 Good afternoon everyone, thank you for joining us. This webinar is an introduction to accessing and using the European social survey data. So a quick introduction, my name is Simon Parker, I work in the training and user support team at the UK Data Service based at the University of Essex and I'm joined here by Alan Humphrey, who's the head of household surveys in the survey research centre at NatSense Social Research. I'm going to hand over to Alan and who's going to be taking you through an introduction to the European social survey. Okay, hello everyone, as Simon said, my name is Alan Humphrey, I am head of household surveys at NatSense Social Research and I'm also the national coordinator for the European social survey in the United Kingdom. Each country has a national coordinator, that's that I look after the UK. What I'm going to be talking through today is as we can see there an introduction to accessing and using data from the ESS. Just to give you a quick guide as to the contents, I'm going to give a quick introduction to the ESS just so you've got some background on how it works and what its purpose is. Many of you will know this already but it's a useful background. I'm also going to provide some very quick basic examples of how the data are used really just to demonstrate how the data can actually be used. But I'm going to spend most of the time actually talking about how to access the data and I'm going to do that in two ways. The main way I'm going to do it is to demonstrate the NESTAR online analysis tool which is operated by the Norwegian data archive who are basically the curators of the ESS data based in Norway. I'll also at the end show you how you can actually download the data sets yourself so you can actually undertake your own analysis. Okay so just to start with an introduction to the ESS. It's been running really since the turn of the century that the funding was first provided back in 2001 and the first fieldwork wave was undertaken shortly after in 2002. It won something called the Daycard Prize for Research in 2005 and really what that is that the Daycard Prize is a prize which is administered through the European Union and it's to reward outstanding achievement in sciences basically for collaborative research projects across Europe and the ESS was awarded that. It was the research aspect of the prize. It was the first social sciences survey to be awarded that prize. Really I mentioned that just to demonstrate that this is a really well regarded and well run institution. In 2013 it became something called an ERIC which is a European Research Infrastructure Consortium. Essentially what that means is that the ESS has its own legal entity. It just means it's more secure and the funding is secure for the next two or three years at any one given time. So the background it is a cross national survey and it is primarily and almost entirely focused at measuring attitudes. So what it seeks to do is to look at how do attitudes vary across time and across different countries so those countries being European countries and it's academically driven. I'll show you how the questionnaire is devised or the process which that happens in a moment but basically it is driven by specialists academics in with particular research interests so survey methodologists and subject experts in a range of different fields. And it's run by Enly so that's every two years. So the first round was in 2002. Later this year 2018 the ninth round will take place across a number of countries. So we've got eight rounds of data completed so far and that covers a total of 36 countries on that in a moment and you could see there 370 thousand interviews and all the data are freely available for non-commercial use. So you just have to undergo a very short registration process at the NSD which is the Norwegian data archive. You can then access the data to your hearts content provided of course it's for non-commercial purposes. So to date in excess of 100,000 people have done that worldwide. Many of them students many of them academics also people in government and NGOs and there's been an excessive 3,000 publications which have used ESS data. Methodology is really important it's based on random probability sampling now this isn't a sampling course but random probability sampling as many of you will know is essentially the most robust method for producing a sample that represents the population in which you're interested in. The fieldwork is conducted face-to-face by interviewers last approximately an hour but that does vary of course from country to country because they're translated into different languages. Most countries use computer assisted interviewing in fact from later this year that will be mandatory for ESS participating countries and the questionnaires are translated for what's called functional equivalents. So what that means is obviously that a prime purpose of the survey is to undertake cross-national research that you want a questionnaire which means the same in every single country in which it's being fielded. So that's functional equivalents rather than literal equivalents. So what happens is that the source questionnaire is written in English then local experts often the national coordinators organize translating that questionnaire so that it has as far as possible the exact same meaning as the English original. As you might imagine there's a very strong focus on standardization of procedures so the sampling procedure is must be random probability. Lots of interviewer instruction and training so to standardize their approach in terms of asking the questions exactly as they appear in the questionnaire and not paraphrasing or not interpreting the questions for example for respondents and that's to ensure that when analysts are using the data they can be confident that as far as possible the questions have been fielded in exactly the same way across all the different countries. ESS countries seek to maximize response to get the highest response rates for the survey and then the data are aggregated once field work is complete in a completely consistent format that's obviously crucial for analysts who want to undertake cross national research so you can be confident that the variables are in exactly the same format for every single country. So just to outline that the countries that participate the map on the screen there shows all the different countries that have participated not going to go into detail here but what you can see is the sort of light yellow bar countries are the ones that have participated in all eight rounds United Kingdom among many countries there so for those countries we have data for every single round since the ESS started in 2002. Now in terms of the questionnaire I said I'd say a little bit more about this the way it operates is that there are kind of two parts of the questionnaire so on the left hand side of the screen there you can see the core topics like crime democracy and politics etc. Questions on those subjects are fielded in every single round of the ESS so that means that you can look at a variable a question it would have been fielded since the beginning usually of the survey so you can get time series analysis from back as far as 2002. That takes up about half of the questionnaire and then in addition each round will have usually two but not always what we call rotating modules and again I'm not going to go through all of these in detail but essentially they focus on one particular topic and go into much more detail so you might have say 40 or 50 questions on one particular topic and what's more some of those modules are actually repeated over time so to take an example in round one 2002 there was a module on immigration in round seven that module for the most part was repeated again so what that means is you've got a detailed set of questions you can compare the data cross-nationally across all the European countries and in addition for that particular model you can also do some really detailed time series analysis so comparing how things have changed in different countries over time so I said I'd give some examples as to how ESS data are actually used and this isn't to sort of talk in detail about what the data actually show it's really just to highlight the sorts of analysis that ESS data enables us to undertake that these are both examples that Nat said the organization I work for has produced so this first one is actually looks at the most recent round of data the round eight data and that was a set of questions on climate change so people's attitudes to climate change etc and this particular piece of analysis focused primarily on well entirely on those questions but just on looking at cross-national so we had a number of different countries participating in round eight that module of questions there wasn't really any time series that can be done because they haven't been fielded in that way before but a really detailed picture of how attitudes to climate change for example climate change skepticism compares to across different European countries and that's one of the fundamental objectives of the ESS to see how attitudes change across European countries and what we could see there is there were some big differences between the countries we could also look within the countries and we could see the attitudes as you might expect varied by age and also by educational attainment actually those relationships were different so they were different in the UK compared to some of the other European countries so that's one example of how the data can be used another example ESS data again this was something conducted by produced by Nat Sen this was attitudes to racism and this drew on a number of different data sources including our own panel on the British Social Attitude Survey this used ESS data on from round seven and some really interesting questions on biological racism so for example and this just focused entirely on the UK we didn't look at countries across Europe but we were very much interested in the UK which is why we also use British social attitudes data and we found that nearly a fifth of people actually in the UK agreed with the statement that some races or ethnic groups are born less intelligent obviously quite a key finding there to just demonstrate that the extent of racial prejudice in Britain and that used ESS data so that's an example of just using one particular country and using one of the detailed modules really looking in depth at one particular topic in one particular country so those are just two examples that just highlight how the data can actually be used so what I'm going to do now I'm actually going to show you how you can use the data yourself and as I said I'm going to do that in two different ways firstly I'm going to use an online analysis tool this is run by nsd which is the the curators of the data the ESS data based in Norway and we're going to go to their websites and look at how we can actually undertake some some relatively quick easy and basic analysis using using that tool so what I'm going to do now is I'm going to go online and I'm going to go to their website and the website that we're going to go to is up here at the top www.EuropeanSocialSurvey.org and if you were to type that in you would go straight to the European Social Survey Data Archive hosted by NSD in Norway you can see there are a range of options on the screen that you can use there what we're going to do first of all is we're going to go to the online analysis tool so I'm going to the menu here data and documentation then I'm just going to scroll down to online analysis I'm going to click on that and that will take us to this new screen which is the Nestar which is the the tool for actually undertaking the analysis it's the Nestar website hosted in Norway and what you're faced with here is a screen with the variables down here on the left-hand side and I'll just explain how these are structured so what you can see is each round here so that's ESS round 1 round 2 round 3 and this basically operates just like Windows Explorer so it's a kind of hierarchical structured set of variables so I'm going to click on round 7 and you can see when I do that that opens up to submenus and within that variable description if I click on that it's a little plus sign it opens up all the various different sets of modules so these are modules of questions I'm going to show you how to use those in a moment so what we're going to do is we're going to produce a quick table and what we're going to say is let's say we're interested in finding out people's interest in politics and how that compares across the European countries so we're going to click on tabulation up here towards the top and when I do that what you can see is that it's opened up a blank table and you can see that it's going to have the the top break across there and the side break down down the side here so we're interested in how attitudes to politics compare across countries now I happen to know that that particular variable is in the politics section here so I'm just going to click on the little plus sign there and that will open up all the different variables that these are all the different questions individual questions that are included within that model module on politics and the top one there is the one I'm interested in today how interested in politics so I'm just going to left click on that and a little box will appear up add to row add to column add us filter or add as measure what I would like to do is I'm going to add that as a row variable it's going to appear in this part of the table so I'm going to see a frequency of attitudes to so interested in politics down the side so I'm going to click on that very briefly it's going to ask for my login which I'm just going to type in here so when you want to use this you will have to go through a very short registration process just to type in that address correctly and straight away there we have our frequency so what we can see here at the moment is that we've got the frequency very interested we can see 12 percent of people very interested in politics going down to 18 percent not at all interested that's the picture across the whole of Europe what I really want to do is I want to see how does that compare by country so what I'm going to do now is I'm going to add country as a variable across the top here so I can see the responses broken down by different country as you might expect the country variable is in here in that country section I'm just going to open that up there's only one variable in there again I'm going to left click on that I get the same little dialog box opening up there this time I'm going to add it to column so I click on that it turns away and here's our table so we now can see how the responses to that question interest in politics vary across the European that all the European countries that participated in that particular round in round 7 however one thing I still need to do is I need to wait the data so this at the moment is an unweighted table and I really need to wait it so how do I do that I'm actually going to use a little icon actually see here so there we go so there's a little icon with some scales on that you can see here when I hover over that the word weight appears and I'm going to click on that and I'm going to add the weight okay now what I'm going to do you can see this opens up a little dialog box which I'm just going to go quickly back to my presentation because I'm just going to explain a little bit about waiting data so just going to quickly move through these screens here that I've just shown you so waiting that the data now this is not a statistics course I'm not going to go into detail about how waiting actually operates but as many of you will know for various reasons for example certain types of people being more or less likely to participate in surveys the raw unweighted survey data may not reflect the profile of the population that we're seeking to represent now on ESS there are three different types of weight and I'm just going to briefly describe each of those to you now so that you will know exactly which weights to use when you are undertaking analysis so the first one is what's called the design rate and all this is is reflecting and correcting for the fact that some people have a higher chance of being selected to take part in the survey than others so as a random probability survey we select addresses we send our interviewers to those addresses and if there's more than one person at that address the interviewer has to make a random choice now what that means is that people in single person households will automatically get chosen people in households where there are more than one person have a lower chance of being selected the design weight simply corrects so it means that everyone has an equal chance of being selected which is a fundamental aspect of probability sampling and don't worry too much about if you don't follow that I'll explain how to use them in a moment the next component is the post stratification weight and that includes the design weight what the post stratification weight does is it essentially forces the sample to be representative of the population we're interested in and on ASS it does this in terms of the age the sex the region and it also to some countries education attainment now for example most surveys tend to under represent younger people younger people are just less likely to take part in surveys so what the post stratification weight does is it correct for that it just weights up the proportion of younger people so that the overall sample is reflective of the general population now the post stratification weight also includes the design weights okay so these two weights are essentially combined in the post stratification weight and then the final one is the population size weights now as I'm sure you're all aware they're obviously very large differences in population size between European countries so for example Germany has a population of 82 million Belgium has a population of just 11 million but on ESS will probably have similar numbers of participants in each of those countries now if we were to just add them together without correcting for that what that means is that the Belgian respondents would kind of have an equal share compared to the German respondents but what we need to do is correct for that so that the German respondents are reflective of the relative population size so they will have a much higher weight using the population size weight okay so so which weight should you use so basically to summarize what I've just said in most cases you will be using either one of two weights so you'll either use the using the post stratification weight which of course includes the design weight or that the also you'll be using the population weight so when you're looking at a country on its own so we're just interested for example in the UK we only need to use the post stratification weight but if we're also looking at two or more countries at the same time we also in addition to that need to use the population weight okay and that last bit the population weight all that does as I said is it just gives that the right proportion to each individual country when you're adding them together okay so what I'm going to go back and do now is show you how to do that so I've gone back now to our live nest our analysis I remember if you remember I clicked on the little scales icon and that brought up this little dialogue box here and in there you can see the various different components of the weight now remember what we're doing here is we are looking across all the different European countries okay and so we are going to use both the post stratification weight and also the population correction weight so I'm going to click on the post stratification weight here and I'm going to select it there I'm then going to click on this little arrow and you will see that that then then deposits it in this box here waiting variables selected in addition to that I also want the population size weight okay and I'm going to click again on that arrow there and it then moves that into the waiting variables selected so I've selected these two variables I then simply have to click on okay and that will apply the weight to the table now just whilst we're here you can see down the bottom of part of the screen here waiting information there is a little bit more information on what I've just said so that summarizes or goes into a little bit more detail about how the different weights are calculated and what their purpose is just in case you need a refresher okay so I'm going to click on okay and after a short delay I now have a weighted table and and how can I tell that the table is weighted well two things mainly the scales icon is highlighted down here and you can see it says weight is on but also if I just draw your attention I can now see some of the bigger population countries for example the United Kingdom you can see that n as in the total sample size at the bottom is quite large 5,300 for the UK Germany which is over here 7,000 at the bottom here so if I do our comparison we compare Germany 7,000 to Belgium which is just under a thousand you can see that Germany is about seven or eight times as as big in this population weighted data set than Belgium and that's just the being implemented by that population weight if we took that population weight off we would see these samples come back to fairly similar across the different European countries so essentially what we've just done in a few minutes is we just produced our table and we can see here how attitude or interest in politics varies across the European country so we can see 16% of the UK population is very interested in politics that compares with just 10% in Belgium and just 3% in the Czech Republic okay and we can actually scroll across over to the far side here and there is actually a a total column at the end there where you can see the overall interest in politics so that's that's how we've produced a simple table and you can produce any number of tables using exactly the same approach of course you can you don't have to do it by country you can break down answers to one question by another using exactly the same process that I've just done so we could have interest in politics cross-analyzed by something different so for example trust in the country's parliament we could add that instead of country if we were particularly interested just to show you one more aspect of this analysis tool which is the subset command we may be interested not only how attitudes interest in politics varies across country but we may want to look at just younger people so we may want to look at a subset of our respondents and we're going to do that using this little icon here so if you can see subset appears I'm going to click on that and then a dialogue box comes up and so what I now need to do is I now need to select a variable to essentially filter this data I want to look at age so again I need to find age in the list of variables down the left hand side here now I happen to know that the age variable is in this block here gender year of birth and household grids so again I click on that little cross there and it opens up all the individual variables in there and I can see there this age of respondent is in there so again if I click on there this time I get the option to add it to subset because I've clicked the subset command so I'm going to click add to subset and then you can see over in the dialogue box over here age of respondent has appeared now we said we wanted to look at just younger people so what I'm going to do is I'm going to filter it on everyone who's under the age of 36 so up to 35 so I type 36 in here and I add that into the value there and you can see this box here has a number of logical operators which I can select just by this drop down arrow here you can see them all appear so we can do greater than less than less than or equal to etc now I've chosen 36 so I'm going to choose less than 36 that gives me up to 35 so we have age of respondent is less than 36 and I'm going to click on okay and then after a short delay I now get a table which is filtered and you can now see that the filter is on and the weight is on and just to show you that there if you remember when we were looking at the United Kingdom earlier I think it was 15.5 percent of people were very interested among the younger people as drop down to 9% and you can also see here that less people are actually in the table so we've got 1700 as opposed to the 5000 or so that we had for the whole population so I've now shown you the way of undertaking pretty much all the sorts of analyses or basic cross tabulation analysis that you that you might want to do now you might want to chart or produce your some tables yourself and there are a number of options at the top here I'm not going to go into these in any detail but there are some charting tabulation icons up there but what you can also do is you can also actually export the data to a spreadsheet and you do that using that little icon there okay and if you were to click on that you would just get an option there of where you would like to actually open an Excel sheet I'm not going to do that now but you can see there if I would click on open this would take me to Excel I might just try and do that if you can see that so opening Excel and there I have my table exactly as it showed on screen in Excel I can now produce charts or I can copy that into a report and take that in any format that I want to moving on from there okay so so that is basically the online analysis tool the only thing I'm going to show you now is the other way of actually getting at the data is to actually download the data yourselves you may want to do that now as you've seen the analysis I've done here very quick very easy and great for doing quick bits of analysis if you want to do something more substantial and you actually want to be able to run it and rerun it you may prefer to actually be able to download the data in SPSS or stator format and undertake the analysis yourself of course you can then write syntax which you can save and you can repeat the analysis you can change it much more easily okay so what we're going to do just going to show you briefly how you do that not going to show you any actual analysis we go back to European social survey.org that takes us back to that screen that I showed you a short time ago again we're going to click on data and documentation at this time rather than going to online analysis we're going to go to cumulative data wizard and that is the system for actually producing your own bespoke data set so here you can see that this is the process of doing this essentially what you can do is you can choose which years you're interested in which countries you're interested in and which modules of questions you're interested in so you can just ask for everything but that would be a very big data set so rather than doing that you may just want to choose the ones you're interested in let's say I was doing a bit more analysis on the politics questions in ESS so I might want to choose the politics module so I click in that box there I'm also going to want to look at some of the socio demographic variables for example age so I'm going to click on that box there and over here I choose which rounds or which countries I'm interested in now if I was only interested in one country say it was United Kingdom I might just click on that country there that would give me all the rounds of data for the UK I could just click on the latest I could just only want one year that the latest round or I may want to look at all countries for round seven and all countries for round one for example and you can see how that operates so essentially I'm just clicking which years which question countries and which question modules I'm interested in now I'm not going to do this because it will then try and start producing a big SPSS data set for me so I'm not going to do that now what you may find is that in order to do this you do need to sign in again here to the data wizard to do that but as I said I'm not going to do that but that is how you produce your own bespoke data set that you can then save on your own networks obviously completely anonymized so you can do that as much as you want now finally just before we finish I'm just going to briefly show you the how to access the documentation that you might want to use now obviously I've picked on variables here that I happen to know because I looked at them before but you may not be so familiar with the data set the data archive has an excellent system for documenting all the different questionnaires etc so you can see data and documentation again it's in here by country by theme by year I'm just going to click by year and there's all sorts of things so you can get the data an integrated file this way if you want you can get the source questionnaire here in PDF format so you can quickly click on that I'm actually going to go into an individual round and show you the more extensive documentation that there is so this is round seven so here's all the files that you can get all the different data files the main file is the is integrated file there and then there's all the documentation here field work documents so there's the main questionnaire you can see the show cards that the interviewers used you can use the project instructions that the interviewers were given so if you want to know any particular assistance that was provided or guidance that was provided to interviewers when asking an individual question you might find it in there the other thing that I think data analysts will find really valuable is what's called the data protocol so this can almost be used instead of the questionnaire so if you know there was a question for example an interest in politics you open up the data protocol and what that's essentially is is a listing of all the variables exactly how they're structured and crucially what the variable name is where they appeared so you can quickly find them when you want to do that online analysis so there's a whole raft of other documentation which is available and might be useful for users when doing detailed data analysis ok and then just finally just for more help that you might want when using this facility I'm just going to go back to the online analysis and let's quickly recreate it's a nice way of doing a recrees and so let's just say we want to recreate that table very quickly so we chose the the year over here we chose tabulation we chose variable description we opened that up we knew that we were looking for the question on interest in politics so we opened up that set we found our interest how interested in politics question we added it to a row we then found the country variable we added that to our column and finally we waited the data by choosing the post stratification weight and the population size correction weight we produced that so there's our table and what I did want to show you again this isn't quite on my screen but over on the right hand side there is actually a little question icon which you can click on I have actually got a picture of that quickly go back to the PowerPoint so that's quickly just move through here it's not on there here we go over here so the little icon the question mark icon over on the right-hand side if you click on that that will take you to the Nestar manual and lots of information about how to actually undertake analysis using Nestar okay showing you the the output to Excel so that's the help icon and that's the online guidance and contact details for nsd in Norway and you can actually should you wish to if you have a complicated query contact them directly so I've shown you basic tabulation there is actually the facility in there to do more detailed analysis for example correlations and regressions that there is a facility to do that I would suspect that most people who want to do that level analysis would probably wish to download the data set and do that themselves in SPSS or status so I'm not going to demonstrate that at all today well thank you very much everyone for attending and thank you very much Adam for this very informative presentation and goodbye