 So, I work for the UK Data Service and so for anyone who doesn't know, the UK Data Service is funded by the Economic and Social Research Council to provide a single point of access to a wide range of secondary social science data, but as well as providing the data and providing access to data, we also provide a whole series of support, training and guidance to help people use that data. So that's part of the reason why we are supporting to run this workshop today. And the main types of data that we hold come in different forms, so we have aggregate data, so that's the sort of data you'd get from census data or some of the international data. We have microdata, so by microdata I mean data about individuals, so lots of that comes from the surveys. And then we do have some other data as well, such as mixed methods data and some qualitative data, or quite a lot of qualitative data actually. But today I'm focusing on the survey data that we hold, so the microdata that comes from UK Surveys. So survey data is a fantastic resource because it's been collected in a systematic way and so it's a representative sample, so we know that the data that we're having is representative of the population as a whole and it's microdata, it's data about individuals. So bringing all that together, survey data provides a very valuable resource that we can use to provide context to some of the work that's going on in our communities. And so I think social survey data is particularly valuable to the VCSE sector because the services within that sector are for the communities and society, but what the funders want is to demonstrate the need and the impact of those services, so they need the context and that's what the data can provide. And the UK Data Service is the resource where you can get that evidence from. And the reason why this sort of data is potentially particularly helpful for the VCSE sector is because it's already been collected. So lots of third sector organisations, as we've talked about already, might not have a dedicated research team or that research team might be one person or it might be the person delivering the service that also needs to provide the evidence. And so if you have data that's already been collected that you can reuse, then that's going to help your services. So some of the pros and cons of reusing data is so some of the data sets that I'm going to show you shortly would be impossible to create on your own as a small from a small resource. They've been collected by large data collection teams. So then for reusing that is very cost effective to anyone doing secondary analysis or reusing that data. The ethical considerations have all been considered already by the people that collected the data, so you don't need to worry about the ethics of collecting the data. And there's no need to recontact the people that were interviewed. And you can reuse the data to make claims of your own. So you can use that data in your own way. Some of the cons is kind of opposite to all the pros, really. You didn't collect the data, so you need to understand how it was collected and why it was collected so that you interpret it in the right way. So it takes some effort to get to know the data. And there still might be some ethical issues that limit the amount of data that you can access. For example, we don't often release a data to load geographies because that might make the data sensitive. So you might not be able to reach or access the data that you're specifically looking for. So you might have to rethink how you use the data and how it can support your service. So the data might not match exactly the research questions or the projects that you have in mind. But you can rethink about how that can be used. So some of the survey data that we have at the UK Data Service just to give some examples, one of the big ones is the British Social Attitude Survey. I quite like the survey. It's interesting and it is funded by the Economic and Social Research Council, but different people also put forward questions which are answered within it over the years. It is carried out every year and it goes to 3,879 people and it's personal interviews and there's follow-up self-completion forms so that people might feel they're able to be a bit more honest in those if they're not asking questions about potentially sensitive issues. And so the type of things that we find from the survey is opinions and views on things like same-sex relationships, austerity, Brexit, all these sorts of questions are asked in the British Social Attitude Survey and the data is collected by NACSEN, the National Centre for Social Research. And this is what the catalogue page looks like on our website. So this is the UK Data Services catalogue and in there you'll find we've got thousands of data sets and each data set has its own catalogue page where you can go and find out the details about the survey. And I'll show you that in a bit more detail shortly. Another survey is the Health Survey for England. So this is a bit different. It's funded by NHS Digital and it's used to shape health policy and improve health services. So it's finding out about the health of the nation. It's a slightly bigger survey than the British Social Latitudes. It has 10,250 interviews, but that covers adults and children. Again, it's administered face to face and there's a self-administered questionnaire as well. But they also do physical and clinical measurements and tests so they actually weigh people that sometimes they take blood samples. You know, these sorts of things are all measured as well. And this is run by NACSEN as well. But UCL work with them to produce the survey and the things we find out from this survey includes information about obesity rates, healthy living, gambling, all those sorts of health issues are covered in the survey. And this is what the catalogue page looks like again. So each survey has its individual catalogue page on our website. Another one I just wanted to touch on was the Quarterly Labour Force Survey. This perhaps is one of the biggest surveys. You can see it's got 69,733 interviews and this was for the second quarter of 2020. It's directly government funded and it tells us a lot about the current labour force. So when you hear about unemployment rates and things like that, this is where the information is coming from. So it comes directly from the Office of National Statistics. So it's very pertinent at the moment. It's the survey team are very busy collecting information about the impacts of the pandemic, but generally it follows things like joblessness, unemployment, redundancies, any changes in the labour market are all picked up by the labour force survey. And that's what the Quarterly Labour Force Survey looks like. And I'm going to show you that a bit more detail shortly. But just to reiterate as well that we do have a lot of resources and support available on our website to help access this data and the surveys. We have a whole series of guides and videos and webinars. So we are here to support you, access the data and have a look at the website afterwards and you'll find out more about that. But before I go on to show you a bit more of the website and we have a little activity at the end, much like Dermot did. I'm going to pass over to Tom. From Independent Age, who is going to give you an example of how their service has used some of our survey data in their research. So thanks very much, Tom. Hello there. Let me just get these slides up. There we go. I assume everyone can see that. So, yes, hello, my name is Thomas Wilson. I'm a senior policy officer at Independent Age and I'm going to talk about our project in focus, which we ran throughout a lot of 2019. We published earlier this year, where we looked at using societal data, but also qualitative data through commissioned work. So both Qual and Quant. I'm going to talk more about the Quant side because that was the side I was more involved with, but I'll try and touch on the Qual as I go through. I'm not going to talk about the results too much because they're all kind of published online. I'm going to talk about more about the structure of the project and how we went about doing it. He says unable to move his slides. Ah, there we go. So, just quickly about Independent Age. So we're an older peoples' organisation, national older peoples' organisation. We focus on ensuring that older people have the opportunity to live well with dignity, choice and purpose. We have a national helpline, including an advice team and a national befriending service as well. We also have a policy and influencing team, of which I'm one. So in terms of the origins of the InFocus project, we experienced a big, big expansion of our policy and influencing team in 2018, went from about three or four people to about 15 or 16. So we wanted to undertake a project that would underpin our policy work going forward and give us an evidence base of our own and our own kind of base of facts and figures that we could use to get insights into the groups that we were interested in and that we thought were kind of being unheard. We were particularly concerned at how 65 plus is often treated as a homogenous group by decision makers and published statistics often tend to focus on the 65 plus age group. It's quite a modernized age group without looking at differences that might exist within that. Particularly also, we were concerned about the subgroups of the 65 plus group being often poorly represented in published statistics. It's very hard to get statistics on 65 plus BAME people, for example. You can maybe get statistics on 65 plus, you can maybe get statistics on BAME, but getting them both together is very, very difficult. Although the data is there, it's just not kind of published in that way. And we wanted to look at three key themes that were important to us. So health and well-being, financial security and social connections. So it's the first stage of the project because it was going to be a pretty big project, so we wanted to make sure it was going to work. We did some scoping. We asked the national development team for inclusion at ETI to test our hypothesis looking at academic literature. So they conducted a scoping review and literature search, which took a few weeks. And their findings confirmed our hypothesis. So in academic literature, subgroups of older people are very under-searched and that the 65 plus age group was used quite a lot in a kind of a modernized way, not really looking at those differences between the smaller age bands. The first thing we did, as we were writing the tenders, was just run it before for a positive project. So we set up a group of older people to advise them. I'm obviously focusing on the qual side. Tom, you're breaking up a bit. I don't know if your microphones become a bit loose. Yes, maybe. Hang on. Is that better? I think so. Okay, I'll continue. Tell me if there's another issue. We'll do. So we set up a co-production group of older people. They mainly inputted on the qual side. We found them to be really, really useful and the group reported quite a positive experience as well. I can talk about that a bit more in the chat if people are interested later, but it was quite interesting and they had quite a big impact on some of our decisions. So we put out our tenders and we got bids back and we were faced with some initial commissioning questions. So firstly, academic versus market research. So a lot of our qual bids are obviously market research, but with the quant, we had market research bids but also academic bids as well from universities. Related question, both projects to one provider or keep them separate? So basically, do we give both the qual and the quant to a single market research provider and kind of have it all done by one person or do we split them up and give one to a market research and one to another market research or another academic? Which qualitative research method to use? There were lots of different ones. Luckily, as I'll say in a moment, our co-production group helped us with that one. And the quantitative methods, sorry, I've misspelled that on the slide, the quantitative method, should we look at lots of different data sets? So look at hundreds and hundreds of different surveys and try and get lots of small insights or one deep dive into one very large data set like the kinds that Patty was talking about earlier. So in terms of our answers that we came up with those questions. So we ended up going with a market researcher for the qual and the academics for the quant. The main reason we decided to go with the academics is partly because we'd worked with them before. We had quite a good relationship with them. We quite liked their methodological approach, which I'll talk about in a second. They were also cheaper with quite a lot of the other bids. And also we quite liked the idea of having a kind of market researcher doing a quant and then the kind of weight of academics looking at our quant. We thought it would give it perhaps a bit more authority. In terms of our qual methodology and who we went for with the qual provider, our co-production group were actually instrumental because they expressed a strong preference for a particular qualitative method which was a kind of photography task followed by in-depth interviews. So that made choosing the provider quite easy. And as I said, we went with that one deep dive to one huge data set, which I'll talk about in a bit. So that's just an overview of what we ended up going with. So we had humankind research conducting our qual which consisted of 45 in-depth interviews and follow-ups as well as a photography task and city university looking at our quant which was a deep dive into a big data set. So the first decision on the quant side was about which data set to use. We saw some examples from Patty earlier. So we were looking at two other ones. The first was the English Longitude and Study of Aging, Elsa. We were tempted to go for this one because it is the go-to and quite comprehensive data set about older people. It's a data set that looks only older people. So the questions are quite focused on older people's issues and can give us a lot of flexibility. But we decided not to because of some of the cons. So because it's only older people, you can't get that nice cross-generational comparison looking at younger age groups, which we quite wanted. And also there were some issues with special licenses where to get some aspects of the data, it would have taken us quite a long time to apply for those licenses. So in the end, we went with a different data set, the Understanding Society or USOC data set. So USOC is a giant survey. It's the largest sample size of older people than Elsa, including the oldest old because there are tens of thousands of people who have done it in the survey. The city team had already worked with USOC extensively, so it was quite an attractive option. And as I said, most importantly, it allowed us to look at the same questions asked to older people and younger people so we could get that nice cross-generational comparison which I'll show you a few examples of later. But we did have to bear some cons in mind mostly because the questions obviously are not designed with older people at nine because they're asked for everybody. And because the survey is very big, the number of questions asked on each topic is lower than Elsa because the survey covers quite a lot of ground. The next thing that we had to come to was choosing our definitions of our subgroups and our indicators so that our partners at City University could run the research. This was more complicated than it sounds because each subgroup needed defining within the data. So for example, to take one subgroup of carers might sound relatively simple, but someone caring for one hour a week and someone caring for 50 hours a week, those are very, very different people who are going to have very, very different experiences of caring and their life experiences are going to be totally different. So trying to look at both of them in the same group can produce you with data that's not entirely useful. So we had to break it up into different kind of chunks of, okay, this number of hours caring, this number of hours caring, this number of carers caring while maintaining our sample sizes is a good enough level that we could get some good insight. And we had to do that with all of the subgroups. So it took some time. We also went about choosing 17 indicators for each theme. There was no special significance to 17. It was just about as many as we could get away with. And that's involved looking at a long, long, long list of all of the questions asked by understanding society that might cover those theme and basically culling them down to the indicators we thought were most important. A key learning from us from this project was just because of our unfamiliarity with the dataset and of doing this kind of research before this slowed us down a bit. I think this could this exercise could have probably been done quite a bit ahead of time than we did it. So that was a key learning for us. This is an example of some of the analysis or example of the analysis that came back. So this is one of hundreds of graphs that's the university presented to us. So I thought this was quite a good illustration of what we were trying to do. So this graph is showing a kind of relative measure of poverty, essentially all relative measure of income. So we've got poorest households in red here, second poorest middle incomes and so on. And over on the left side of the graph we have these broad age groups, the young adults, mid-working age, late-working age, older age. Sorry. Over here we have these smaller four-year age bands. I thought this was quite a useful graph to show you because if we look at this older age group here, 65 plus, they seem to be doing okay. If we look at poorest households, they're one of the least likely to be in that poorest households category, certainly when compared to mid-working and young adult. But then when we look at the four-year age bands over here, we can see that this average is actually disguising a lot of variation within that 65 plus age bracket. So if we look at the 65 to 69 year olds, well, they're one of the best stuff of everybody in terms of likelihood to be in that poorest income group. But then if we look at 85 year olds, they're one of the worst stuff along with 16 to 19 year olds and one of the most likely to be in that poorest income bracket. So, you know, we were quite happy with this data because it showed all kind of partly proved our hypothesis that these larger age group categories are disguising variation that exists within those smaller age groups. Tom, one minute please. Okay, thank you. I'll try and speed up. This is an example of one of our subgroup presentations that came back. So this is the without children group looking at indicators for social connectedness. So you can see all of our indicators down the side. Purple bar is without children. Gray bar is kind of all older people average. And then when there's a black heading, that shows that there's a significant difference between the two. We really liked this because it allowed us to just kind of add a glance, look and say, okay, there's a significant difference here. There's a significant difference there. That's interesting. Let's dive in more in the data and into our quail and look at that in a bit more detail. We also did a kind of phase three where we took the quail and quant, got them both together, got them to present their results to each other and then went off and gave them some follow up analysis to do based on what each other had presented. This was an attempt by us to kind of smooth over one of the problems with giving both things to different providers where they can be a little bit disconnected. So we tried to join them up a bit more at the end. And this is an example of what we ended up with in our final report. And I quite like this one because it shows us presenting the different data alongside each other. So you've got the quant data being presented here and kind of infographics and highlighted statistics, the quail in these case studies and quotes, and then our analysis over here on the left. As I said, I quite like this because it just presents everything together. It kind of highlights the mixed methods approach and it just gives the work kind of a bit of authority and then we're not just making it up. We have the kind of both the quail and the quant data to back up that analysis. And that's the end of it for me. Hopefully I'm relatively bang on time. I'm happy. I know I kind of went whistle through, whistle stop tour through there and there's a lot to cover. If anyone would like to follow up directly with me and my emails on there, you can also find more about our in focus work on the link attached. I think these slides are going to be sent around. And yes, happy to answer any questions in the chat as well. Sorry for the microphone breaking up, by the way. That's great, Tom. Thank you very much. So I'm just going to show you now a little bit about how to find some of this data and how to start to understand it a bit more yourself. And then much like Dermott said, we're going to have a look at giving you a task to try and find some information. So we're going to look firstly at the UK data service catalog and then look at the survey documentation and then look at how you can start to explore the data sets about yourself as well. So I'm going to try and share with you the website. So just bear with me for one second. This one here. So this is the UK data service website. Hopefully you can all see that when you get to the landing page for the website, the big black box in the center of the screen, that's our data catalog. So if you know what data set you're looking for, you can type it in there or you can type in a keyword. You can also go up to the top of the screen to these tabs and you can get data via key data. So it's the big data sets, the main key data sets that we hold or you can look for data by theme. But I'm just going to show you as an example, the labor force survey. So this, as I mentioned earlier is one of the biggest surveys that we have. I'm going to look at it by series because it's a series of surveys. So this is the labor force survey. It gives you a general abstract of it here. And then if you access the data, it tells you all the different types of access that we can have. And I'm looking at the quarterly labor force survey. And it's a bit complicated at the moment, the labor force survey. I mean, it's a very complex survey just generally, but since the pandemic, they've been releasing data on a rolling quarter. So every month they've been releasing three months of data, but generally they do it on calendar quarters. So we're going to look at April to June 2020 as an example. So that's the second quarter of 2020. And so this is the catalog page that I mentioned earlier. So on the catalog page, you'll find all the basic information for the survey. So you see you've got every survey has an individual survey number. It tells you how to access it. Safeguarded means that you have to agree to our end user license before you can access it. But that's a relatively quick process to sign up and access data. Further down, it tells you what topics are covered in the data set. It gives you an abstract for the data. And then I quite like these bits down the bottom under coverage and methodology where it summarizes sort of the key details of the survey and how it was collected. So you can see when the field work took place, where it took place, who the data was collected from. So in this instance, it was from individuals as well as households, how many people. So we can see that 69,733 cases there. And it tells you how it was collected. So face to face or telephone interviews. The way the labor force survey is collected, they go back to the same people every quarter for five quarters. The first time they see them face to face, then they bring them up to follow them up at the following quarters. But every quarter they're bringing new people into the survey as well. And then to look at documentation, that's the second tab here. And in part, I'm showing you the quarterly labor force survey because it is one of our more complicated and comprehensive surveys. So there's a lot of documentation attached to it. So most surveys won't have quite this much detail and it can be a bit overwhelming. But the things to look out for in the documentation list is things like the questionnaires. So this will give you the questionnaire for the survey. So this is the survey used or the questionnaire used by the interviewers when they go to the participants and talk to them and ask them all the survey questions. You can see down the side, there's all the questions. There's 223 pages to this questionnaire. So at the beginning, when I was saying, you know how valuable the resource is and how it isn't something that most people collect themselves, that's kind of highlights that point. For example, if we've gone down to main job here and it asks, did you do any paid work in the last seven days? So yes or no. And this working word here, that's the shorthand for the variable that you'll find in the data set. So that's the name of the variable in the data set. But if you're trying to find what was asked in the survey, I often use the find function. So control F or command F on your computer to bring up find and then say you want to find out what they asked about ethnicity. You could search for ethnicity and it will bring that up in the questionnaire and you'll find out what questions were asked. So you can see here, I've got 13 matches for ethnicity and you can go through and find out what was asked. And so that's the questionnaire and it can be explored in that way. We can read through the whole document. The other documents that are interesting to look at include the user guides and the technical reports. But this is going back to the documentation tab. The one that comes with all our surveys is the data dictionaries. And the data dictionaries are quite good because they give you all the itemized variables from each data set that you download. Often downloads as a zip file and that comes, I'm not going to download that now because I've already downloaded it and I'll show you what it looks like because I've got it ready for you here. So this is what the data dictionary looks like. At the top, it tells you what the data file was, how many variables in it. So in this case, there's 809 variables for those 69,733 people. So again, I find the easiest way to search for information in the data is to go to the data dictionary, find something and for example, use ethnicity. Again, you select on that and it automatically takes you to the variable for ethnicity and you can find out what's been collected. So that's how to use the documentation or how the documentation can be quite valuable when you're exploring the data. But then the last way that you might want to look at if you actually want to see what the data tells you is by going to the access data tab. So if you go access data, you can add the dataset to your account and load it up or for lots of our data, you can access them online. So I'm going to do this. This is a tool that we've got called Nestar. So it's an online tool where some of our datasets are and you can explore the data from here. The quarterly labor force survey, as I mentioned, you have to agree to an end user license to look at that data. So we're not going to look at that today. Instead, if you go up the top, you can see on the left hand side, you've got all the research datasets and then you've got some unrestricted datasets and some teaching datasets. So the unrestricted datasets, more limited datasets that have been put together just really for the purposes of teaching. And if we look under the unrestricted teaching datasets, you can see there is an example of the labor force survey there. This data is actually from 2015. So it's a bit older now, but you can open it up and it tells you a bit about the survey and then going under, drilling down to variable descriptions and household variables, you can straightaway look at some of the statistics that are coming out of the survey. For example, you can look at full-time and part-time and main job and automatically it's giving you the statistic for that answer from that survey. So we can see from the quarterly labor force survey for January to March, 2015, 30.5% of the sample were working part-time. This is of the people working. And so the other 69.5 were working full-time. But for this question, we've got 5,000 respondents that it's actually about a little under, about 20%, 25% of the sample, pretty much 20%. We're not actually working. So that they are not included in this question. And so you can look at the different categories and so you can go to qualification and all the statistics come straight up. Using Nesta, you can also do basic cross tabulation and even some regression analysis. So Nesta is quite a nice tool just for exploring the data and looking at some of the percentages of the distributions of the data. So what I've shown you here has been the catalog page, the documentation and Nesta. So I'm hoping that you might be able to explore some of that yourself now. So again, if we go to the website where we've got all the information for this workshop today and if you scroll down to societal data and there's the activity down here and I've put in some links for the catalog and for Nesta. And if you follow that through, it's the scavenger hunt thing that Dermot's introduced. I've got six questions in there that I'm hoping you might be able to find the answers to. So I'm going to give you just under 10 minutes to try and find those answers. Feel free to put questions and comments in the chat in the question box and we'll come back in 35 minutes past to share your answers and see how we got on with that. Right, so hopefully you've found some of those answers on those different pages. I realized that you had to go to like three different places to find the answers. So it's a bit of a challenge in the time that you had but maybe put in the chat some of your answers to those questions. So the first one was what is the UK data service study number for the crime server for England and Wales, 2018 to 2019? And that's not the answer that you've got there. I've got the wrong answer on that page. There we go. Those are around the wrong way. But the study number is the 8608, 8608. So hopefully some of you got that. See, I've got the chat coming up here. Yes, we've got that answer coming through. So that's brilliant. And then the next question was about how many adults were interviewed in the survey? So how many adults were interviewed? Hopefully the answer is 34168. So that was a 163. That was from right down the bottom on the catalogue page under the methodology and number of units and all that data right at the bottom of the catalogue page. The next question, how many variables in the data set? Oh, that's good. Yes. And so we've got the same answer here. You'll see number of cases, 3,401, 34,163. But we've also got the number of variables which is like an astonishing 2,215 variables. It's an awful lot of variables. But that includes all the derived variables that the data producers would have made. So it's not just the questions asked of the participants, but it's then how they've been used to make more data out of those questions. So finally, hopefully some of you were able to, first of all, asked what was the variable name for the question about vandalism. So hopefully some of you were able to find that in the data dictionary. And the name I was looking for was Vandals. And then hopefully some of you were able to get into Nestar and find some statistics coming from the survey. And we're looking at a different version of the crime survey for England and Wales then. But you should have hopefully come up with something like this. So what percentage of people think that the lack of discipline from parents is the one main cause of crime in Britain today? I found this quite interesting actually when I looked into this for this presentation. So that's the most popular category for what people think is the cause of crime in Britain today. So we had 31% of people thinking that lack of discipline from parents was one of the main causes of crime in Britain today, closely followed by drugs. And then finally asking people what percentage of people are fairly worried about being physically attacked by strangers. And so this was the distribution that we've got here. And quite a lot of people are fairly worried, I think. So that's nearly a quarter, 23.4% of respondents were fairly worried about being physically attacked by strangers. So hopefully you managed to find some time to explore some of those dataset, that dataset in Nesta. And that just gave you a taste, just a very quick brief taste of how to find data and what you can find out from it. So that kind of brings us to the end of the case studies and the different sectors. So we're hoping just to use the last 10 minutes for further discussion and any questions that you have that's been raised by all three of the sections. So we've got the question and answer box so you can use.