Interview: UK national statistician on how data has helped the PM – and the public – in coronavirus response

To mark World Statistics Day, Professor Sir Ian Diamond talks to PublicTechnology about how the work of statisticians has been crucial in supporting government’s pandemic response, as well as his thoughts on the media and politicians’ use of numbers, and how the ONS is preparing for the census

Professor Sir Ian Diamond appears on the daily Downing Street briefing in May    Credit: PA Video/PA Wire/PA Images

Statistics have, this year, become a crucially important part of the fabric of daily life.

We have experienced and understood almost everything that has happened to the world in 2020 by tracking the numbers, and deriving from them significance and meaning.

There are the obvious examples of infection, hospitalisation and death rates, which continue to be updated by the government each day. Inherent in these statistics is information on trends in growth and decline; the rolling seven-day averages and week-on-week comparisons in which we are now all well-versed.

More deeply embedded is also information – that has come to the fore in recent weeks – on where the virus is proliferating most dangerously, and how cities, counties and regions differ from one another.

Elsewhere we have come to seek out statistics that answer the questions we have about the broader impact of coronavirus: how is the economy coping? How many people are on furlough? Have people really changed their behaviour? Is our mental health suffering? How much am I likely to be ripped off for two bags of rice and some hand sanitiser?

Helpfully, all of this information is located in one place.

Throughout the pandemic so far, the Office for National Statistics has delivered the data that has become so familiar to government, its citizens, and the media. 

“I think that we will be doing more work in the future to identify where transmission is taking place… and there needs to be more work in linking that to behavioural surveys. Another area we are looking at is the impact of ‘long covid’.”

Working with ministers, and policy and delivery professionals, the ONS has provided a wide range of statistics to calculate and elucidate the societal, economic, and public-health impact of the pandemic.

Talking to PublicTechnology to mark today’s World Statistics Day – an annual initiative led by the United Nations – the UK’s national statistician, Professor Sir Ian Diamond says that, during the coronavirus crisis, the role of the ONS and the wider statistical profession has been not just to support the rest of government, but also the citizens it serves.

“World Statistics Day has the strapline ‘Connecting The World With Data We Can Trust’, and that is something that has been very important during the pandemic,” he says. “It is incredibly important to get the best data, in the most timely way, to government, to [help them] make decisions. But it is also important that the public – who are being asked to make profound behavioural changes – are informed as well.”

Citizens can access information on the ONS website, where a wide array of data related to coronavirus is collected on a dedicated webpage

This includes not only statistics on the spread of the virus itself, but also data sets on issues such as the prevalence of depression among different age groups, the social impact of the pandemic on various demographics, and the fluctuations of online-shopping prices for in-demand goods such as dried pasta, paracetamol and toilet roll.

Incidentally, the latest data set shows that, by mid-August, pasta and paracetamol were still more expensive than they were pre-lockdown – but, after spiking in April and May, toilet paper is now 2.5% cheaper than it was in March. 

Diamond picks out the survey that measures the prevalence of the virus in the population as among the most important data sets compiled by the ONS during coronavirus response, as well as statistics on what proportion of the public has developed antibodies. He also cites a survey that examines “how people are responding to the advice” issued by the government as being an important input for policymaking.

Data has been gathered using traditional surveys, as well as “radical and innovative” new sources such as using location information from Facebook and Google to look at patterns in travel and movement, and the density of people gathered in various places. The work to understand the impact of the virus on BAME communities also saw the ONS – using aggregate and anonymised data – map information gathered from death certificates onto census data.

 


Coronavirus in numbers

The collection of coronavirus-related data sets gathered in a dedicated page on the ONS website contains a wide range of information to help map and understand the impact of the pandemic.

There is, of course, information of infection and death rates; this data can also be broken down by local area, by setting, and by ethnic background.

But alongside the clinical information are surveys and data collections on how the virus has impacted the totality of life, and how its effects have changed over time.

The most recent edition of the social impact survey – which has been published by the ONS on a weekly basis since April – shows that 34% of citizens had worked from home in the last week. This compares with figures of as high as 50% six months ago.

The study also reveals that 98% of the population indicated that they have worn a face covering in the last week as an anti-virus measure – with only 1% claiming that they have, effectively, chosen not to do so.

Just three months ago, these figures were 52% and 45%, respectively.

The ONS has also tracked the impact of coronavirus on the jobs market by collating data from online jobs site Adzuna. The data set published by the government statistics agency measures how current numbers of vacancies compare with the average for 2019.

In mid-May, the number of available positions being advertised online was down by almost two thirds on 2019 levels. The most recent stats show that this has rebounded somewhat, and that there are now 63 opportunities open for every 100 that would have been a year ago.

A one-off data set published by the ONS in August, meanwhile, shines a light on the impact of the pandemic on mental health.

In June 2020, 5.3% of adult survey respondents reported symptoms of depression – compared with 3.3% that typically did so in the nine-month period before lockdown. 

All age groups have seen a rise, but people aged 39 and under represent the biggest spike in both volume and proportion; depression rates in this group have increased from 3.9% to 6.8%.


Throughout the pandemic, statisticians have worked with policy and political professionals to “co-create” a plan for what data should be gathered.

“We are in a constant conversation with colleagues in the Cabinet Office, No. 10 and the Treasury about the sorts of data that they would find useful and we co-create the data requirements in some very, very positive conversations,” Diamond says. “The collection of data has been a really good example of government working together.”

The national statistician cites the example of the ONS working with the Departments for Work and Pensions and Health and Social Care on the construction of a survey to understand the impact on the lives of people considered to be clinically extremely vulnerable – many of whom have spent much of this year shielding. This collaborative process has delivered “some really informative data” for all parties.

Politicians and the public
The role of statisticians does not end once the numbers are handed over; Diamond and his colleagues “have worked very closely with the No.10 communications team” on the presentation of data through channels such as the daily Downing Street television briefings that took place from March to June.

The national statistician post comes with the remit to serve as the head of both the Government Statistical Service and the Government Analysis Function.

And Diamond is keen to point out that “we do not just collect the data, we analyse it as well”.

“We produce a first-cut analysis; but, in any situation, the best analysis is always co-created,” he says.

This is achieved by discussing findings with other stakeholders and, wherever possible, by making data available publicly for further expert analysis.

When asked whether the urgency of the pandemic has required any trade-off in terms of statistical rigour, Diamond responds with a firm “no”. 

He points to data provided on economic indicators and the survey collecting information on changes to people’s lifestyles as two of many examples of data sets that are designed to be produced weekly as a matter of course.

The national statistician is equally assured in his response to the question of whether politicians have used statistics well and responsibly during the pandemic.

“Absolutely,” he says. “They have used them really very effectively. I think the pandemic has also been transformational in bringing statistics to the public. I think the public is increasingly aware of some of the challenges in making comparisons using statistics. And I do think the media have been good at trying to understand and trying to communicate what the statistics are saying.”

He adds: “I am not saying that everything is perfect, and there is always more to do in improving statistical literacy in this country. But I think we have made a really good start; it is incumbent on us to keep improving in communicating with the public about statistics.”

The long-awaited National Data Strategy, published in July by the Department for Digital, Culture, Media and Sport, sets out an ambitious vision for improving how information is used by government and throughout the economy.

“I think it is great, Diamond says. “We need to work together right across government and across stakeholders to ensure that the National Data Strategy has the impact that it deserves to… in building public engagement and public trust, and in increasing and improving the ways in which data can improve the lives of all our citizens.”

The national statistician agrees with PublicTechnology’s suggestion that, supported by some of the prime minister’s closest advisers and ministerial colleagues, data science currently seems to be having something of a moment in the sun in government.

“Ministers have used statistics really very effectively. I think the pandemic has also been transformational in bringing statistics to the public… And I do think the media have been good at trying to understand and trying to communicate what the statistics are saying.”

But he adds that “data science is only one part of the way in which we use data”, pointing once again to the broader statistics and analysis professions, and the importance of all of these working with colleagues across the civil service.

“Data scientists and analysts do not do their best work alone; they need to work with policymakers to make sure that data and analysis can have a proper impact on policy,” he says. “The crucial thing is to work very, very hard to make sure that these conversations are happening – and I am personally looking forward to having lots of conversations with (Defra permanent secretary) Tamara Finkelstein, who is now heading up the policy function.”

Diamond says he is “heartened” by the extent of data-sharing across government, and the ease with which it takes place. He claims that many of the barriers to doing so are, in essence, logistical ones – resulting from the inherent size and complexity of government, when considered as a single entity.

“What I am seeing from government is a real commitment to working across departments,” he says. “Many problems are, ultimately, systems engineering problems – government is a big system, and sometimes we do not know which lever to pull.”

Making census
While the ONS has played a key role in supporting the public-service response to the pandemic, it faces the same challenges as any other organisation in adapting its ways of working to the coronavirus world, and planning for a mid-term future that seems to grow less certain with each passing day.

That is especially tricky when next year will see the agency deliver the decennial census, in which the aim is to collect key information on every single person in England and Wales. Even with next year’s edition being “digital first” – which, for the first time, will see as many respondents as possible encouraged to complete the form online – the ONS still expects to need 30,000 temporary staff to gather data through field work.  (Pictured right are forms for the 2011 census ready to be posted out)

Diamond says: “I am really looking forward to the census – and the job of delivering it has already started. Are we on track to do a very good census? Yes, absolutely. Has it been impacted [by coronavirus]? Yes, because we have had to think through and model many challenges. But I am quite confident that we have thought through the challenges for every kind of impact that could happen. It has been a fantastic performance by the team – they have done a really great job.”

He adds: “I think the census is a fantastic statement for our country; everybody in the population has a voice, and gets counted. And that is a fantastic thing.”

Other priorities for the ONS’s work in the year ahead include “trying to enhance and expand our economic statistics, and taking forward a lot of the momentum on the sharing of data – in a secure way”.

In the short term, the ever-changing demands of supporting coronavirus response will see the statistics agency tasked with providing new types of data.

“I do think that we will be doing more work in the future to identify where transmission is taking place… and there needs to be more work in linking that to behavioural surveys,” Diamond says. “Another area we are looking at is the impact of ‘long covid’, which is increasingly an issue.”

The statistics produced by Diamond and his colleagues seem sure to remain something that the public – as well as politicians and policymakers – are looking at for some time to come.

 

 

Sam Trendall

Learn More →

Leave a Reply

Your email address will not be published. Required fields are marked *