Information Pandemic: Initial explorations of COVID-19 coverage

The Media Cloud team has started using our repository of news publications from across the globe to look at the coverage of the coronavirus (COVID-19) pandemic in various digital media ecosystems. It will take time to answer most of the questions that we have, but we can share some preliminary data.

It seems obvious that coverage of the coronavirus and its implications has taken over media agendas across the world. We were interested in seeing to what extent and how that had happened. In particular, we looked at media across Western Europe and the US and their reaction to the arrival of the virus to those parts of the world. We used Media Cloud’s National and State & Local collections in the US, Italy, Spain, Germany, the UK, and France to look at media coverage and we used Worldometers data on coronavirus cases to draw the backdrop for our analysis.

Daily data from Worldometers starts only on February 15th, when there were already a few cases in each of the countries (15 in the US, 3 in Italy, 2 in Spain, 16 in Germany, 9 in the UK, and 12 in France), but it clearly shows the inflection point of the epidemic (with Italy’s starting first and spreading further).

In parallel to the spread of the virus, the media in all these countries started paying increased attention to it. The graph below shows the percentage of all stories published in each country that mention the coronavirus (from Jan 1st until mid-March).

A few things are worth noting in that graph. The first is the unprecedented focus of all media ecosystems on a specific topic. By March 18th, over 70% of all stories published in Italy, the UK, and Spain mentioned the virus. In the US and France it was slightly lower (about two thirds of all stories), while Germany seemed to be the only country with room for other issues (about half of the stories mentioned the virus).

As points of comparison, we looked at the media attention paid to (1) Trump’s election by US media in November 2016 and (2) the Brexit referendum by UK media in June-July 2016, and compared them to (3) the attention paid to coronavirus by US media over the past month. Despite the high peaks on the day of the US presidential election, and the day of the Brexit referendum, the most newsworthy political stories of the past years pale in comparison to the virus’ coverage. The dominance of this story in media agendas is literally unprecedented in our experience.

The second notable fact is a similar pattern of attention across studied countries, with the exception of Italy. While it is true that each country’s media ecosystem has its own variations, in  five of the countries we considered, we see a quite similar evolution: increased attention being paid starting in the fourth week of January, a spike in coverage during the last week of February, and a burst in attention starting the second week of March.

The first surge, in late January, coincides with the first confirmed cases of coronavirus in Europe and the US. The spike in late February coincides with the outbreak in Northern Italy. The surge in the second week of March, while no doubt related to the pandemic declaration by the WHO, actually precedes the declaration. It could be said that media coverage showed three stages, the same way that the virus expansion went from isolated cases, to outbreak/local epidemic, and to pandemic. The following graph, showing Spain’s media coverage since January 1st, clearly displays these three stages:

The third element that stands out when examining coverage over time is the pattern followed by Italy (see graph below). Its distinctive evolution, though also showing three steps, is much sharper than in all other countries. No doubt the fact that Northern Italy was the first place in Europe where an outbreak of the virus was identified contributed to the sudden increase in coverage in late February. On February 20th there were only 4 confirmed cases of the virus in Italy, by February 23rd there were 157. As a result, media attention (measured as the % of stories mentioning the virus) increased by 240% over the previous day on February 21st, 63% on February 22nd, and 43% on February 23rd. All other countries saw an increase in attention to the virus in the days following the Italian outbreak, but none experienced such a sudden and steep surge. It would take about three weeks for the media ecosystems of the other countries to reach the levels of attention paid to the virus by Italian media. At that point we were already in the middle of a pandemic.

It's difficult to say exactly when a pandemic reaches a country, as the virus may have spread for weeks before tests identify patients. We might consider 100 cases, though, as a moment where it's clear that a nation is facing a significant public health challenge. Italy reported its hundredth case on February 23rd, and quickly experienced a media tipping point, with more than 50% of stories mentioning coronavirus the next day. Attention drops, returning to the 50% mark on March 6th, 12 days after the 100 case point. We see similar patterns in Spain and the US (100 cases on March 2nd, a majority of stories on March 12th). The cycle in France and Germany is slightly slower (14 days between 100 cases and 50% of stories) and faster in the UK (eight days), perhaps because the progress of the disease in nearby countries had become obvious.

There's lots more we want to do in understanding coronavirus and media attention, including expanding this analysis to include other countries, and beginning to understand how political partisanship may have affected coronavirus coverage in the US and elsewhere. We will keep sharing analysis as we complete it.