Communicating uncertainty in data without undermining trust

This project aims to identify a range of methods to communicate uncertainty to suit different topics and audiences in a manner that can be easily applied in real-world situations, such as in official statistics, journalism and education.

Uncertainty is inherent in all evidence, whether it is due to inevitable inaccuracies in measurement, because the subject is inherently probabilistic by nature or because of uncertainty regarding scientific knowledge. It is vital therefore that decision-makers know the uncertainty around evidence in order to come to an informed and calculated decision.

It has been found that expert consensus about evidence or a course of action has shown to be a powerful means of reassuring people about the validity of information. However, there is a danger that in attempting to communicate uncertainty around a figure, it may unwittingly undermine a person’s confidence in that figure. The communication of uncertainty even with a consensus could be challenging not least because it is often regarded as an aversive state that people are motivated to reduce.

Finding a means of communicating uncertainty that enhances rather than undermines public trust is an issue of paramount importance. A combination of online, laboratory and field experiments across a span of topics and fields, and analysis of the reactions to uncertainty, comprehension, feelings of trust and decision-making will form the basis of this project. This will inform the team’s aim to identify and further develop methods of communication that will reduce, and ideally transform negative societal stereotypes regarding uncertainty.

After developing a successful methodology and publishing findings the team are extending the project to explore facts and numbers relating to COVID-19. The topic is extremely emotional, and every country across the world is receiving a large amount of communication of data and facts relating to the pandemic. The team will carry out further in depth studies, both in the UK and internationally, addressing the following questions relating to COVID-19 information:

1. Does communicating uncertainty (verbally or numerically) affect different subgroups of people differently?
2. Does communicating uncertainty (verbally or numerically) affect people differently when they have high, versus less high, fear and emotions regarding the topic?
3. Does making a distinction between aleatoric (statistical) and epistemic (systematic) uncertainty change the effect of communicating it?

The findings will be shared with journalists and other professionals.