What’s in a statistic?

How can we provide unbiased data to the public, in a way that can't be manipulated to support specific agendas and how much is it the experts’ role to contextualise it?

We’re very lucky in the UK that we have the UK statistics authority, and that’s because the authority is entirely independent of government and has been since 2009. It means that I do not report to a minister, I report to a board, and I have complete independence in everything that I do. 

That doesn’t mean to say we don’t talk to government about what statistics would be useful, but it does mean that all statistics that come from the ONS, when calculated, are entirely independent.

We also have a code of practice with statistics, which we should be incredibly proud of. Across government, there’s a requirement that all statistics produced are produced independently. And that there’s a policy of no pre-release access, by which I mean official statistics, when produced, are released to ministers at exactly the same time as they are released to the public.  

It is unbelievably important always to remember that underneath any statistic you ever produce, is a person. And in the case of Covid, where that person sadly may have succumbed to this dreadful disease, it is absolutely critical that we always remember we are talking about people in social statistics and that there is context underlying each of those data. I think if we lose that context, then we lose not only the beauty of the data but the importance of them. There’s a real responsibility in social statistics to remember the context in which the data were collected.

How, therefore, does the expert really play a role here? I believe that as a social statistician it’s incredibly important that you understand the context in which you are working – that doesn’t mean in the main, in my view, that you work alone. It means that the very, very best social statistics are created alongside the experts in the field with which they are working and that’s why I believe in government, that the best policy is made when the analysts co-create policy with policy experts who are deep in the world of what can and can’t be done and what the implications of different policies are. Co-creation has to be the most successful way forward and I’ve been privileged in my career to work with people from many different disciplines and to co-create work.

 

What about international efforts to collaborate on this pandemic? What has worked well and what could we improve on for the next global crisis?

It’s so important to have internationally comparable statistics in the future for pandemics and beyond. I recognise there have been challenges, and when we make international comparisons of the number of people who sadly have died of Covid, it is the case that there are different ways in which Covid deaths are reported in different countries and the speed at which those deaths are reported also varies. 

That means that we have, in our work looking at international comparisons and mortality due to Covid, worked on ‘excess deaths due to any cause’. We take the average age-standardised mortality rate in any week over the last five years and then look at whether the age-standardised mortality rate in 2020 or 2021 is significant higher or lower than the average. For the UK when there are peaks in excess mortality rates this broadly correlates with peaks in COVID-19 deaths but does also include deaths that are caused by the broader impacts of the pandemic. But it means it is incredibly important when you make international comparisons you really work hard to make sure you’re comparing apples with apples. 

There are opportunities to continue to do that in the future and I think that’s important. For example, I’ve been highlighting to the Organisation for Economic Co-operation and Development (OECD) that there needs to be work around how different countries have taken into account changes in health and in education during the pandemic in the way they measure their Gross Domestic Product. That is because at the moment there are variations, and it seems to me that if we’re looking at international comparison of GDP we need to make sure we are all comparing the same thing.

It wasn’t an issue before Covid because you have education and health included in the GDP, but if you close schools and healthcare provision changes how do you change your calculations? E.g. What proportion of education are you putting into home schooling? And as we saw in the first wave, there was a massive reduction in the number of people going to A&E, for instance. This is non-trivial, as not all countries are treating those reductions in the same way. 

However, I would say global collaboration in statistics is extremely good. I meet on a regular basis with national statisticians of a number of countries and I speak to other national statisticians very regularly to talk about ways of collecting data – we’re also talking about the ways in which our data collection systems have been impacted by the pandemic. 

For example, we (as well as almost every other country) have had to pull our forces out of the field. That means there is no one, for example, in shops around the UK checking on prices and therefore how do we calculate inflation? There is no one doing interviews for labour force surveys, therefore, how do we calculate unemployment? These are challenges that have had to be faced across the world by national statistics agencies and we’ve worked very hard to make sure we are sharing our challenges and working towards best practices. 

In the future, we need to continue to think about how we can share data in a really imaginative way. I think that much of the data that we have collected should be open and that various data will be incredibly useful across many countries for secondary analysis in order to understand what is going to be the future. 

I also think there is a real role for national surveillance studies to monitor for diseases in the future. There are real opportunities for global collaboration in how best to set up those surveillance studies which will be for public health and be easily allied into any national health emergency that comes. Global collaboration to set these up will be incredibly important. When we’ve been thinking about proposals in the UK, we’ve drawn on colleagues from across the world to get advice.

How have you seen attitudes towards data change during the course of your career?

I’ve been either educating or working around data since the 1970s, and I have to say that over that period of time there have been, at times, a need for there to be a much better understanding of data across the whole of our population. We’ve had some wonderful mathematicians and some wonderful statisticians in the UK, but that’s only the tip of the iceberg. We want every person to have some understanding of data.

I think that many people in our country feel that data means maths, and they can’t do maths. So I do believe that there is an enormous need for upskilling of our country in data numeracy. That said on its own suggests it is just the issue of people not wanting to. On the other side, there is a responsibility that has not always been accepted by educators to make the teaching of data and numeracy really relevant to people. 

For example, I think that we need to make sure that throughout all education, people are studying data and numeracy, but they are doing so in a way that is relevant to their own interest. So, a trainee hairdresser should also be learning data and numeracy in ways that are relevant for running a business. I feel exactly the same about any part of any subject. I don’t think we’ve always, as educators, made our subject as exciting and relevant as it is to people like me. 

Having said that, the last year has really brought home to the nation as a whole, the real need for data. And I think we exist at a moment in time where, firstly, the UK public understand the need for data and it’s just brilliant to hear a prime minister saying ‘data not dates’ – it’s just fantastic. And the second reason is that we’re at a point in time where we have the data visualisation skills to be able to visualise data and communicate them in an exciting and informative way - in a way that has not necessarily existed before. 

So there’s an opportunity to display and visualise data in a really clear way. It’s the role of the national statistics institute, like ONS, to really take that role in communicating data effectively and innovatively - and beautifully - in a way that the public not only can easily understand, but are able themselves to communicate it across their families. 

For a long time in my career, I felt we weren’t moving forward at the pace I would want. We now have the opportunity to move forward at real pace and the tech to help us do it. We just need to commit, as a nation, to improving our communication in data and to making ourselves a data-led nation. 

 

Bonus question: Do you have a statistic that particularly sticks in your mind?

It’s: 27 per thousand babies will die before their first birthday across the world. Of course, one is too many, but this is still seven times less than it was in 1950. And that says to me – that if you really have major international efforts, (for example in immunisation, etc.) then you can make real progress. In my lifetime there has been roughly a sevenfold decrease in that statistic, and it seems to me that’s a cause not for complacency, but for celebration.