New section makes statistics understandable
‘Statistics? I hated that hard subject in school’, statistician Sanne Willems is often told. And that is a pity, she thinks, because statistics doesn’t necessarily have to be difficult. Together with two former fellow students, she set up a Statistics Communication section for good communication on statistics.
Start with yourself
Nowadays, everyone has to deal with data and statistics, according to Willems. For example, when visiting a doctor or reading a newspaper. 'The problem is that for many people statistics are difficult to understand, but statisticians are not always aware of this.’ Together with former fellow students Nynke Krol and Nikky van Buuren, she founded the Statistics Communications section within the Dutch Association of Statistics and Operations Research. The three want to encourage colleagues to think more about how they can make their research understandable to the general public.
Because that still often goes wrong, according to Willems. 'When it comes to figures in the newspaper, information is often lacking in how they came about, conclusions are very brief, or figures are even taken out of context or misused. It would be good if people knew more about statistics so that they could recognise these kinds of things for themselves.'
This is important because data is everywhere these days. It is often used to make predictions, also in health care, for example. 'So if you don't really know where those numbers originate from, it's likely to affect you,' says Willems. 'Even if you talk to your doctor: the choice of a certain treatment is often based on statistics. It's useful if you understand them. Is the doctor's recommendation specific to you, or to all women, or only women in your age group? It is important to know where it comes from, and also what the degree of uncertainty is.'
To achieve this, Willems, Krol and Van Buuren start at the statisticians themselves. By means of symposia and workshops, the section wants to share knowledge about the communication of statistics. 'A great deal of research has already been done into communication about our field. We want to spread that knowledge and discuss it,' says Willems. In the workshops, skills such as blogging, visualisation of data and creating interactive plots will be discussed. The three also hope to collaborate with other parties, such as journalists, teachers and doctors. ‘The most important thing is to realise that statistics are not so self-evident if you don't work on them every day.
Already at the first symposium of the Section, the enthusiasm for the topic appeared to be high. Within a few days, the registration for Communicating statistics to different audiences was full. CBS gave the first presentation,' says Willems. In recent years, they have been working more and more on their communication. They told how they try to make statistics more fun and understandable with videos and other visualisations.' Research consultant Nel Verhoeven then talked about statistics education and the main pitfalls of statistics research and communication.
Finally, keynote speaker Gerg Gigerenzer of the Max Planck Institute for Human Development in Berlin spoke about risk communication in the healthcare sector. 'According to him, for example, it is better to talk about absolute risks rather than relative ones. If a very small chance increases by 50%, the chance is still very small. To avoid confusion, it is therefore better to also explain what this increase means: for example, that the probability then increases from 2% to 3%. Then you can see that the chance is still relatively small: something that is not immediately clear when there is a '50% increase'. With those kinds of little tricks, you can make it simpler.'
An experiment at the beginning of the symposium for statisticians showed that even seemingly simple sentences can be interpreted in different ways. Those present had to answer how they interpret the statement 'most marbles are red', when talking about a container with different colours of marbles. One half of those present thought that in this case, more than half of the marbles were red, the other half thought that the red marbles only formed the largest group (more red ones than other colours). In this example they are marbles', says Willems. But when it comes to people ('The VVD gets the most votes in Eindhoven'), careful formulation becomes a lot more important!