Ten Data Challenges We Face With COVID-19 (Pt. 1)
On March 12 the WHO announced that "More cases are now being reported every day than were reported in China at the height of its epidemic," marking a new phase of the Coronavirus pandemic. Every day there seem to be more numbers circulated via every medium, with social media playing a prominent role. Although this proliferation of information can seem like a boon, there are certain challenges we need to address if we want to ensure these data have a positive impact of the overall trend of the infection. These are the first five.
1) Forget data literacy, numeracy is the first hurdle. Percentages feature heavily in reporting because they provide insight into both the scale and severity of the crisis. Yet we know that in certain countries adult numeracy skills are worryingly low. As few as one in four US adults with only a high school diploma was able to correctly answer a question relating to percentages. What does this say about citizens' ability to parse data on infections and make informed decisions as a result? This is why clear explanations matter.
2) It is hard to feel a sense of urgency when you are only thinking linearly. The problem is that humans are good at it. If you get paid $100 every week, it is not hard to understand that in four weeks you will have accrued $400. Yet viral infections do not follow linear but exponential curves. This matters because it means that we will always intuitively wildly underestimate the speed of contagion. As this 3Blue1Brown video explains, at the current rate of infection we will see a ten-fold increase in cases roughly every two weeks.
3) Providing context to numbers takes time but is essential. During times of volatility everyone is after the latest data. Yet in a world of fragmented communication channels it is easy for numbers to be taken out of context. For example, understanding the current number of infections is of limited use if you do not know total population size or the growth rate. One reason this post by Tomas Pueyo has done so well is that he does an excellent job of providing context to the numbers through clear visualisations and detailed explanations.
4) We tend to accept data at face value. Rarely do we stop to question what they mean and whether they are accurate. For example, while numbers of total infections are bandied about liberally, it is important to note that these are only confirmed infections. Given the limited number of people tested in many countries, including the US, the true number of total infections is likely to be much higher. Ensuring that everyone understands the nature of these numbers is important when articulating the urgency of the response required.
5) Having data does not automatically mean acting rationally. Over the past two weeks it has become very clear that the number of cases will continue to accelerate if we do not institute significant preventative measures, such as social distancing and self-quarantining. Despite these truths experts believe to be self-evident, even certain heads of state are refusing to acknowledge the severity of the crisis. Framing data in a way that increases its relevance or emotional resonance is essential in helping the message land.
What has now become clear is that the situation will worsen before it improves. How we act now based on the data at our disposal is going to make the difference between whether tens of thousands will die – at this point likely inevitable – or whether we will face losses of life magnitudes greater.
– Ryan