 So welcome everyone to Guy Goodwin, who most of you will know or have heard of, who's going to give the keynote presentation. So I will give you Guy's introduction. Guy has been Chief Executive at the National and Scottish Centres for Social Research since 2016, with experience of leadership roles in statistics and social research over 20 years. He's been responsible for publications on society, social issues and people in a range of policy contexts from population and migration to crime and from tourism to health and well-being. Guy is a data collection expert, a former director of social surveys and the 2011 census analysis phase and has specialist knowledge in education and geography. His organisation, Natseng, conducts many of the country's large social surveys and longitudinal studies, including the gold standard British social attitudes survey, which provides a window of a time into public attitudes across a full spectrum of issues. And Guy's background is that he studied statistics to post-grad level at the London School of Economics and he's a fellow of the Academy of Social Sciences and the Royal Statistical Society. He was chair of the UK Data Forum, has a strong interest in linking survey and admin data. So welcome Guy, the floor is yours. Okay, hello everybody and I hope you're having a great day. What a wonderful time to be involved in health studies and I hope you realise quite how important you all are at the moment because we are in a period, aren't we, significant public health challenges? Whether that's, you've been discussing some of the issues today, whether that's with the ageing of the population, with our healthy life expectancy, with healthy inequalities and some of the multiple disadvantage we're seeing, whether it's with the challenges of obesity, people's poor mental health, well-being, our sense, I guess, as we all change to, you know, if we weren't operating at home before different ways of working, different lifestyles, it is a remarkable time. And of course then on top of all that you've got this magnifying, amplifying impact on those issues caused by the current pandemic and alongside a suite of whole new issues from, you know, the backlog, the NHS to long Covid, the trauma of working long hours on infection wards, you know, changes in lifestyle again during this pandemic, it really is a challenging time. So it's a fantastic time to be a health researcher or indeed a health policymaker, health user. So you've been discussing some of those issues today. I was asked to come in at the end really to say a few words against this remarkable backdrop on what I'm observing with data collection and in particular social surveys during this pandemic period. And so that's what I'm going to talk about for 20 minutes and then we can open things up for questions. And I guess the first thing I would say is if I sort of split you all the 64 of us into two very unfairly actually into two groups, what I would call the traditionalists in terms of social surveys and data collection and then the modernists. And by traditionalists, I mean, those who lean towards being in favour of a traditional sensors who are rather enthusiasts around the survey and the social survey rather like face to face interviewing and believe that the data collection methods we've been using largely over the past 50 or more years are, you know, both sensible but also pretty relevant today. And then if I call the modernist people a bit like me, although of course the whole thing's a spectrum continuum and it's a bit unfair to group us, but by call the modernist people who feel society has changed a fair amount that believe that we have to look differently at our data collection methods as people live their lives differently, more online and so on, who worry about falling response rates when perhaps we ought to focus more on bias are attracted to multi mode survey approaches, rather like the idea of using administrative data. Let's not collect data twice and see great opportunities in so called big data, sort of unstructured or less structured data bucket loads of it and who want to use the power of technology to harness to mine all of that data for the benefit of society and to plonk it on our policy makers tables in real time rather than a year or two after the event. So it's a crude way of defining us I guess as traditionalist and modernist but we probably see ourselves on a continuum all of us in in that and I guess when face to face interviewing paused and we all stopped doing it at March 2020 with the pandemic. Many of us inevitably asked the question is this really the turning point for social surveys is this when administrative and big data really come to the fore. Is this when we move to multi mode survey approaches for good and face to face interviewing is binned or becomes much more marginal and no longer necessary or practical. And the answer to that question I guess is going to almost certainly be a resounding no. Or at least that is how I see it as we approach the end game of the pandemic. I sense traditionalists are pretty upbeat about things. And right now and it's probable face to face interviewing will be returning for most of at least the complex national surveys including the health ones they're queued up waiting patiently commissioners and there's even advocacy isn't ever a traditional census in five years time not 10 the need for longitudinal studies seems to have become greater not less during the pandemic. And no doubt I think that will continue after it. I mean did you see that wonderful word cloud that showed the estimates of the number of covert 19 cases including asymptotic cases in the UK using social media data. Well no nor did I run being honest. And whatever our individual beliefs and preferences might be for the future. I think today we should inevitably look at how we have actually acted in this pandemic and be prepared to turn to the evidence on what that pandemic experience tells us. And I guess when the numbers and trends in infection cases became so important and grew so much higher. Where did we as a community turn to and and primarily we turn to the O and S covert 19 infection studies as some people call it. Which is effectively a great whopping national survey of the population conducted face to face. And the reason we did it of course is we needed to be pretty sure as or as sure as we could be of where the numbers and the trends were. And although it sits alongside other studies including of course the react study the administrative data that we produce data so on and we look at the suite of measures. It is very noticeable that the national status national statistician made that decision I believe correctly in deciding that's how we were going to play things. Yes it's been used alongside other sources including administrative data. But guess what I mean those administrative data don't collect asymptotic cases. We have had to refine definitions as we've gone along such as the number of deaths of people within 28 days of the first positive test. And the data even increase when we do more testing in affected areas again amplifying a magnifying potential change. And those that have experience of administrative data including many in this conference listening today know full well the trial and tribulations of doing so the comprehensiveness of GP lists. The fact that you know one guidance notes going out to police forces giving instructions can affect your crime survey. Your your recorded crime data or the consequences of departmental staff illnesses on weekly recording of say national insurance numbers of the foreign born. So we understand the deficiencies of administrative data and of course you fully understand it in health. I mean GPs don't know how many of us have poor mental health from their sources. They look at survey sources and that's because it's not all diagnosed and is being handled in a variety of ways by families schools and so on. Do GPs really know how many of us are obese or overweight from their people going to GP surgeries. Not a chance at the moment is the reality. And so you in health studies of course realize some of these issues. Perhaps more interestingly though it is. It's fascinating I think that where on this has turned to when looking for fast turnaround data. You know are the public going to be compliant with vaccinations is our mental health getting worse and so on. And they've turned to surveys as well haven't they and take a bow in particular those of you who work on the opinions and lifestyle. Survey I mean what a tremendous job you've done during this period. And it again reinforces that social surveys have other advantages to in the sense that you can actually tailor things to the precise question you want asked and ask it when you want it. And and more to the point I guess the turnaround is being pretty fast doesn't it and countering one of the perceived big weaknesses of social surveys which is slow turnaround. I mean the other obvious ones being cost and coverage or at least sample sizes and whether we can get geographical and subgroup coverage. So in some ways this pandemic period as we look at it we started it questioning I guess and saying really is this almost the turning point in the end and we coming hopefully to the end of the pandemic. And I think what we have seen is many of the sort of data collection method concerns with other alternatives being seen and reinforced over this period so I just chuck out a small number of observations. The amount of innovation on national surveys and adapting during the pandemic has been pretty remarkable and it just goes to show we can do it. And I would love to see that continue into the future. A lot has happened in collaboration with customers and users and a lot we can be really proud of you know as a community. I worry still though you know it was 2003 when I suggested we move the annual population survey to a multi mode survey when I was at O&S. And you know we're almost 20 years on and have we put in enough investment into data collection methods so we can deliver well enough in in a pandemic or in other scenarios and I sense we haven't. And I think in my judgment we need new impetus on methods and it's funding in this area. I think there's a question of how well we are doing with our networks and engaging with policymakers, influencers and practitioners is a bit mixed. I still think we can do more as a community. We should ask the question are we largely talking to ourselves rather than to policymakers and users. I always tell a story when I was at Ofsted with Chris Woodhead was HMCI in those days and I was his chief statistical advisor but you weren't supposed to go to the top floor and talk to him really if you were working in statistics. And I decided one day I would went up in the left went to see him and I said I'm your new statistical advisor and he said I didn't know I had an old one. But what I learned as from that experience was the afternoon he phoned up wanting some statistics and then he wanted a questionnaire and then he wanted to help with this and that. So there's a great opportunity for us as users, as producers, as researchers, as statisticians to really make an impact by being bold I think with some of the key influences. It's interesting to see that the so-called knock to nudge which is where you knock at the door and quickly back away and you know somebody opens the door and you try to persuade them to answer the questionnaire on telephoning or whatever mode has produced pretty solid response rates actually. It's interesting to me how close they are to the to the normal response rates, but of course it's only slightly cheaper. I think over time we are going to have to increasingly provide a suite of products to commissioners and pick and choose between different modes. And I think certainly if the pandemic keeps going that will become, you know, pretty normal. So really I think that's an important takeaway from the pandemic. I think some of the technological innovations have worked in certain scenarios some haven't. I mean video interviewing seems to have worked rather well on some of the longitudinal studies, the British cohort study I think trialled it. It hasn't worked so well on the health survey for England and you know limited interest among participants of telephone interviewing was offered as well. I think we're using it on the European social survey and it'd be interesting to see what happens there. If there are no alternatives in a path based survey and some of this work needs to be brought together because we've got some very important health surveys coming up, haven't we? We've got Natsale next year, the sectional attitudes and behaviors survey. We've got the adult psychiatric morbidity survey coming up and we really do need to be able to do these surveys in whatever context we are facing. The pandemic I think has shown the real importance of surveys. I mean the mental health of children and young people survey are longitudinal studies like understanding society being analyzed. It's great to see some of the work academics and others have done analyzing some of our data sets, the PPE studies, the opinions and lifestyles survey I mentioned earlier. I think the relevance of surveys has come across, but I think that relevance does depend on us increasingly even though social data look are not like stock market data, are they? I mean they don't need to be in real time, often annually will do, but you know we really do need to turn these data around faster than we have in the past. I think the days of 18 months or two years waiting for survey data are very limited. The most of the methodological challenges and weaknesses we know about social surveys have been proved over this pandemic. You know the absence of good quality sampling frames, poorer or differences in response rates, mode effects and so on. We can see differential response rates. So on the English Housing Survey for example we can see we're getting more homeowners and fewer renters during this pandemic period. You can see some of those biases arising. I mean they can be corrected potentially, but samples become less efficient and our job becomes a little bit more difficult. It doesn't seem to be any easier to focus on harder to reach groups to participate in the studies at the moment. So a whole load of evidence is emerging which you can bring together of course over time to sort of take a stock take almost of where we are on data collection. But my overall feeling is that the survey is going to come out bouncing out of this pretty well actually in comparisons to some of the other methods which really have some way to go yet until we get them to where we want. Finally and just to conclude good policymaking is essential in a democracy and I believe the work you do. I mean social research statistics really does have the power to make life better. I mean health is the big, big example of this at the moment. You're therefore very important people. I think you have a duty in the coming months and years to really argue and make the business cases. The cost benefit analysis and so on in the coming years so that our policy colleagues succeed in not just addressing the pandemic but the whole very bold. Government agenda around leveling up and I think there are big challenges in health there including data collection. You know we need mental health data more regularly than we get them at the moment every seven years won't do I don't think as we move forward. So please leave today with excitement renewed momentum on the importance of your studies your health research and data collections now and in the future and thank you all of course for coming to the conference today. I mean the UK data service is a wonderful thing in the UK and it's great to see you all here today. Thank you.