I frequently see variations of these graphs on websites looking at the covid rates in different countries.
Individually each graph does a great job of showing the changes over time for each country. However, these are structured to encourage the reader to look for comparisons and that gives a very different story. The title states it shows the countries with the highest rates, using a comparative word like highest in the title encourages the reader to look for comparisons. Ordering them from highest to lowest and putting this week’s numbers on reinforces the comparison between countries, so the primary message that this conveys is this.
However, unless there have been some significant changes between this data and the latest data this isn’t the 16 countries with the highest cases in Europe.
Now when drawing comparisons it is important to compare like for like. It is important to remember that this data shows people who test positive not new infections. Every country has different testing criteria and availability and reporting. If we were to show the same graph for Tanzania for example it would show zero infections since 8 May 2020. They haven’t found a cure, they have simply stopped testing and reporting.
We see the same in our industry. Everybody is really keen to compare customer survey scores with their competitors, however, these scores reflect not just customer satisfaction, but how you ask, what you ask, when you ask, who you ask and most importantly who you don’t ask. Every company measures slightly differently and these differences can have a bigger impact than the differences you are interested in. There are many ways we can gain value in these surveys but using them to benchmark against other surveys isn’t one of them.
But lets assume for a moment that we are comparing like for like testing methods; it is still misleading. It is clear to see that France has the highest cases – but do they really? Surely a larger country should have more cases. What happens if we look at cases as a percentage of population. Suddenly the story really changes.
Now we can see that the Czech Republic, Hungary, Poland, Serbia and Sweden all have much higher infection rates than France but you would have no way of seeing this from the way the data has been shown.
When sharing data we need to ensure that we are clear what story we want to tell, but equally important what stories we are not telling, then ensure we pick the right measures and visualisation techniques.
It is easy to find fault with the work of others, but that doesn’t mean it is easy to do better. The more we practice and the more we and others review our own work the better we will become, so why not see what we can do.
I have taken the latest data (as of 30 March) from the John Hopkins website and overlaid this with population data from the Worldometer website and put this in the attached excel spreadsheet. What story is this data telling you and how would you communicate this? Please share your examples.
If this is a topic that interests you then don’t miss the final workshop of our 2021 Customer Strategy & Planning Conference, Telling a Stories with Numbers at 10:30 on Friday 30th April.
Author: Ian Robertson
Date Published: 7th April 2021