How to read numbers: An audience with Tom and David Chivers

6th December 2021

How to Read Numbers is a short, practical, timely guide to the tools you need to understand the numbers we read in the news every day — and how we often get them wrong. This event was an opportunity to quiz the authors, Tom and David Chivers, on reporting numbers, the common pitfalls, and what we can do as PRs to help journalists report stats correctly.

The event opened by highlighting the need for communicators to look for absolute risk (the size of your own risk) rather than relative risk (comparing the risk in two different groups of people). Relative risk is often communicated as large percentages, when the risk to an individual is actually quite low.

“I think at the heart of one of the things we’ve been trying to do with this book is encourage people to look for the absolute risk rather than relative risk”, Tom explained.

David then described one of the motivators behind The Book as being the increase in data that we are exposed to in our daily lives: “We’re so good at reading the written word, and we can understand grammatical mistakes and spelling mistakes quite quickly. But our numerical accuracy is not as good. And that’s a problem. We’re getting all this data that’s being given to us all the time, and especially so in the pandemic. We feel that this is such an important thing, which is why we wrote The Book.”

Tom commented on discussions around COVID data and numbers that get used to describe the impact of COVID. Some report the number of people who have died within 28 days of a positive COVID test. Separately, there is the excess death — the number of people who have died in total, compared to the baseline average of how many people have died at that time of year. Tom made the point that these are very different things that will come up with different numbers and they have different uses.

Other examples of reported research that could mislead included a study showing that shouting swearwords could increase strength. However, the samples size was small, less than 100 in total, and so the chances of random affects occurring is very possible.

They also highlighted the need to think about the sample population, commenting that if you measure average height in a population outside a basketball convention, you may not get the average height of the whole population!

The event was a great recap on why we need to be careful when communicating number and their book takes you through lots of headlines and articles to show how numbers can go wrong in the media. Well worth a read!

© Copyright Stempra