Communication during the pandemic: a reflection from the statistical community

Last month, the Royal Statistical Society (RSS) hosted the first in a series of events inviting members of the statistical community to discuss the response to the COVID-19 pandemic. Attendees gave their views on the effectiveness of data providers, government representatives, the media and commentators when communicating data to the public. A recording of the event can be found on the RSS YouTube channel.

The meet-up, chaired by Sir David Spiegelhalter, included academics, government statisticians, fact checkers and prominent journalists with a range of expertise and insights. Attendees included John Burn-Murdoch (Financial Times), Robert Cuffe (BBC), Will Moy (Full Fact) and Hanna Ritchie (Our World in Data) to name a few. Representing the ONS in person, we had Sarah Crofts from our Covid Infection Survey Team and Sarah Caul from Mortality. We also had Beth Rounding, Sam Rickards, Marcus O’Brien and myself from the Coronavirus Latest Insights Tool in attendance. Others from ONS were also in attendance online. 

There was generally a complimentary attitude towards the ONS pandemic response from speakers but with some useful critique around what we, and the wider statistical community could have done differently.

David and others spoke about the need to communicate uncertainty and some current inadequacies around it. Arguments were made for clear language and openness around uncertainty to tackle the frustrating but understandable tendency for journalists to lead with a single number. A good point was raised, particularly relating to modeled projections (for example, the projected impact of the Omicron variant on numbers of deaths), that the fact we don’t truly know how bad a situation could get is in itself an impactful statement.

Great points were raised about the need for more intelligent curation and presentation of specific measures to best answer questions. Too often, crude metrics such as confirmed case counts and the number of deaths were lead indicators, when excess deaths and infection rates were more insightful. Forums like the daily TV briefings were cited as examples where the data could have been better contextualised if we had sometimes led with more appropriate metrics. This felt like a culture we need to foster as much as possible at the ONS going forward. Good curation and prioritisation of messaging as well as context and understandability are crucial. 

Will Moy in particular spoke about the need to challenge misinformation, and the need to signpost and add context to our data to stop information being used inappropriately. He gave practical advice about avoiding confrontation and not always resorting to policies of censorship, but instead engaging in conversation, addressing misunderstanding and replacing misinformation in a more productive way. This is an area where I think we at the ONS could use the Coronavirus Latest Insights pages to good effect with some myth busting content tailored for citizen users. 

Attendees also spoke about data availability and openness. Although slow off of the line in some respects, it was recognised that UK government statistics in response to the pandemic have been world class. When tracking the pandemic within and between countries, open data and data in a consistent format has been crucial. The gov.uk dashboard along with many examples of publications from ONS and other departments have been successful in this regard.