Corona – Why media reporting is misleading

 

[UPDATE: case in point from El Pais – Spain’s graph attached at bottom of post]

Media – whether traditional or social – is doing a less than good job of reporting on Corona.

Instead of analyzing the trends, how the growth of the virus and number of deaths evolve over time, most of them report exclusively on individual daily fluctuations, in an alarmist manner.

Obviously, it’s a good way to gain “clicks” to paint a sudden day-to-day increase of the numbers as doomsday is coming, but it has very little to do with reality and truth. 

An example from yesterday as point in case: for several consecutive days, the number of deaths in Italy had been decreasing. Which obviously is a very good thing. But then, for yesterday (Monday) the number of deaths took a jump from 525 to 636. And several media – established media among them – immediately sounded the alarm bell in capital letters: “deaths in Italy growing again!”.

While that statement is true in a very narrow technical sense, it does a lousy job of explaining how the virus, and it’s resulting deaths is evolving, and what we might expect for the near term future. 

Technically speaking, media loves to report alarmist point data, while they instead should be reporting on trends, in particular, growth trends.

Let’s have a look at two cases, Italy and Sweden:

Italy:

First, let’s look at how the growth factors for confirmed and deaths have evolved over a number of weeks:

Growth_factors_for_Italy

What media yesterday reported, in a very alarmist manner, was that almost invisible last jump upwards for the red graph, representing the growth factor for deaths. What media completely ignored to report was the big picture, that over the past 25 days, the trends for both confirmed and deaths have been going steadily down in Italy.  Yes, with some random jumps upwards now and then, but overall, a clear decline of growth.

Next, let’s take a look at the graph for Expected Number of Deaths for Italy:

Italy_violinplot

Here we can see that over the past 6 days, including yesterday, which was reported as “sudden raise of number of deaths!”, all daily death counts in Italy are extreme outliers on the low side, that is, the process is definitely slowing down.

Still, media reports with “Breaking News” on individual, day-to-day random fluctuations.

A better way to report on processes with huge daily random fluctuations is to average or sum over a longer period, e.g reporting on weekly totals instead of daily totals. So let’s see what Italy looks like when the data is presented as weekly totals:

corona_weekly_sums_Italy

Here we can clearly see that both processes (growth of virus, growth of deaths) have been declining since early March. With this type of summary over a longer period, the daily random fluctuations no longer distort the Big Picture – and the Big Picture is what we should really care about, since that is what should determine what actions to take next.

Let’s look at the same graphs for Sweden, which is an interesting case, since almost all the world is questioning Sweden’s strategy. As usual by now, yesterday the number of deaths took a steep jump upwards in Sweden, as it “always” does on Mondays (see my earlier post on the Swedish “Weekend Effect”), and immediately all the “experts” in media – social as well as non-social (ignorant/alarmist journalists) – started panicing:

Growth_factors_for_Sweden

That last almost vertical jump upwards does indeed look scary, and our ignorant (but perhaps well-meaning) journalists immediately jumped. But the question is: is it really representative for the overall trend…?

Let’s dig further into the data:

Let’s look at the Expected Number of Deaths:

Sweden_violinplot

Yes, it indeed looks like yesterdays number was an outlier, so perhaps doomsday really is here… ?

Not necessarily – remember that all daily numbers are impacted by random fluctuations, and particularly for Sweden, the “Weekend Effect” plays a huge role. So let’s smooth out the daily random fluctuations by looking at weekly sums:

corona_weekly_sums_Sweden

What we see here is that the growth for the virus over the past week has been declining, despite the huge jump yesterday.

What is true, however, is that the number of deaths per capita in Sweden is high, compared to most other countries, and that IS a cause for concern. However,  the reason for the high number of deaths in Sweden is an interesting topic in it self, and I hope being able to return to that topic in a day or two. But just to throw a bone here to wet your appetite:

Sweden has the lowest per capita number of Intensive-Care-Units (ICU’s) in all of Europe. And it has been so now for many years, long before the outbreak of Corona… Basically, despite often being painted as a “Humanitarian Superpower” having “World Class Medical Care” by our politicians, the sad fact is that Medical Care in Sweden of today is way below bar, it lacks money & resources of all kinds, and that has very little to do with Corona – people have been dying waiting for surgery for e.g. cancer and other deceases for years now.

So, at least to me it’s been clear for a long time that medical care in Sweden would have very slim chance of dealing with a major outbreak of any epidemic, and now Corona seems to prove that. In addtion to that, our Swedish politicians also decided, after the fall of the Berlin Wall, that it was no longer necessary to maintain stores of critical goods, such as face masks, medicine, nor food or energy (oil etc); everything should be purchased “just-in-time”… Well, if I’ve learned anything from this Corona outbreak, it is that all Global Delivery Chains and Just-in-Time processes come to a halt very quickly in a global crisis – when the shit really hits the fan, you can only rely upon your own, national resources.

Summary:

In order to understand what’s really going on with the evolution of Corona, you can’t just take what the journalists and self proclamed experts on social media present to you: both groups seem to either have a hidden agenda, or then are simply so ignorant about facts and maths, that they simply don’t understand what they see. 

Spain:

Spain_violinplot

 

 

 

About swdevperestroika

High tech industry veteran, avid hacker reluctantly transformed to mgmt consultant.
This entry was posted in Big Business, Data Analytics, Epidemics, Probability, Society, Statistics and tagged , , , , , . Bookmark the permalink.

2 Responses to Corona – Why media reporting is misleading

  1. Joe Marasco says:

    Excellent post! Use alarmist instead of alarmistic.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s