Statisticians vs journalists: how can we explain data better?

Explaining complicated statistics simply is a daily challenge in my work as a data journalist. My work needs to be both factually correct and easy to understand – two requirements that can be conflicting. In this talk, I will show you how I use data to tell stories to a young audience, and which dilemmas I face in the journalistic process. Together, we’ll discuss questions like: how can we explain statistics better? How can we use smart metaphors? What role does visualisation play? And what ‘shortcuts’ is a journalist allowed to take without angering the statisticians (too much)?

Wouter van Dijke, Data journalist at NOS op 3
Wouter is a data journalist at NOS op 3, a news channel aimed at people aged 18-35, specialised in explaining complicated news stories. As a data journalist, his job is to explain data, and use statistical methods to do journalistic research. Previously, Wouter has worked for a number of news organisations and taught at the School of Journalism of the University of Applied Sciences Utrecht (HU). He has also created funny viral websites like Woonplaatsguesser and CatGPT.

Forensic evidence evaluation - statistics in court

In criminal cases we use statistics with a twist - 1) there are two hypotheses, but neither is the 'null' and 2) the judge, not the scientist, is making the final decision. Forensic scientists thus often provide a likelihood ratio - specifying how much more likely the evidence is under one hypothesis than the other. But how should the judge combine this number with the other (non-quantified) evidence in the case? And how can we best help them to do this well? I will discuss some of the successes, challenges and informative pit falls in this fascinating interface between scientists and lawyers.

Rolf Ypma, Principal scientist and Forensic data scientist at Netherlands Forensic Institute
Rolf works on criminal cases and scientific research in the forensic data science domain. He is interested in how forensic statistics can be better communicated to lawyers, and open to academic collaborations on this fascinating topic.

Basing decisions on data: understanding is key

The purchase of rolling stock unit (or: train) costs NS multiple millions of euros. After the purchase, the maintenance of these units is yearly a big expense for NS. Thus it is very important to make efficient decisions when it comes to rolling stock. We don’t want to purchase too much, but also not too little. And the rolling stock that we own should be at the right place at the right time, so that there are enough seats at moments when there are a lot of passengers that want to take the train. To make these decisions we rely heavily on passenger forecasts.

Just as all forecasts, these forecasts have uncertainties. We need to take into account these uncertainties when making decisions. Understanding and interpreting these numbers is not easy by itself, and there are even multiple different quantities with their own uncertainties that should be combined in decision making. I will discuss how we (try to) make the numbers understandable for decision makers so that they can feel comfortable in their decisions. I will also give some examples of times that we could not explain it well enough, demonstrating the importance of not only having good forecasts, but also communicating them well.

Simone Griffioen, Business Consultant at NS
Simone works in the department where decision making models are developed for planning and operational purposes at NS. She is educating on the use of these models and working on the implementation of these models within the business processes. Simone likes to help other people understand complex matters to enable them to make well informed decisions. In her spare time she enjoys rock climbing.