Event Details

Kevin McConway, Emeritus Professor of Applied Statistics at The Open University (http://www.open.ac.uk/research/people/kjm2), will be delivering a training session to STOR-i students on communicating uncertainty to non-statisticians.

As statisticians, we are used to working with and talking about uncertainty. Arguably, that’s our job. But we need to get uncertainty across to other people – fellow researchers who are not statisticians, people and organisations for whom we do consultancy, journalists, the general public. This is not made easier by the fact that some of the numerical information we might need to get across is in itself not easy for others to follow, even without taking uncertainty into account. What’s meant by a percentage change in a percentage, to someone who is not necessarily strong on what a percentage is? What’s an odds ratio, or a hazard ratio, and why and when might they matter to someone not involved in the research? And that’s before we get onto confidence intervals, margins of error, or p values. I shall describe some of the pitfalls, and some evidence on what does and does not work. I will briefly cover some aspects of how the mainstream and other media work, and why it’s important for statisticians to understand that. I will describe some of the research findings on communicating uncertainty and risk. Many important insights come from psychology and graphic design, for instance, rather than from statisticians, but statisticians need to know something about them.