of the different waves of the pandemic we are experiencing, the media usually report how critical the situation is in hospitals due to the shortage of ICU beds. Well, in most cases they give the number of occupied ICU beds, but do not indicate the percentage of occupation, so it is as if they were not saying anything at all. What difference does it make whether a province has, say, 450 or 900 occupied ICU beds; if we don’t know how many it has in total, that data is of little importance. It would be more useful, that is, they would give us information (and not just a figure that is useless) if they always gave the percentage of ICU beds occupied by COVID-19 patients.

I think this happens because there is a general tendency in the media to attract attention, they need to capture our interest and to do so they believe that the higher the number (not the data), the greater its capacity to have an impact. The number 450 (e.g. ICU beds) is bigger than the number 61 (e.g. percentage of ICU beds occupied). They believe that saying 450 instead of 61 impresses the audience more and captures their attention. However, that 61% of ICU beds are occupied by COVID-19 patients is more impressive than saying that there are 450 ICU beds occupied by COVID-19 patients if we do not know how many ICU beds there are in total.

Another example of this, also related to the pandemic, is the data from Brazil. The media often portray a dramatic picture of the South American country. But is the situation so different from our country? Not really. The media portray Brazil with two figures: deaths (368,749 as of 17 April) and infected (13,832,455 on the same date). However, if we put these figures in relation to its total population, we see that the percentage of people who have died from COVID-19 in Brazil is 0.17% compared to 0.16% in Spain on the same date. In the case of those infected, Spain is worse off than Brazil, with 7.23% of the Spanish population infected, compared to 6.55% of the Brazilian population on the same date. It is true that a single life lost is dramatic and that sometimes when we put percentages, we seem to relativise the seriousness. But if we do not do so, we run the risk of misinforming instead of communicating and giving correct information to citizens so that they can draw their own conclusions.

Communicating data is not about attracting attention, it is about providing information to deduce the right consequences.

The media need news, headlines, figures, statements that are historic, astonishing, exceptional, record-breaking. As with football matches, where every season there is a match of the century, the media need the figures to have as many digits as possible in order to be able to use a superlative adjective. Often forgetting that they are not saying anything.

As I tell my students, a number is not data. For that to happen, the number has to convey information. Someone should remind the media of this. Communicating data is not difficult, but it needs to be done well.