Use the typology search to search our podcasts using terms from the NCRM research methods typology.
Dan Woodman on mixed and qualitative longitudinal approaches
Catherine McDonald, Dan Woodman (25-01-23)
In this episode of the Methods podcast, host Catherine McDonald talks to Dan Woodman, Associate Professor of Sociology at the School of Social and Political Sciences at the University of Melbourne. Dan is an internationally recognised authority on conceptualising generational change and the social conditions impacting our young adults.
Dan discusses explains what drew him to his area of study, why iterative models are so important in longitudinal research and how reciprocity can help reduce attrition. He also talks about his approach to writing and being wary of allowing the data to simply say what you want it to say.
This series of the Methods podcast is produced by the National Centre for Research Methods as part of the EU Horizon2020 funded YouthLife project, and is looking at how researchers can do better longitudinal research on youth transitions.
For further information on the YouthLife project, visit www.EUqualimix.ncrm.ac.uk.
Catherine McDonald 0:00
Hello and welcome to Methods a podcast from the National Centre for Research Methods. In this series as part of the EU Horizon 2020 funded Youth Life project, we're looking at how researchers can do better longitudinal research on youth transitions. I'm Catherine McDonald and today I'm talking to Dan Woodman, Associate Professor of Sociology at the School of Social and Political Sciences at the University of Melbourne, and an internationally recognised authority on conceptualising generational change, and the social conditions impacting our young adults. I began by asking Dan to tell us about his research.
Dan Woodman 0:37
Basically, for my entire career, I've been working on a project that tracks two cohorts of young Australians as they finish school and go through their 20s and 30s. So really, with that that kind of transition to adulthood, and looking at how it's changed across two groups that are in the kind of pop cultural language, they're members of the millennials group, and then the generation before them the Gen X's. So we're being able to follow some young people as they've gone through those things like starting new education courses, getting that first post school job, partnering up, trying to get into an ever more expensive housing market. Sometimes de-partnering, having children doing all those things as Australia has changed alongside it. And we were very lucky recently to get some more support. And that's one of the challenges with these longitudinal studies is finding the resources to do them. But we were lucky to get some more funding from the Australian Government to recruit a new cohort who were finishing high school in Australia this year. So we'll follow them through their 20s and into their 30s as well. And they're part of that group, Gen Z, so the next generation, so in 10 years’ time, we'll be able to say something about how that transition to adulthood has changed across three generations of young people in Australia.
Catherine McDonald 2:00
Gosh, having heard all that I've now got so many questions I want to ask, I think the first one should probably be what drew you to that area of study? What made you want to track young people in that way?
Dan Woodman 2:12
I was drawn to sociology, partly to understand what was happening to my own cohort is, you know, gender, relationships, work patterns, all kinds of things, particularly going to higher education had just expanded hugely. So it was partly trying to understand my own world that drew me to what was happening to young people. That study is mixed methods, it collects survey data that's primarily quantitative. So we, you know, we get people to tick boxes that turn into numbers. And then we do interviews with a subgroup from our larger sample of about 1000. So there's 50, young people that I've got to talk to others in the team have helped at various points. But for that generation, that generational cohort, in particular, it's, it's been one of my jobs, to go back and talk to them over the years, the stories that lie behind some of our numbers with these 50 young people across time.
Catherine McDonald 3:04
So to sort of drill down a bit, then how do you see the relationship between theory and qualitative research?
Dan Woodman 3:12
So I think theory is part of every type of research endeavour. It's not only qualitative. And there's different views on how you do theory, including some people who try to do qualitative research that mimics or models after what they see as objective type research that goes on in quantitative studies where you, you have a hypothesis, and then you kind of put the theory aside, or try to kind of ignore the theory until you've collected the data. But most people who do qualitative research, say that's not really how it works, whether it's for qualitative research or quantitative research, that there's a series of lenses that we take to the questions we ask that shape what that question is, and the way we approach collecting data on it. And you can either be reflective and think about that, or you can not. One of the things that I think separates the kind of research that we do in qualitative studies, longitudinal qualitative, but also the best academic quantitative studies is we think at a deeper level about what our data means. We live in a world that's just saturated with people collecting information on us that, you know, everyone by now has probably heard someone say the line that you know, when you're on Facebook, or Twitter or Instagram, or Tiktok, we think we're the customers of these platforms, but actually, we're the product. They're collecting data about every click and interest and everything we like, everything we share, every interest we have, was then shared to advertisers and others to get an understanding of us. And that's driven by data that is just of a profound scale, you know, billions of pieces of data coming in all the time. But for those big questions, the big questions researchers ask about anything to do with humans, how we live together in groups? What's happening to people's understanding their lives? Why they might want to buy that product? Anything that goes beyond the kind of questions you can ask just with big data with a clear end point where you have so much information that you can just pull out the patterns, you need theory. So theory is there behind every piece of research we do. One thing I will say about longitudinal qualitative research, is that it brings in a temporal dimension that needs to be thought about reflexively, so actively thought about its implications for the questions you ask and the way you design your study, it maybe a way that some other types of research that are cross sectional, maybe don't require in the same way.
Catherine McDonald 5:52
So that leads me on nicely then to the question of what sort of research questions or issues do you see as best suitable for qualitative longitudinal research?
Dan Woodman 6:02
I think there's a lot of questions that are amenable to this than in pretty much anything that involves a dimension of thinking about time and change over time. And as I said, that's one of the things you really need to think about, when you do this kind of research is amenable to qualitative longitudinal research. But like every type of research, I guess, sometimes people let the method they want to apply drive the way they do the research, when it's really getting the method that fits your research question, that is ideally the way to go. And that can mean that, you know, some research questions aren't for you, or that you have to do some more skill building or often that you build a team. But yeah, there's any kind of questions where you want to find out what's happening to people over time, longitudinal research jumps out anything that's to do with experience, or trying to get the processes behind things, often qualitative data comes into it.
Catherine McDonald 6:59
So you've given us glimpses into this already, but what sort of model or best plan of research methodology do you like to adopt? And why?
Dan Woodman 7:08
In some ways, I just don't think that you can have a research design or methodology that you pull out of the world of doing research and say, this is the best. The research question you've got, at the time, really drive what's the best methodology, but we also have to be aware that we have certain skills and abilities that we can develop, but that that will also help us think about which research questions we can best contribute to. And there's, there's resource implications. So the ideal research design is one for a world that doesn't exist where you have the ideal amount of time and resources to do it. But in the case of this research I do in Australia called Life Patterns, we are interested in following people as they transition to adulthood. So we've got that dimension of change over time, we're interested in comparing that process over two different cohorts that were born, you know, 15 years apart. So we've got this kind of another temporal generational kind of cohort dimension here. But it's also part of a world that's changing around people. So there's a kind of third dimension of change. So we've got this panel of, you know, people. So that means a panel study where we follow the same group of people over time, it's a cohort study, it's people who had an experience, you know, finishing school and starting on post schooling time. And so there's two of those cohorts in a panel. And it's mixed methods. So we do ask people, you know, what are you studying? What work you're doing? What kind of housing tenure do you have? Where are you living and who are you living with? Do you have kids? And then we follow that as it all changes over time. But we're also dealing with a world in which what it means to be an adult, what it means to be a parent, what a good job looks like, might change over time. So just asking people what they're doing, and comparing the patterns isn't going to get us to that path of change. So that's why we have kind of iterative model where the questions we asked in the survey, we want to keep a lot of them consistent over time, because that's what allows us to make the comparisons across time points within a cohort, but also across the two cohorts. But we also add new questions in and do change things because the world has changed. So we asked a lot more about social media than we used to. And another example that jumps to mind as I think about this is we've got some great data that compares what people are looking for in a good job across these two generations when they're in their early 20s. So just out of school and when they're at 30. And we use that to show that quantitatively. There was a lot of similarity across these cohorts and going against a lot of the generational narratives. Now both cohorts our Gen Xs, our 1991 levers and our millennials, our 2006 levers say they were looking for a secure job most important thing when they were in their early 20s. And when they were 30. So in some ways, not what fitted with the rhetoric at the time we were doing this that said, flexibility and opportunity were more important now than they used to be. And people would trade off job security for that. We didn't find that at all, when we use this quantitative pattern, to say, actually, even for people in their early 20s security matters, in some ways, it mattered even more in some ways than when people were at 30. And it was consistent across these two generations. But we also really detailed qualitative data talking to these two, two cohorts at the same age. So when they're in their mid 20s, about what work meant to them, what they were getting out of work, what was good and bad about work. And I was able to do a journal article, a paper that showed these similar quantitative patterns, and then showed that underneath that the meaning of job security had changed quite a lot. So for our 1991 Gen X cohort, they were in a workforce that was going through a big recession. And they were quite angry that what they thought were the kind of jobs that will be available to them after following government, parents, teachers, telling them that in the world they were entering, you needed to, you know, at least finish high school, and then maybe even do some more they were that first cohort, first generation that really did that in Australia in large numbers. So they finished high school then went on, and then came out into a job market that was not strong, the youth job market was really tough, there were not jobs that were as secure as they had been. And they were quite angry about that. And then 15 years after that, people still wanted job security. But they kind of gone through 15 years where the idea of getting a job when you were 20, that might you know that that was always partly a kind of nostalgia or false picture of the world of people starting in the mailroom and ending up in an executive role. No one thought like that anymore. And they were talking about job security in a very different way, which was a job that built a set of skills and opportunities for them, that they could then take on to the next opportunity that might not be with the same place. So a good job wasn't necessarily one that will promise you a job 10 years in the future, but one that was investing in building you. But it also had shifted to a bit more of an individualised kind of perspective about a set of skills, you were building more than the job. So we had these two things where we could show this really profound and important continuity, a sense of security, and real security matters in a job just as much as passion or flexibility or other things. And it allows you to do many other things in your life. And despite what some employers were saying in the media at that time, it wasn't true that younger generations didn't care about that at all. So we could say that, but then we could also say, well, we do have to be careful that we don't assume what all these things main haven't changed for people in this different world in which they're living in. So we could really show these two things together about continuity and change in a way that was powerful to intervene not only in academic discussions and debates, but in the broader melee of Australian policy and media and politics at the time, in a way that you know, I still think of it as one of the better pieces of research we've been able to do.
Catherine McDonald 13:44
Again, we've touched on this already but what's your approach to comparison and generalizability?
Dan Woodman 13:51
Generalizability is a term that's not so often use in research that has a qualitative component. With longitudinal research, you are often looking for a representative sample when you recruit a panel or a cohort of a population. So generalizability people can use it in all kinds of different ways and define it more clearly or not. But technically, it kind of means that you've got a sample that has been sampled correctly from a population you can define, so that you can use inferential statistics to look at the sample you have and make claims about the broader population it comes from. So very rarely in qualitative data. Are you trying to do that the big challenge with longitudinal data that everyone struggles with and has some strategies, statistical or otherwise to try and deal with is attrition. So longitudinal research brings so many benefits to looking at causality change over time, all kinds of important things. But it's really hard particularly when you do youth research. and many other things to just hold on to people. Because people's lives change, you lose contact with them in various ways. But well, and I guess I should say that attrition is never kind of even across the board. So in our study, for example, we've found that much easier to hold on to younger women, particularly younger educated women, then the young men, particularly the less educated young men, so it's been a real hard slog to try and hold on to those groups. So attrition is a big problem in terms of generalizability with longitudinal studies. With qualitative research, you are interested in transferability. And often with quantitative as well, when people talk about generalizability. They slip from talking about is the sample, truly representative sample of a population? To transferability do these patterns hold in other places? And so you can do comparative research, we ask the same questions quantitatively to try and do that. So that kind of cross-cultural research and other things with qualitative research, often, it's transferability ties back into theory or concept development, you can use people's stories and understanding of what they tell you in a contextual framework where you can say, these factors about the world they're living in can help us understand why they're having these experiences, or why they're telling us these things. Don't generalise up to a population in the same way representative sample, quantitative study does, but they can give us good hints about maybe we're thinking about the question in the wrong way. Or maybe we're asking the wrong things, or we need to look at this with a different lens. So as I was talking about about that changing meaning of job security, or the different ways people were creating that, that's something that I don't know, if holds in other places, but I would suggest that it gives us pause to think about whether we're making assumptions about things holding constant where we shouldn't be. So it's through theory and concepts as much as saying this data applies to a bigger group, that you can make those transfers in a way that might be valuable to other people.
Catherine McDonald 17:10
That's really interesting. And of course, you mentioned attrition there. How do you keep your participants on board? What top tips would you have for longitudinal researchers, when it comes to attrition?
Dan Woodman 17:21
It's hard and having lots of resources, again, is one of those things that people listening will be well, thanks for that one. So you know, being able to send people vouchers for quite a bit of money for their time, having a team that can keep in regular contact, or yourself having the time resources to make that regular contact. But also, it is about a reciprocal thing, recognising the contribution people make sharing what's coming out of the research, so people know why. So we do share reports, so people can see how what's happening to them might be similar or different to broader patterns, we also share with the participants that influence the study has had. So if there's media coverage of it, or we can make a case that has had a policy impact, we like that to let the participants know that. And otherwise, it is keeping in contact. I think having some continuity in the staff, or the researchers working on projects can help a lot. But it's not always possible. So your team can change. And there's nothing you can do about that. But as much as possible when people do change, making sure that the new people coming on to say do the interviews for the qualitative component have time to read the previous transcripts, do a handover with the last person who was talking to them and acknowledge that to the participants that there's a change and say why and say what you already know about their lives. So those kinds of things that were possible, use your resources and your staff continuity, to have a good relationship and a reciprocal relationship. And where you can't you're clear and honest with your participants, give them a sense of the limitations, and then also show them what you have done with the research over time. The other thing I would say about attrition is, and this might segue into some other things we want to talk about, about the ethics of this research that it's important to have a plan that you're honest about with the participants about how they can stop taking part, you know, all the things that a good research ethics review process will make you tick off but you should do anyway, which is about being clear with the participants about how they get out. And you can try your best to build a good relationship and reward and reciprocate. But you've also got to let people go when they really want to go. You know, that's important.
Catherine McDonald 17:29
So it's sort of the threads through all of that, in terms of what I'm hearing is that the importance of just keeping it a two way street so that they feel very much part of something. That's a lovely point to have made. So you obviously touched on ethical dilemmas there. Have you had experience have a particular one that you think it would be advantageous to share any sort of tips around dealing with ethical dilemmas? Although I appreciate obviously, they're all going to be different.
Dan Woodman 20:11
And I guess they occur in different ways. Just recently, we've been dealing with one of the sectors of education in one of the states of Australia, I won't say anything more than that, who were not happy to give us permission to talk to any of their students. Because we said we wanted to keep the data over time. It used to be common to say in Australia, I'm not sure what the rules were in other places, that the data would, you know, you'd analyse it, and then the data would be destroyed five years later. But that doesn't really help with longitudinal research very much. Few people would say that today, they'd say, we're gonna keep the data safe. You know, it's accessible to these people through these rules, anything that's available will be anonymized and all those kinds of things, but that you'll keep it over time. But the Ethics Committee through this particular part of the schooling system in Australia said, we won't let you keep this data, you've got to tell us when you'll destroy it, at what point and when we talk to them, while we're currently talking to them. What I think is happening is that there's an understanding that we now live in this world where data is being collected from people, not exactly surreptitiously, but you know, they're ticking a box that they're not reading the details of the terms and conditions properly. And just data is being collected from everyone all the time. And then sold on to other people, as I was talking about earlier, the way that online, we think we're the customers when we're often the product, and they're worried about, you know, signing their students' up it's being collected, that will be used, on sold to other people, or kept in a way that can be analysed later on. So just being clearer with them, hopefully about the institution were part of the rules that govern that, why it's useful to keep that data over time. The other thing that comes to mind is I have the flip side of building relationships with your participants, where you visit them over time is that they get a sense of knowing you, and to organise interviews with these participants I've used my personal mobile phone number. And just in one case, I did have a participant that started to contact me about non-research related things and started to develop a sense of a friendship or relationship that wasn't healthy. So that was something I had to kind of manage in a way that got it back on track. And it was fine for me because the seven other social positionality is and power in the relationship. But it is something that the researchers should be cautious about the participant relationship, you want it to feel reciprocal and respectful. But sometimes participants, particularly if they have a contact for you might start to go beyond the boundaries that you want to set. And just the resetting the boundaries, in a nice way is all that's needed. But sometimes, you know, hopefully, others have institutional support or other things that can go to if that does that to happen. One thing that this question did remind me of is that there's been times where I've been almost tempted by data, particularly qualitative data that just says all the things I want it to say, and treating people's data ethically, is also to do with analysing. And I think in a truly and in a respectful way. So you are trying to understand where answers might come from, doing comparison across all the participants, not ignoring the contradictory cases that don't fit with the pattern, but we're trying to work out what's going on.
Catherine McDonald 23:40
Yeah, that's really interesting how we just have to remind ourselves to constantly take step back and see things for what they really are.
Dan Woodman 23:46
Yeah, and I just add to that, from that kind of ethnomethodological perspective in sociology, social sciences, all of us are trying to understand social life doing some amateur social science, amateur sociology including our participants. So there's a kind of reflexivity and an interaction in the interview that's going on on both sides that you have to account for in your method, but also your analysis in one way or another. And there's heaps of different ways to do that.
Catherine McDonald 24:18
So on to analysis, then how do you go about analysing the data? And why do you choose the approach you choose?
Dan Woodman 24:26
My view on this is that you try to analyse the data in different ways. So common things for me is doing a kind of thematic analysis and comparison across cases, to particularly when you're dealing with big data sets. And we often think of qualitative as, as not as big as quantitative data sets, but they are really big if you kind of think of what a unit of data might be as a sentence or an idea. An hour-long interview has so much data to it just so much and collect think that from 50, sometimes more participants but even a small number, and then you're doing that iteratively. Over a number of years, your data set gets really complex. So you can use somatic coding. But I find, you need a way to manage that complexity that you can visualise. So it's not exactly turning it into quantitative data. But I do use spreadsheets or tables to look at patterns across time and kind of labels or codes across participants. So do that as well as that kind of traditional thematic coding. And I also build kind of narrative case studies around all the participants that are just about them over time. And using some analytic categories or things we're interested in, writing a bit of a biography of that participant. Doing the analysis before we have to collect the next wave, documenting well, what the steps in the analysis we're, having really clear descriptions of codes. And then you know, making sure that all team members and new members can get a kind of induction into that, when they step into the project is what leads to an analysis that builds over time in a sophisticated way that has integrity and rigour, that you can then also with the right rules and protocols. And another thing we do in the team is have clear guidelines on who can use the data and when in which circumstances that is that is available to the participants as well, you end up with a data set that is valuable more than your little team potentially.
Catherine McDonald 26:31
And what about the writing up? Then have you got any sort of top tips to pass on when it comes to the writing upstage?
Dan Woodman 26:39
I've managed to over the years do a lot of writing. But I must admit every single journal article, every book chapter, the books, everything is felt extraordinarily hard at the time it was being done. But I guess the flip side of that is I did get there. And for me, the strategy was to kind of iterate back and forth between interesting data, trying to reflexively write about the research question theory. So I, in all my writing kind of move back and forth between different parts. And sometimes it's been to really, really get that question that others I think are struggling with in the literature right. By writing it out. And other times the paper really develops through doing the analysis. That's what really drives it. And my advice to others is to find what works for you. And I've seen people do it in a 1000 different ways. So whatever works for you to get that writing on paper, have something to rework is what you should do. But I must admit, I don't feel like I have any profound insights. Because it is, for me tough to do writing, it always feels tough.
Catherine McDonald 27:47
I think just knowing that there isn't a set way of doing something and you know, to have the freedom that you can choose a way that works for you. And to hear that coming from you is probably a top tip in itself, to be honest. So I've just got one final question that I'd really like to ask and that is, what would you say to your younger early career researcher yourself? If you could go back and give one bit of advice?
Dan Woodman 28:09
I think I would say that research, particularly in some of the disciplines in the social sciences and humanities really can feel like an individual project and for some people that really works, but to tell an early career version of me that an academic career and all research is teamwork one way or another. And you don't have to be the best at every part of the research project. You know, when you're doing your PhD, you'll probably have to be okay at most parts of it and have an understanding supervisor to help you through the bit that you struggle with. But a lot of your academic life and research work will often be on working on projects that aren't just yours, so you won't have to be great at everything. And recognising your limits and recognising and respecting and bringing your strengths to a team project where you can draw on the strengths of others is a great way to make academic life less isolating. In my case, I think more enjoyable.
Catherine McDonald 29:11
My thanks to Dr. Woodman. The Youth Life project is funded by the EU horizon 2020 Research and Innovation programme, and it's a Twinning initiative between the Universities of Southampton, Tallinn and Bamberg, and the Netherlands Interdisciplinary Demographic Institute. You can find out more about the project at euqualimix.ncrm.ac.uk. This was a Research Podcast production. Thank you for listening and remember to subscribe wherever you receive your podcasts.