 Felly ni oes, dwi wedi'u bwysigol ydym ni'n mynd i gweithio i ni. Mae'n ddiddordeb yn ddiddordeb i ni, ddim wedi'u brifuddwyr. Mae'n ddiddordeb i ni i gweithio i ni i'n ddiddordeb i ni. Dwi'n… Dwi'n… Dwi'n… Dyna'r wneud i'r ystafell a gennym eu rhai cyfnodi'tau a dwi'n gwneud y Llywodraeth Cymru, ynddyn nhw'n cyffredig pwnghㅗs i'r sefydlu y bwysig cyddur. The previous3ver, astud yn llwydiol, byddwn yn fwy ffoci feddwl i'r cychwynfyr yn ymgau bydd y Llywodraeth Cymru, i'n fath o'r fath o'r tydd i'w newid. I'm also doing a linking of two different data sets, which might be interesting for some people to learn about as well. So just a really quick introduction to myself in my kind of context in which I'm doing this research. The fourth year of my PhD, hoping to finish quite soon, and my entire PhD looks at low pay and progression in the UK. So you might know that in the UK there are quite a lot of low paid workers, and there's also some evidence that the ability out of low paid jobs into slightly higher paid jobs, better jobs, is relatively low in the UK. So I'm interested to find out more about what sort of things influence that progression from low pay, and particularly when it comes to the local labour market environment that the workers are in. So I'll talk to you a little bit more about the specifics of this study that I'm doing at the moment, which I've looked at another aspect to do with local labour markets in progression. In the first part of my PhD I'll talk to you about this study here for which I'm using the LS, and I'll talk to you about why I chose to use the LS in combination with the LFS. And then some of the five experiences of applying for the data using it and so on. This probably will mirror to some extent what you've already heard, but anyway, and I'll share some early results as well, and tell you a bit about what I'm currently working on and planning to do next. So this graph hasn't quite, you know, worked out the way that I am. Let me see if I can just resize it slightly so you can read. So the kind of general premise of my research is that in the UK and also in many other countries over the last couple of decades we've seen this sort of hollowing out of the labour market in terms of the proportion of employment in what you could call sort of intermediate or middle wage occupations as slowly declined relative to employment in relatively low paid and relatively high paid occupations. So you can see it, this is based on our data, just divided all occupations into some broad categories. So we've got low paid here on the left, we've got intermediate, we've got social professional, major and professional, and the average wage associated with each of these groups tend to sort of go up. As you move forward to the right and you can see that from 2001 to 2011 there's quite a big relative decline in the employment share in intermediate occupations. And this has happened since about the 1980s and one of the reasons, one of the main reasons is because of the influence of computer technology on the workplace. So this tends to affect occupations like sort of administrative, secretarial occupations as well as some types of manual occupation where machinery is replaced. A lot of tasks that workers used to do, there might be other causes as well, but I'm not so much interested in the causes. I'm interested more in what are some of the implications of this trend. Obviously the trend in itself has been studied quite widely and it's interesting just to observe the sort of changing structure of the labour market. But given my interest in progression and low pay, I wondered, well what does this mean for workers who are in relatively low paid jobs who might be looking to kind of move up the occupational ladder is the fact that we now are perhaps relatively fewer intermediate jobs available potentially. Does that mean that they are finding it more difficult to achieve that sort of upward occupational ability? And given also my interest in local labour market areas and recognising that different areas of UK are quite different occupational structures, this way that I've thought about investigating this is by exploiting the geographic variation in this sort of polarising trend that has happened. So this sort of job polarisation of employment in both low and high paid occupations relative to middling occupations has affected different areas of the UK differently. And so you can compare if you like occupational mobility from low paid occupations between these different types of areas. So if you live in an area where there's still quite a lot of these intermediate jobs, is your mobility affected in a positive way compared to if you live in an area where there's been more of a hollowing out of these intermediate occupations? So to investigate this I needed to find the right data and I sort of in my head divided it up into two main types of data that I needed to suppose. So I needed to first of all look at occupational mobility that's kind of my depending on outcome variable. And then I also needed to have some data on the degree of job polarisation at the local level. So for the first one the occupational mobility data obviously I needed longitudinal data because you need to observe changes in occupations and it needed to have quite a large sample size because I'm looking at a specific group of workers, workers who are initially start off in quite a low paid occupation like I don't know being a sales assistant or a cleaner or something like that. And because there's relatively little occupational mobility if you just look from one year to the next you need to look at it over quite a long period of time. Again this wasn't really formatted very well is it? Anyway I'll just talk. So you need to look at it over a relatively long time period especially workers in low paid occupations they don't tend to do jobs all that often and occupations even less frequently. I needed to also have occupational data obviously and preferably occupational data that is consistent over time. So you need to, those of you who've worked with occupational data before might know that over the years there's been various different classifications. So the first introduced standard of occupational classifications has been like in 1998 or 1990s or something and they've updated it since. So there's a 1999 version, there's a 2000 version and there's a 2010 version. So as nature occupations change they update you which is a good thing because you need to reflect the occupations that people are actually doing. But it makes things very difficult when you're trying to look at changes over time because you want to observe them according to the same classification system and both the start and the end of your period to see whether they've made a change in occupation. And the last thing is that I needed to know where people lived in terms of what their local labour market area was in order then to identify whether their local labour market was polarised or not. And then for the area level data I needed similar criteria so also relatively large sample size because small areas so you need to have a sufficient number of observations within each area. Obviously it needs to go over the same time span as the other data set. Also you need occupational information and also you need to have the information available at the same geographical areas. So why did I choose the LS for the occupational mobility part of the research or the data requirements anyway? Well it kind of meets all the criteria that I just set out. It has a lot of sample size because it's collected every 10 years even if you just use two ways. You already have 10 years worth of observing people's lives so in that period quite a lot of people change occupations which is good. They conveniently have occupations in the same classification system although it's imputed data for one of the ones I think. But it allows you to observe that mobility that I was talking about and you have quite relatively small area level geographic variables. I'm using travel to work areas in my case which were available in both the 2001 and the 2011 data set which was great. And then I chose to use the LFS for the other part the local level data about polarization but also because it meets all the requirements that I set out. You know relatively large sample size and so on especially because you can pool data for multiple years. And unlike the LS it also has earnings information which was important because I wanted to look at the wages associated with occupation in order to kind of construct my measure of polarization and a big plus is that you can access it through the secure lab which I think someone mentioned before this is the online sort of secure system that you can use wherever you are. You just need to get it set up which just take a long time but then you can use it wherever you're having to study your work which is more convenient than traveling to London. So I probably don't have a lot of time left but I'll just sort of reiterate the things that have already been said. Obviously there is an application procedure to access the LS so you have to take some time for that but it didn't take nearly as long as my secure lab application for the LFS so I just wanted to mention that it's relatively good. And as others have also mentioned this is quite good documentation about the variables and so on available which will help you in your application. The sales and staff are excellent. So in my application I had to also explain to both my application for the LS and the LFS what my plan was to eventually link the two data sets together and that I guess just introduces a slightly extra layer of complexity because I was linking at the area level it's relatively straightforward to do but the only slight issue is that although I'm currently doing the analysis of the LFS in secure lab at Coventry University where I study they will not allow any data to be exported from the secure lab or to be released from the secure lab. So I've had to also apply for a secure research service on SRS application for the LFS to then replicate my analysis there on the LFS. So it's a slightly complicated procedure so if anybody is thinking of doing this linking of the data sets it's worth the thinking through in advance how you're going to do that and talk with the support staff they will be able to help you to figure out how to do this. So I haven't yet done this but hopefully very soon I'll be able to do the linking and should I probably stop. Oh okay that's fine. So as I've already said before travelling to London is not ideal I do know that I can't send two files but in my case I really need to be there and look at the data and figure out exactly what there is what needed and what did they look like and my coding is not quite as meticulous as maths is probably so I have a lot of mistakes. So at the moment I've been coming in to do my analysis. I live in Birmingham so it's not that bad travelling to London but obviously just take a bit of time but once you're there the SRS environment is quite nice to work in it's always quite quiet so that makes it a good working environment. Plus the fact that you're on PCs that you can't use the internet on is really good for your discipline because there's no distractions. So I'm super reductive when I go which is great. So as others have also mentioned the LS is a really great data set. It's got a great range of variables. It's got very low attrition compared to I've previously worked with understanding society where you do see higher attrition rates even from year to year. So it makes things a lot easier that like the majority of your sample in one census is still going to be there the next census. And just really briefly I've just looked at the occupational ability of workers who start in a low paid occupation just to sort of get an idea where they ended up ten years later basically. So this is all only for those who are still in employment in 2011. And you can see that like actually the majority are still just over half are still in a low paid occupation ten years later. So this goes back to the point that I was saying there's not all that much occupational ability. And then there's quite a lot of ability to intermediate occupation. You know biggest occupational categories that they tend to move into. But when you add these up like you it's still higher than this. So there's still a relatively large amount of ability for low paid occupations to also you know occupations of the sort of higher end of the sort of skill spectrum and weight spectrum as well. Which is quite encouraging to see. So if anyone has a question. I was just wondering if you've dealt with self-employed at all because I'm thinking the LFS doesn't have wage information on self-employed. The LFS. No I'm looking only at people who are in paid employment. So I can't make things a bit easier. I don't know about the LFS what information they might have on self-employed earnings. I know that it's obviously not as good and reliable as any wage data. Yeah although in the LFS you do observe people over multiple ways. So in terms of that you could maybe do some adjustment if you were interested. But I don't have any experience with that. So I'm not doing the best person to ask. I'm afraid. I was going to say in your sample is that you look at men age 16 to 64 and women age 16 to 60 I guess. In both 2001 and 2011. So obviously those English older in 2001 is going to fall out of your sample because of age. But that also goes the easier in the two. It might help you go to differentiate what you thought sort of like natural progression. So those people who enter the labour market at young they will be a natural progression of going from low to intermediate to those who have been directly affected by this criminalisation process. So what I'm saying is that in any specific analysis of young workers versus primary versus older. Yeah exactly someone at the start of the period is already in their 40s might not necessarily be expected to move all that much anymore at that stage. Although having done some previous work on this with women and men it's slightly different because some women after having children they might return to the labour force and then sometimes they do still have some ability after that. Especially if maybe they go back to education as well. But yeah that's definitely something that I'll be looking at and I did do a little bit more of finding detailed analysis. You know different age groups and so on so you can see differences definitely and you might also just be able to see those differences in the effect of polarization as well on the different age groups. So that's a very good point.