 Hello, everyone. Welcome to this webinar, an introduction to the British Social Attitude Survey. My name is Sarah King Heal and I work for the UK Data Service in user support and training, and I'm based at the University of Manchester. I'm going to be telling you a little bit later on about how you access the BSA. The main speakers, however, who work for NATSEN are going to be Miranda Phillips, who's a research director working on the BSA, and Eleanor Taylor, who's a researcher for NATSEN, also working on the BSA. OK, now I'm going to hand over to Miranda, who is going to give our first part of the talk. Hello, everyone. So, thanks for the introduction, Sarah. OK, so we're going to talk to you today about the British Social Attitude Survey. We're going to cover a few things around that as well. So, this is just to explain everything we're talking about in the webinar, starting off with who NATSEN is, and before we get into the details of the survey, we want to reflect a little bit on why we would want to measure attitudes in the first place. I'll then talk briefly through the methodology for British Social Attitudes, and I'll hand over to Eleanor to give an example of the kind of analysis that you can use BSA data for. We'll finish off by talking about dissemination of findings and some practical tips about carrying out your own analysis using BSA, and then we'll hand back to Sarah who'll talk through accessing the data from the UKDS. If you've got any questions while we're going through, just type them in as we're talking. We won't handle them until the end, but when we get to the end, we'll try and cover all of them. So NATSEN Social Research, which runs the British Social Attitude Survey, is the largest independent social research agency in Britain, and we're a not-for-profit organisation. We have lots of different clients and funders that the majority of our work comes from central government clients. We do lots of different types of research, probably the best known are our big surveys, but we also do qualitative work, evaluations, secondary analysis, and in lots of different policy areas such as education, welfare, work and health. If you're interested in finding out more about what we do, our website's a good place to start. It's got lots of information about our reports and our current projects. Okay, so I wanted to start off just reflecting on why we'd want to measure attitudes. There are some people, and it's probably fair to say this used to be the case more than it is these days, but some people who feel that attitudinal research isn't legitimate. These fall roughly into two types of concerns. The first is that attitudes don't matter. Under this heading that would encompass the view that behaviour and experience is more important. It's not what people say, but what they do that really matters for policy makers. A related point, attitudes and behaviour aren't always consistent, and that's true. So if you know what someone thinks, it won't necessarily tell you what they'll do. And lastly under this heading that we don't want government by opinion poll, and this is really a worry that knowing what people think may mean and that government just does what the people want, rather than taking tough decisions and leading from the front. So the second group of concerns is really about how you go about measuring attitudes, and the feeling that it's very difficult. And it's true that there are challenges associated with attitude measurement, but we'd say the same is true of any survey, and if you think about measuring something that you might call an objective fact, such as alcohol consumption for example, you can see that actually there will be challenges around the wording and the context and the order that the questions are asked in the questionnaire. And it's true that for attitude questions especially, but again I'd say for all surveys, that how you word your questions will have a crucial impact on how people respond. And what we'd say is that we want to acknowledge that and then address it by the way that we design the survey. So these are legitimate concerns, but it won't surprise you to hear that we would argue that attitudes do matter. Our attitudes and values are an important part of our social world and if we can't measure them reliably and describe them, then we're going to miss out on a big part of the jigsaw. So see it not as an alternative to behavioural and factual surveys, but as a complement to them. And they matter to policy makers too. Policy makers need to be aware of what people think and whether it's changing. We know that attitudes and legislative change can go hand in hand. So for example, we know they've been changing views about same-sex relationships and that has really led the way for allowing legislation for same-sex marriage. That wouldn't have been conceivable 30 years ago. We also know that policies can fail if they hit attitudinal barriers. It's not just about what people know and what they care about. It's also what they think and feel. Another argument for measuring attitudes is that if we don't do it, that people will guess and make assumptions about what people think and they'll often get it wrong. We also know that when that happens there are competing claims and that there are dominant voices, politicians, journalists, people with access to the media and it will be their voices that get heard. What we want to do is make sure that the public view is collected and known about. And when it comes to solving the measurement and question design problems, that's something that we address through spending a lot of time and effort crafting our questions, testing them and refining them so that we have measures that are robust and reliable. Let's turn now to looking at the British Social Attitude Survey. I'm going to talk about an overview and also some methodology. It's one of Britain's most well-known and frequently used surveys and one reason for that is that it's very highly regarded. It's seen as an independent and authoritative source of information about contemporary British attitudes and how these are changing over time. The quote on the slide from the FT just really illustrates that. This is your at-a-glance overview of BSA and I'll talk about some of this in more detail in what follows, but just to quickly run through it. The British Social Attitude Survey was initiated by Natsen and it began in the early 1980s and we run the survey annually. It measures social and political attitudes and moral values. The key aim of the survey is to look at change over time. Many of our questions are repeated so that they build up a time series, but it's also important that we keep the survey up to date so each year we include new questions that reflect current areas of interest. In terms of methodology, it's an annual cross-sectional survey so it's important to note it's not a panel. It's a random probability sample and I'll say more about that later. We interview roughly 3,000 respondents each year. That's adults aged 18 or over living in Britain, that's England, Scotland and Wales. We don't cover Northern Ireland, but it's worth just telling you that we've got a sister survey in Scotland so if you're particularly interested in the policy context there, Scottish Social Attitude has a much bigger sample for Scotland. The interview itself comes in two parts. There's a copy or face-to-face interview and also a self-completion booklet. We have a range of different funders for the survey and these change from year to year. It includes the big government departments, charities and also grant-giving bodies. We've got a wide range of users from government and charities to other social researchers and journalists. It's also used a lot by academics and students. I used it in my university course and perhaps some of you did too. Just to give you a flavour of what's in the survey, I want to show you some example topic areas. First of all, we cover some of the major policy areas such as health and education and housing. These are typically funded by government. We also have more academic topics, if you like, less regularly on the survey, often funded by the grant-giving bodies and research councils and that gives you a flavour of the kind of thing that you might cover under that heading. Some of these topics are on the survey most years, others are more occasional. The important thing to note is that when you're trying to understand what BSA is about, the topics change from year to year and it all relates to who's funded it and what they're interested in. You always need to do some scoping out to work out which questions are on each year. Just going to have a poll to ask you how long you think it takes to do the fieldwork for BSA. Just to remind you, it's face to face and we get about 3,000 interviews. The answer options you can see in front of you, one month, two months, four months, eight months. Okay, thank you. We've got an equal split on the winning answers which are that 38% of you think four months, 38% of you think eight months and the rest are saying shorter durations. The answer is four months which is quite a long time in social research time scales but it's important for the way we collect the data and I'll say a bit more on that later. So thank you for taking part. I'll show you what the survey timeline looks like. Around the start of the year we're finalising our funding and then we start with question design and sampling. We continue question design and pilot our new questions and then we program the final questionnaire. And then here's that four month period for fieldwork. So between about June and September we're in the field and then after we've collected the data there's a process of editing and derived variable creation before we deliver data in December. And then we spend at least half probably of the year following the survey, doing analysis and reporting and various different types of dissemination. Currently, talking to potential funders for the 2017 survey and beyond so if any of you are on the commissioning side of research or if you've got an idea for a grant funded project that would be a good fit for BSA if you get in touch with us we're very open to new funders and collaborators on the survey. So I mentioned that the first thing we do really in the design, in the whole process of the survey is question design. And this is for any new questions. We also run repeat questions each year to continue our time series. And the process that we go through is longer and more detailed than you'd find on most other surveys. So we have an initial stage where we're talking to the client working out what we need the question to do, what the objective is and have a number of design meetings. Once we've got some draft questions that have been revised and refined, we'll test them on two pilots both of which are with real interviewers and members of the public. And what that gives us is data from the questions that we can use to test whether there are any particular issues. So for example, high levels of non-response or a code frame that doesn't seem adequate because we get lots of specified answers. But also feedback from interviewers about whether any questions were difficult to read or too sensitive. And they're very good at suggesting improvements to wording as well. The third stage is that after we've refined the questions again, after the piloting, we'll have final questions and we then need to look at the order that we asked them in, both in particular modules and between modules and the flow overall. And we also have to think about the mode for the questions. So as I mentioned, we've got a face-to-face interview and a self-completion booklet. And for new questions, we have to think about what's most appropriate. We typically use the self-completion for batteries of questions that work well laid out on paper and also perhaps sensitive questions or those that would have a socially desirable response. Sampling is happening around the same time as question design. What we want from our sample is that it's representative of our target population and large enough to give us robust estimates. So for BSA, this is roughly 3,000 people. That's not huge compared to other surveys but it does allow us to get a clear picture of public attitudes across the country. What we can't do is very detailed subgroup analysis or regional analysis. So we use random probability sampling which is a statistical method and it means that everyone in our sample frame has a known chance of being included in the sample. We use the postcode address file which is a list of all private addresses in the UK. And we then do some multi-stage sampling that first of all we select postcode sectors, then addresses, and then in the field the interviewer will select a flat if necessary and then the actual person that we need to take part in the survey. So the way that field work works is that the office sends out an advanced letter to our respondents and you can see there's just a picture of our materials on screen so you can see they're quite visually attractive. We do have a voucher, a 10 or 15 pound voucher to bank people for their time and the idea of this is to maximise our response rates and to help our interviewers because they can refer to the letter and make it clear that we're not just cold calling. Also helps that respondents can make an informed decision because we've given them that information up front. The interviewer will then approach and select a respondent and probably make an appointment to do the interviewer at a later time and then they'll carry out the interview in the person's own home and transmit the data back to the office securely and leave behind a self-completion booklet too. There will always be people in our sample that we don't manage to talk to. We get about a 50% response rate which is fairly common for attitudinal surveys and we wait the data at the analysis stage to correct for this. It's worth saying here by our four month field work period that this is because we make repeated attempts to secure interviews with the selected person and that means that we don't just interview the most available or most willing or least busy people because they won't be representative of the public as a whole. It's really important for our robust estimates that we make those repeated attempts to get the most out of our sample that we can. I'm going to hand over to Ellen and I who's going to give you an example of how you can use the data. Hi there. I'm going to take you through an analysis example. This will just illustrate some ways in which the BSA data can be used. This example uses some of the funded questions that are on personal relationships and it was done as part of our 30th annual BSA report. The focus was really on how attitudes have changed since 1983 and this area in particular is somewhere where we've seen quite a lot of change over those three decades. The two questions that we're interested in here are these ones on screen. The first is about premarital sex and the second about same sex relationships. For both, the respondent would be presented with a show card that listed five response options. Always wrong, mostly wrong, sometimes wrong, rarely wrong and not wrong at all. So in 1983 we found that 28% said it was always or mostly wrong to have sex before marriage. That compares to just 11% in 2012. If you see the chart here, you can see that there was a really steady decline from 1983 when about 3 in 10 thought premarital sex was wrong. Steadily declining to 2012 where just over 1 in 10 thought premarital sex was wrong. We've got a poll now. So in 2012 we're now thinking about same sex relationships. In 2012 28% of people said that same sex relationships are wrong. What percentage do you think said this in 1983? So the poll is now open. You can vote for A42% E52% or C62%. I'll just leave that a little longer to make sure you've all had a chance to vote. Okay, great. So again there's a split there between B and C 52% and 62% with just a few of you saying 42%. So 2012 is 28%. In 1983 it was 62% of people who said it was always or mostly wrong. So if we look at that in the chart here, so the top purple line shows and the percentage of people saying that sex between adults of the same sex is wrong. So this declined from a high of 74% in 1987 to 28% in 2012. So rather than seeing that steady decline that we saw with the premarital sex line below, there was an increase in the late 80s of people saying this was wrong and that was really at the height of fears around the AIDS epidemic. This is a really big change over time and shows a shift in public attitudes since the 1980s and this is actually an area where we've seen one of the biggest changes over time. So we can use BSA data to explore whether these changes reflect a change in attitudes across society or if certain groups are driving that change. We did some analysis by Religion, School Party and also by Age. So first of all looking at religion and views on premarital sex, this chart shows the percentage of people saying premarital sex is wrong among various religious groups and over time from 1983 to 2012. So in general we know that people are less likely to belong to a religion now than in 1983. So we might then hypothesize that views have shifted because few people are now religious. However when we look at this chart and we compare the views of the religious groups across time we see that views have changed within each religious group. So you can see for example that among Roman Catholics in 1993 32% said premarital sex was wrong and that's now defined to 11%. So you can see some of the larger shifts have been among the religious groups while those with no religion they didn't actually have that part to go because only 11% said this in 1983. We found a similar pattern when we looked at views on same sex relationships by religion as well. So the fact that many religious people have changed now more liberal than they once were suggests that there are other forces that are more important here than just the decline in religion. Okay so now looking at political party support and this time we're looking at views on same sex relationships. So there is still a difference in 2012 between the major parties with conservative supporters being more likely to say that same sex relationships are wrong. Llyfdemes are most liberal in their views and as well as those who don't have any affiliation. But within each group again we see that change over time. When we looked at premarital sex there was very little difference nowadays between groups. Okay so now looking at these results by age there's a slightly more complex pattern when we look at this by age so I'll take you through the differences first of the 2012 only. This chart shows a portion saying that these kinds of relationships are wrong by age. You can see that for both same sex relationships and premarital sex the oldest group is much less accepting of these kind of relationships than the young groups. So for example nearly 20% in that oldest group who think premarital sex is wrong and about half think same sex relationships are wrong. It's worth noting that for premarital sex the middle age groups there appear to be more accepting of premarital sex than the youngest age groups. When we did analysis of this we actually found that there wasn't a significant difference between these groups overall. Looking at change in age groups over time it's interesting to examine by age cohort rather than simply looking at age groups. So this is where we select a cohort and look at how their attitudes change over time. This is not a real cohort because this is a cross-sectional survey as Miranda's already mentioned. So we're not following the same people but we are able to follow the same age groups and see how their attitudes change over time. This will really help us tease out different effects in changes over time. So in this chart we see generational change in attitudes of premarital sex over time. So we have a line for each birth cohort and we're divided by what decade they were born in. So for example with top line is people who were born in the 1920s second line is people born in the 1930s and so on. You can see that the top line doesn't extend all the way to 2012 and that's because we don't have the sample size in those later years. You can also see that the black line is people born in the 1980s. They started joining in when there's a big enough sample size of people who were old enough to take part in the survey. So that's in the early 90s. So in general you can see that the lines are stable across time and that means that each cohort has not changed its view over time on premarital sex. The younger generations have more liberal attitudes and the older generations have less liberal attitudes and those attitudes do not tend to shift as people age. What this does mean is that as younger people grow older and replace the older generations then attitudes across society as a whole will change and you'll see the patterns that we've seen in a decline in people thinking that premarital sex is wrong. So this chart is generational tech change in attitudes to same sex relationships. Now here you can see there's a slightly different pattern. So when we look at how cohorts have changed their views over time all of the age groups have shifted that there was an increase in people saying same sex relationships are wrong in the late 80s and then there has been a decline since then. So if you look at each individual line they do all follow that pattern. So this shows a different pattern to when we were looking at the premarital sex. Although there are differences between the cohorts of the generations are less liberal than the younger generations we also see a pattern of change over time within each cohort as well. Okay, so just to sum up on that we've seen huge shifts in attitudes to same sex relationships and attitudes to premarital sex since the 1980s and we've seen changes in attitudes across lots of different groups in society. There's little difference between groups and their attitudes now towards premarital sex except noticeably for religion there are still certain groups now that's more likely to see same sex relationships as wrong. However, when we look at analysis by age that's shown that different cohort trends there are different cohort trends between these two questions and that highlights the different patterns that could be hiding underneath changes over time. So despite huge shifts there is still a sizeable minority who say homosexuality is wrong, 20% however from the patterns we've seen we can see that it's very likely there'll be a further decrease in the proportion saying this. And just very quickly here's a slide with a few bits of publicity that we got when we released those findings so we got picked up across lots of different papers and radio and TV news as well. And that leads on nicely to talking about dissemination and analysis of BSA. So a good place to start to get an idea of what BSA analysis is out there is our BSA website and there's a screen grab of that on the screen now this hosts all our online reports and we've been publishing our annual report here for the last few years there's also some standalone reports on there as well. So when would it be appropriate to use BSA? Well BSA is a fantastic resource to understand what the public thinks about particular issues to explore differences between groups so we do analysis by key demographics and also to examine trends over time in attitudes and you can do looking at trends over time by different groups as I've just shown you in the analysis example there. However, there are certain circumstances where BSA is not appropriate so one key example here is doing detailed geographic analysis so if you have a research question that's really trying to pin down into a certain area I'm afraid BSA isn't the best vehicle for that because of our sample size and because of the clustered nature of our sample so there are other surveys for example something like understanding society it's a huge survey, it's got lots and lots of respondents so that might be a better resource for if you're really going to look at very detailed geographic analysis. Another example of where BSA is not so appropriate it's not appropriate at all is if you're interested in understanding how individuals change their views over time so that is longitudinal analysis so if you're following individuals rather than looking at cross-section of the time again understanding society or any other longitudinal survey would be a better resource. So waiting and dealing with missing values if you're doing any analysis on BSA you should always use the weight which is called weight factor, WT factor it has the same name in every year so on every day setting will be called that and you just need to apply that before doing any analysis. Unlike many other surveys we always include don't know and refusal responses in our analysis these are valid responses to attitudinal questions so if you're analysing attitudinal questions then you need to include those most other surveys will set these responses to missing or negative value but on BSA data sets you'll see that they are not treated as missing value so they will be a positive value. Note that that doesn't apply to every single question across the survey if you're looking at background variables such as age or sex that you might want to exclude the missing values there and there will also be genuine missing values on all questions so for example that question wasn't rooted to a particular group of people they will have a missing value there. So just some practical issues for analysis here if you are doing time series analysis you should note that there are separate data sets for each year of BSA and all variables will have the same name across those data sets where they have been asked in exactly the same way if there's been a slight change in the wording we will change the variable name Repeat questions will be repeated every year it's likely that they'll be repeated every few years and if you're looking at changes over time a good tip is to focus on the overall trend rather than focusing on small year to year changes you should look at them in the context of the overall pattern Just on a more general level the questionnaire is split into three versions versions A, B and C Each version is asked to about 1,000 people so 1,000 respondents A lot of questions are asked of the full sample they are asked to all three versions there are roughly 3,000 responses but some questions are asked on one or two versions so they will only have 1,000 or 2,000 respondents This might have implications for your analysis so it's just worth looking into that and you can have a look at the questionnaire documentation to get more details I will now hand over to Sarah who is going to talk about accessing the data This is what the UK Data Service website looks like You can see it's www.ukdataservice.ac.uk We provide access to the BSA microdata amongst many other datasets The quickest way to get to the BSA data is to go to the get data which I'm clicking on now It's moving a little slowly I think it's a slow computer Key data on the left hand side screen Then you'll come to this screen here that shows you all the different kinds of datasets that we have available We have a lot of UK surveys These are the large scale surveys often commissioned by the government and many of those We also incidentally have cross-national surveys longitudinal studies including cohort studies We have international macro data including data from the OECD, UN and many others UK census data some business micro data and also qualitative and mixed methods data as well as some administrative data Anyway, sticking with the UK surveys I'm going to scroll down slightly and then click on the British Social Attitude Survey This should take me through to the series page Any survey that's repeated with similar questions or with similar methodology has a series page Not all surveys, of course, are repeated but surveys like the BSA are repeated and so have a series page Here's the series page There's a bit of information a little bit of an overview in the series abstract If you scroll down, getting started just general information about how to get started with our data in general frequently asked questions which is specific to the BSA and some related resources and related case studies and some external links as well The key thing is if we go back up to data access and click on the data there you can see here are all the different data sets that you can get, all the different BSA data sets I could just click on download order which is to the right of the screen but I'm not going to do that because I think it's more important to go via the catalogue page which has got all the information You can also download the data from this page As you can see it says catalogue at the top the name of the survey this time it's the 2013 To the top right you can just click on that download order button in order to download the data Before we do that we'll have a look at the title information which tells you who has deposited the data who actually did the work on the data and who sponsored it If I go down a little further the citation so obviously if you use this data you'll need to cite it and it should be pretty simple basically copy and paste that and that's the citation for this particular data set If we go down a little further the abstract has the key information so if you're new to a survey and you're not entirely sure whether it's for you or not this is a really good place to go because it gives you all the background information often gives you the aims of the survey it draws attention to other surveys that are related in this case the BSA has a module for the International Social Survey Program that I think every year I expect Eleanor and Miranda will correct me if that's wrong so that means that you could if you're interested in that particular module look at analyses not just for Britain but for other countries if they're included in this program Towards the end of that there's also main topics this is something that's quite important it gives you a little bit of an overview of how the data are collected and also some of the key questions again coverage universal methodology this is good because it shows you exactly which countries for example were involved which spatial units again it's just an overview if you're not really sure that this data set might be for you you're not entirely sure and want to check now moving down this is the key thing really documentation so this has got all the information that you should need in order to be able to use the BSA properly so this is the documentation that's provided to us by NATSEN and I can just briefly show you for example I think there we go I've clicked on the questionnaire and what you can see is it has all the information it has a list of contents and then if I scroll down it has information about each of the variables and this is searchable if you use control F and then search through you can find information about particular variables and if I go back there's also obviously you'll be interested in potentially all of these things but the user guide again this is searchable if I want to look at waiting using control F just search and that should take me to all the places in the guide that tells me about waiting so if that's the thing I'm interested in okay now to download the data you should just click on this download order button I'm not going to go through that what you will see is that you'll be expected to log in to the UK data service in order to do that you need to be registered with us and anybody can register so it's a quick process should only take you 5-10 minutes to do that and to download the data you need to say what you're going to use the data for it doesn't have to be a funded project or anything like that you can use the data for any research purpose and what we'd like is for you to just say what the use is so that that's in and their funders are also aware of how the data is being used and what will happen is that that will take you through various screens where you'll agree that you're not going to for example pass the data on to any other person and when you get to the end you'll be able to download it in SPSS data or probably tap delimited format so that's the those are the sort of stats packages that you're likely to want to be able to use this data in another way of accessing the data is to access it online so if you look near the download order button just to the left of it there's a button that says access online and I'm just going to click on that and this is an alternative way of really looking at the data I wouldn't suggest you would use this for proper analysis of the BSA it's something that could be useful if you're planning to teach students about the BSA or if you just want to explore it a little bit yourself but don't want to actually download it so if I click on the left hand pane and variable description I can see different clusters of variables here if I click on the one that says newspaper readership and internet so this is on the left hand pane near the top I click on that you can see on the right hand pane a list of the questions come up and if I click on the first question in the left hand pane it says do you normally read any daily morning newspaper at least three times a week and that's question 221 and on the right hand pane you can see more information about that particular question so it has the literal question it shows the coding so one is yes, two is no it is don't refuse and nine is a refusal and it shows the numbers of responding to answers to those queries and the percentages note that these are the unweighted percentages however okay there's some summary statistics so the number of valid non-missing cases, the number of missing cases and the universe so in other words this question was asked to everybody if we look at the next question we can see a follow on question which morning paper do you normally read or which most frequently and you can see the list of the papers and you can see at the bottom it says minus one is not applicable and those 2064 people are the people who obviously don't read morning papers so that's pretty straight forward and you can see at the bottom the universe is now if yes at ReadPap which was the question that we looked at just before I'm not going to go any further with Nesta there are other things that she can do with it obviously you can look at all the variables I've just done in that way you could if you near the top next to description there's something called tabulation so I could fill this in and you can create a cross tab that way but you do need to be already logged into the UK data service in order to do that so I would suggest if you're going to do that login first and then go into Nesta there are other things you can do this once you've created your cross tab you could show it in terms of a graph and you can see various things along the menu bar at the top there on the right hand side you can, once you've logged in you can also wait the data you can subset it and you can also in fact download the data I wouldn't suggest this is a good way necessary to download all the data because it's probably just easier to do that on the main catalogue page but one thing that it does allow you to do here is to download a subset of the data so you can download particular variables of interest if you're going to do that then make sure you get all the variables you need including the waiting variables and any other administrative variables that you need so I'm going to stop there and I think we're going to move on to questions okay so you can see on this slide the NATSEN contact should you have any questions and you want to contact Eleanor Moranda or just more generally is BSA at NATSEN.ac.uk and I forgot to mention actually something important if I just go back to the UK data service website there we go okay so I'm back on the UK catalogue page but you can do this from anywhere on the website the top right hand corner if you click on help that takes you through to some various help pages but in particular if you click on get in touch you'll go through to a web form and this web form will allow you to send a query direct to us at the data service and we provide email help desks help basically to any of our users who have questions about the data the kinds of questions that we're able to help with is if you're struggling with say registering with us or accessing the data or if when you come to look at the data you're not entirely sure what you're looking at or you don't really understand the documentation these are the sorts of queries that we're able to help with if you have a more advanced query about analysis then we're not really the right people to contact about that a query that suggests maybe that there might be some problem with a particular variable or something more advanced than that than we would contact then to ask them to help with that so I think that's the end then I'd just like to thank Miranda and Eleanor for their presentations and I'd like to thank you all for coming and I think I will end the webinar as I say the recording should be available on our website in the next few days ok thank you very much bye