Hi,
"What to read about corona numbers?" I asked in my last newsletter, dated Thursday 12 March. It was, in hindsight, the day when shit started hitting the fan, at least in the Netherlands.
I was in Leipzig for a lecture which ended up being cancelled at the last minute. Slightly bewildered, I walked around the Stasi Museum as I kept refreshing my news app. In my inbox I read that, at The Correspondent, we now had to work from home until further notice.
It’s bizarre how fast things can go. In the meantime, many countries, including the Netherlands, went further into lockdown. Schools closed, events were cancelled, and fines can now be handed out to people who don’t adhere to the social distancing policy.
By now, I’d better ask: what not to read about corona figures? Because the calculations and graphs are coming thick and fast. "[The] number of virologists in the Netherlands increased to 12 million", joked de Speld, a satirical Dutch news website. "Twitter in particular is proving to be the source of infection." Buzzfeed also wrote about the "rise of the coronavirus influencers". You’ve probably seen them, with messages that start with "I’m not an epidemiologist, but...".
Last week, I wrote a guide to make sense of the numbers you see flying around. What does ‘R nought’ mean? What is exponential growth? What do we know and, above all, what don’t we know when it comes to corona cases? I also wrote an explainer about group immunity.
I’m not an epidemiologist, but...
Meanwhile, my timeline is full of all kinds of messages with calculations, such as projections of how the spread of the virus will proceed. To be honest, I find it difficult to relate to such calculations, or find comfort in them. I think it’s important to understand the numbers and I love it when people do the math.
But it’s often underestimated how complicated modelling this situation really is. If you want to predict the course of this pandemic, or the effects of the measures, you have to make all kinds of assumptions.
For example, how contagious is the virus? It is now thought that someone infects on average two other people, but that’s still uncertain. And then that ‘R0’ also differs per person. People are not equally contagious, and there is the possibility of "super-spreaders". In reality, the contagion also depends on the measures taken. Soon enough, your model becomes more and more complicated.
Hence, watch out for back-of-the-envelope calculations. An interesting STAT news article explains more about the intricacies of modelling. But I would like to read more about it. Do you have any reading tips on epidemiological models? I’d love to hear them.
Rotten eggs
Even if you don’t fully grasp the subtleties of statistical models, you don’t have to be an epidemiologist to recognise the bullshit in the messages you see about numbers. As Dutch science journalist Hans van Maanen once wrote: "We can’t lay an egg, but sometimes we can smell if it’s rotten."
For example, it’s tempting to compare one country to another. Last weekend, a table went viral that contrasted the trend in deaths in the Netherlands with that in Italy. It looked very similar. The message: we’re going the same way as Italy! But there are good reasons not to jump to that conclusion.
For one, it matters on which day you start to count. If you count from the first death, then the situation in the Netherlands – in absolute numbers – looks better. If you only start counting at the 100th infection, it looks even better. It all depends on the approach you take.
But more importantly: you are comparing apples and oranges. The Netherlands has a different healthcare system and a different population structure. We’re only now slowly beginning to see the effects of the measures taken, because you have to take into account the incubation period of the virus.
All of this doesn’t mean that the Netherlands will definitely not follow in Italy’s footsteps. But you can’t make that claim based on a simple table.
Before you go...
...We’re not done with coronavirus yet, unfortunately, and neither are we done with the numbers. So keep sharing your tips and questions! Last week was so hectic that I couldn’t respond to everything, but I really appreciate your contributions.
Prefer to receive this newsletter in your inbox?
Follow my weekly newsletter to receive notes, thoughts, or questions on the topic of Numeracy and AI.