Methods Matter Podcast – Social Network Analysis

Leah Fullegar:

Hello again, and welcome back to the Methods Matter podcast from Dementia Researcher and the National Center for Research Methods. This show navigates the murky waters of research methods, so if like me you didn’t pay enough attention during your methods lectures, this is the show for you. In this series, we’ll be looking at five different research methods with an expert from the field and a dementia researcher that has put the method into practice. My name is Leah Fullegar. I’m a PhD student from the University of Southampton, and I research dementia care and faecal incontinence. This podcast came about when I got to my methodology section of my thesis and realized I don’t know what I’m writing. So together, we’re going on a voyage of discovery.

Leah Fullegar:

Today, we are getting neighbourly, leaning over the garden fence, and having a good gossip to discuss social network analysis. And bringing all the news we’ll need today are two awesome guests. In the expert corner is Dr. David Griffiths from the University of Sterling. His research focuses on social connections and social advantage. And what tool does he rely on to get to the heart of the issues? You guessed it, social network analysis, along with social survey methods. Hi David, thank you for coming.

Dr David Griffiths:

Hi, thanks for inviting me.

Leah Fullegar:

It’s lovely having you. And our hands on researcher for today is joining us all the way from Sydney, Australia. In a recent interview, she was quoted as saying nearly everything about our brain function and brain health has to do with social relationships, and is formed by the broader social community, which means she’s the perfect person to be joining us today. We’re delighted to welcome Dr. Anne-Nicole Casey from the University of New South Wales. Hi Anne-Nicole, thank you for joining us.

Dr Anne-Nicole Casey:

Hi Leah, it’s a pleasure to be here. Thank you.

Leah Fullegar:

So, thank you so much for starting. I’ve already said, I have no idea what time it is there, but thank you so much for starting outside of normal times. Perhaps we should start by doing our own piece of social network analysis and see if you both already knew each other with some sort of six degrees of separation test. David, do you have any connections to Australia?

Dr David Griffiths:

I’ve got a colleague, Dr. Hannah Graham who’s originally from Australia, so that’ll be my quickest route to Australia.

Leah Fullegar:

So, there is a connection there. How about you, Ann-Nicole? Any connections to Scotland or Social Advantage?

Dr Anne-Nicole Casey:

We do have dear friends who are Sterling locals, though I don’t believe they’re associated with the university. As far as Social Advantage, no direct connections, but definitely an interest in social determinants of health, and how social ties affect brain health.

Leah Fullegar:

So, what do I know? We begin each podcast with me giving a summary of what I understand to be the method we’re discussing, which of course today is social network analysis. And today, I have to admit I’m not going to fare very well. So, when I think of social networks, I actually imagine Facebook and Twitter, and social media. And I had a very embarrassing conversation with one of the higher ups at National Center for Research Methods, where we talked about social network analysis, and he asked if I knew what that was. And I said, “Yeah, that’s Facebook and that, isn’t it?” And he just sorts of looked at me and went, “No. No, it’s not.” But I now realize of course that it isn’t just about digital technology. It’s about understanding communities and how they work, and how they change, and how they might contribute to overall health.

Leah Fullegar:

Oxford Bibliographies tells us that in its beginnings in the 1930s, social network analysis has emerged as a major paradigm for social theory in research. It is used by researchers in such disciplines as sociology, social anthropology, social psychology, political science, and pretty much every other area that exists. At the heart of it are three insights: those social relations are more important than individual attributes in understanding human society, that the structure of social relations is more important than their content, and that social relations can be represented by graphs of points and lines, which can then be analysed visually.

Leah Fullegar:

David, having put to bed my rather terrible understanding and then the rather stuffy formal description, can you give us a better idea of the technique and introduce us to the method, and tell me that social media is at least part of it?

Dr David Griffiths:

That was a good description. Social media is obviously a part of network analysis, in the sense that it gives us fantastic datasets that we can use, and it is a very good tool for studying social media. But network analysis is much bigger. It’s really what we think of as a network is just any system that’s got some form of actors in, and when I say actors, I might mean individuals or companies, and lots of people are doing this network analysis, looking at for instance animals and trying to understand better the way that animals communicate with other animals, where you can apply network analysis. So, when we talk about actors, we can be talking about anything. We can look at how concepts are related to each other. But the main focus of what we think of as being a network study is when we’re interested in some way in terms of the connections that occur between actors.

Dr David Griffiths:

So, we’re not viewing, for instance from a sociology perspective, we’re not viewing the individuals exist within a system, and the system has influence over the individuals. We’re thinking about the way that the relationships within that structure might make a difference. So, one example that I often give is thinking about whether or not you’re likely to attend a fancy-dress party. So, if you’re the sort of person like me who hates fancy dress parties, it’s something that you need to spend a lot of time mulling about whether you ever get invited to one or not. But if you find that none of your friends are interested in fancy dress parties, chances are you’re never going to get invited to one. But if you find that most of your friends are fans of fancy-dress parties, chances are you’re going to get invited at one point. So that’s where we think about people’s connections can then have influences, where the structure of your friends can make a difference to you, whereas if you had a different set of friends, you might be sort of less likely or more likely to attend such parties.

Dr David Griffiths:

But it’s not just thinking about the number of friends. It’s also thinking about the way that those friends are structured. Because if for instance you’ve got 10 friends and five of them are big fans of fancy dress parties, if those five happen to be the five people that you spend a lot of time with and you all know each other, then chances are that big group of people, that clique, is at some point going to come up with the idea of a fancy dress party, and because you know lots of people at that party, it’s going to be very difficult to get out of it. Whereas if your big clique of five people all dislike fancy dress parties, and it’s five people quite separate from each other. So, for instance, someone you used to work with, and sort of your sister’s best friend, and just five people that aren’t connected are fans of fancy-dress parties, then there’s possibly less chance of you being invited to one, and if you only know one person there, then your ability to get out of it is much higher.

Dr David Griffiths:

So, these are the sort of things that we can think about from a network perspective. It’s first of all, how do the connections you’ve got in the immediate area have some form of influence or control over you? But also how is the actual structure of those wider connections, and what does it mean if your friends know each other, and if you’re for instance an organization? What does it mean if your trustees are sitting on the boards of other organizations? So, we’re thinking a little bit about the individual level, and a little bit at the more connected network level.

Leah Fullegar:

That’s a fantastic example, and that really just made sense in my head, because I really don’t like fancy dress parties. So why would someone choose this method? What sort of questions is it answering?

Dr David Griffiths:

There are all sorts of different questions you could look at. So sometimes you might be interested in thinking about the way that somebody’s network is structured. So as an obvious example, we might be interested in whether people have got some level of social support. So, we might be interested in asking them questions about the people around them, because we want to make sure that their support network is full of the resources that they need. So, if for instance somebody needs some level of support, for instance someone going to do their shopping for them and they’ve not got anyone who can provide that service, then sometimes we’re interested in that at the more sort of obvious sort of level of what your network can do for you. But we can also think about it in terms of the way networks are structured themselves. So, if we’re wanting to look at for instance the way that organizations work as a collective, then trying to look at how they work together is going to be very important to how you’re looking at the system.

Dr David Griffiths:

So, I recently had a PhD student, Winston Serrano who did a study about the system of dementia care in central America. And what she was studying is what the organizations were doing but speaking to people within organizations to see what other organizations they were connected to, to try and work out whether the dementia field was well connected or whether it was completely separate from each other. And by having that perception of whether organizations seem to be working together or working in competition with each other gives us a nice idea about what each organization’s doing individually, and also how the wider sector is working, so we can understand a lot more about the way that some social structures are working by thinking about these issues of interconnectivity. So sometimes we’re interested in the network as being an explanation of the individual behaviour, and sometimes we’re interested in the individual behaviours as an example of how the network is working.

Leah Fullegar:

Anne-Nicole, I know your pretty expert at this yourself. Can you tell us how you’ve used social network analysis in your research?

Dr Anne-Nicole Casey:

Certainly. I first used social network analysis in my PhD, and I explored the friendships and social relationships of older people who are living in residential care facilities, focusing on people who have a diagnosis of dementia. And I looked at the networks, perceived social isolation and support. I collected observational data of multivalenced relationships, so whether the relationships were positive or negative or ambivalent. And in order to do this, I conducted interviews with people who had the diagnosis, and staff members, and I talked to family members in an informal manner. And I have to chart data and used standardized assessments like the Lubben Social Network Scale.

Dr Anne-Nicole Casey:

And I collected observational data directly in real time, and the residents who were able to would answer the questions about who their friends were, how they felt about the relationships, what friendship meant to them. I didn’t ask them directly when they didn’t like people for ethical reasons. But what the method allowed me to do was look at the size of the networks, the direction of the ties. If someone indicated that they had a friendship with another person, did that person reciprocate that relationship? Looked at the density of social networks, so how interconnected the relationships were. Did people know one another and have groups or form groups? And also, the structure of that allowed us to see the potential flow of social support through that structure. But unfortunately, the findings highlighted that resident unfortunately felt isolated, and showed a lack of engagement. Even though most people retained a clear concept of what friendship meant to them, they reported small, sparse networks, and indicated that they had few if any friendships. Many people had no positive relationships, and most people felt isolated, so it was a bit bleak.

Dr Anne-Nicole Casey:

But people did wish to reach out in positive ways and connect in meaningful ways to them, and what we found is that they were applying long held friendship schema. So, the way they had thought about friendships and social relationships all their life, they were still thinking in the same way, but the environment was drastically different and the people around them were different, and they weren’t choosing the people around them. Usually when you have a social relationship or a friendship, you’re choosing one another, and people in this setting were being forced together. And often, because of different effects of the dementia, they felt they didn’t know one another. The people that they lived with perhaps would do things that they didn’t like, or enter their rooms, or take things. And of course, they weren’t doing it on purpose, but yeah, people still were applying the same concepts of social behaviour and what you should do and what you shouldn’t do.

Dr Anne-Nicole Casey:

So that was a bit difficult and given the few friendships and the common negative relationships, our results suggested that it was really necessary to monitor and cultivate relationships between people who live together in care to create a social environment that supports the personhood and wellbeing of residents with dementia, and that person centred care should include a greater attention to the broader social context in which the residents lived. And of course, with the COVID pandemic, we’ve seen this play out.

Leah Fullegar:

Yeah, I was going to ask, has sort of, has the pandemic changed how, or helped others to understand the importance of social networks?

Dr Anne-Nicole Casey:

I think it has, unfortunately a bit belatedly I would say. It’s slightly frustrating as a social network researcher and a dementia researcher to realize that this type of research and evidence was going on at least almost a decade ago, and it just wasn’t a consideration. And the social relationships between people that are forced to live together 24/7… it’s very important in their lives. It’s meaningful. It shapes their whole, each day and their entire time together, and their relationships with staff, and their relationships with visitors, visiting family members and other healthcare professionals. I have a friend at this Dementia Center for Research Collaboration who’s doing her PhD, and she looked at social professional networks of people with dementia who are living in residential age care. And her results indicated that the people she spoke with were encountering anywhere from 80 to 90 different people.

Leah Fullegar:

Wow.

Dr Anne-Nicole Casey:

In where they live. And it’s something you wouldn’t normally think of when you think of someone as being socially isolated, and yet completely surrounded by people at the same time, but people that they’re not really connecting, and maybe they don’t recognize or don’t remember or don’t know very well. So that’s, yeah, that was my experience applying social network analysis.

Leah Fullegar:

It doesn’t sound like a particularly happy one.

Dr Anne-Nicole Casey:

It wasn’t, but there was hope. There was light. It was nice to see that some people were forming relationships and friendships, and it gave us an opportunity to look at what could be improved.

Leah Fullegar:

Yeah. I will ask one question, because you said about, obviously ethically you can’t ask about people your participants don’t like. But is that something that’s important to know in social network analysis? Is that useful information?

Dr Anne-Nicole Casey:

I believe so, yeah. And Dave could probably comment on this more, but particularly in dementia care research, I think it’s very important to understand when people are unhappy with their social relationships. I could have asked people directly, but some people volunteered the information on their own, and it was often quite obvious when people weren’t getting on.

Leah Fullegar:

Yeah?

Dr Anne-Nicole Casey:

Yes. So it is, and it’s difficult for care staff and family members to know what to do in that situation. Because often, they can’t really change the person’s living arrangements or who their roommates are. Where they are is often dependent on where there’s a bed available, where they can go. So, it’s important in so much as, if there’s an issue, if someone’s unhappy, if it’s a trigger for them, which understandably a lot of the times it would be for certain reactions, it’s important to know that, if someone’s being bullied or someone’s being harassed, or someone’s being aggressive. There are very serious ramifications to that, and if someone’s constantly harassing you and you respond, and it’s in a way that appears as agitation or aggression to someone else, then you might endure some sort of restraint. Unfortunately. We hate… you don’t want to see that in modern care environments, but it happens.

Leah Fullegar:

So, whilst it’s an ethically challenging question, it is something that’s important to know and observe even if you can’t ask it. And so now we have a description of what the method is and an example of how it’s been used. So, let’s get into detail and provide some top tips for anyone who is new to using this method. In this segment, I’m going to ask some quick, straightforward questions to both guests on how to put this method into practice. So, David, these first ones are for you. How should someone prepare?

Dr David Griffiths:

The most important thing is to think about what type of data you want to be correcting. So if we take an example like Anne-Nicole’s just been talking about, then if we were in a residential setting, would we want to get information on everybody within the setting and the connections between those people, so then we could have a very defined set of people because it would be the residents and the staff and we don’t need to come up with some way that we can join them together in terms of do they like each other or do they interact with each other.

Dr David Griffiths:

So, we could do it in that respect, of thinking from a very sort of more formal network, where we can produce a nice visualization. Or we might be interested in doing it in a bit more of an informal way, of just taking some of the residents themselves and trying to get some information from them, who they perceive their network to be. But it’s a bit more complicated than just saying somebody who’s your network. You then need to come up with normally five or six trigger questions, where you say things like who you enjoy talking to about particular things in your life. So, if you ask people to name up to five people in those areas, you can come up with a network of the people that are important to them anyway.

Dr David Griffiths:

So, you can think about different ways, about the way that you go in to really approach the subject. And one thing that I’d always say about network analysis is it’s much more of a toolkit that has lots of different things you can use, rather than it being a very sort of recipe type way of doing things. So sometimes you might just be wanting to take some of the concepts and some of the theories that are involved, and sometimes might be wanting to take some of the nice visualizations, the way we can draw, way we detail pictures that can normally tell us more about a structure than what we can tell from words. So, to give an example of that, a lot of people look into grooming networks for monkeys and apes, and they’re able to see there’s a hierarchical structure. So, if you draw a little picture, that this ape groom this ape, that then grooms this ape, you can see the structure of how that little community’s working in a really, really nice visual way that would take quite a lot of explanation.

Dr David Griffiths:

So sometimes the visualizations are important, and sometimes you want to use some of the more metrics involved, so there’s a lot of particular statistics that appear in that work analysis that don’t appear elsewhere. So, the important thing is to think about what type of network analysis you’re doing and what you want to take from it, and also what is the meaning of the things that you’re observing. Because we think about things that can flow between a network, if we’re thinking about money flowing between a network, then once it goes to one person, to another person, the first person doesn’t have it.

Dr David Griffiths:

Whereas if we think about information, if I pass information onto somebody else, I’ve still got that information myself. So, when we’re thinking about the element of flow, we need to be thinking about what flow actually means. And when we’re thinking about this in terms of social isolation, obviously the benefit that you’re getting from being friends with somebody is a benefit that that person’s getting as well. So, what we need to be as our first up is just thinking why this pattern of interconnectivity matter does, but also what is the meaning of it, for the respondents or for the actors within the network.

Leah Fullegar:

That was a really in-depth answer. Next question I will ask is, what form does the analysis take, and are there any standard metrics?

Dr David Griffiths:

Okay, so this could be a longer answer, but I’ll try and answer a bit quicker. So, as I say, it’s a bit of a toolkit, so different studies we use different things. So, when we’re looking at an entire network, we can use measures like centrality measures that enable us to tell us particular positions of power and influence within a network. So sometimes we’re looking at how many connections you’ve got, or how much you control the flow of resources, or how closely you can connect to other people.

Dr David Griffiths:

When we look at what we call an ego analysis, which is looking at the network around one person, sometimes we’re looking at how well connected their friends are. Because if you’re part of a network where everybody knows each other, that’s completely different to being part of a network where none of your friends really know each other. So, there’s different types of measures we can use in those respects, and it’s really just a case of, there’s so many different ways we can apply network analysis that I could either give a relatively quick answer, or I could be sat there for six weeks answering.

Leah Fullegar:

Could write an essay on it. And how do you deal with the taxonomy?

Dr David Griffiths:

So, what do you mean by taxonomy?

Leah Fullegar:

So, the classification of different relationships and things. Because it’s all quite subjective, surely.

Dr David Griffiths:

Yeah, it can be very subjective when we’re thinking about what a tie means, and that’s where we need to be defining these things quite strongly at the start of our projects and our ideas. So, if we’re thinking about, for instance, our two organisations connected to each other, then we don’t want to be in a situation where we think they’re connected because they happen to do one thing or happen to do another thing. So, if we say they’re connected by, whether they share a director or whether they’ve got a formal contractive agreement to trade with each other, then that’s very different than thinking, well, they do follow each other on social media, and they do sort of appear in quite a lot of photos together.

Dr David Griffiths:

So, what we need to be thinking about is the way we’re defining a tie, and from in that when we’re thinking about what is meaningful in terms of hierarchy, we need to be thinking about what is the message that we’re using. So, are we interested in influence in terms of being able to control information passing from one person to another, has to flow through you? Or are we interested in how quickly you can get information to everybody else within the network? So again, it’s more the preparation stage.

Leah Fullegar:

And Anne-Nicole, it’s your turn now. Are you ready? So, what skills should someone work on developing if they want to use social network analysis?

Dr Anne-Nicole Casey:

I agree with Dave. The responses are endless. But I guess conceptually, it’s important to hone your skill at visualizing and understanding social networks as complex, dynamic systems. And the system, they exist in three-dimensional space, and back and forth in time. And that sounds a bit mind boggling, but if you think about your own relationships, you have relationships with people that you’ve known for ages, and relationships with people that you’ve only known for a very short time. And they can exist over the internet, or they can exist in real space face to face, or in written documents. It just does go on and, on the way, people interact with one another.

Dr Anne-Nicole Casey:

And the way that these networks expand and contract, and relationship ties strengthen and weaken, influences a person’s feelings, and their beliefs, and their behaviours past present and future. So, it’s good to take a moment and wrap your head around that or take lots of time and wrap your head around that when you’re working with social network analysis. And always remember that social networks are, they’re changing. They’re dynamic.

Dr Anne-Nicole Casey:

And I guess beyond that, the next thing I would say is if you’re going to conduct research of any type, but social network research in particular, it’s good to develop a skill for convincing your chief investigators and funding bodies to incorporate social network measures into the designs and methods of large, well reinforced, inclusive longitudinal studies. Both cohorts would be great if we can work that in. Yeah, an article that every dementia researcher has probably seen or heard of is the 2020 Lancet Commission Report on dementia prevention, intervention, and care. And the report lists social isolation as one of the 12 top contributors to dementia risk in later life, but the authors in that report and other reports have found that social isolation data, which is largely social network data, was not explicitly measured in most studies, and authors have to use proxies like marital status or cohabitation when you’re looking at prevalence.

Dr Anne-Nicole Casey:

So, if you could use your skills of persuasion as a researcher to persuade chief investigators and funders to put social network measures into the longitudinal studies, that would be fantastic, and it would go a long way to answering some of the gaps in the evidence.

Leah Fullegar:

Well, there’s a call to everyone listening to this now. Use social network analysis. So, if someone is mining data to perform this research, are there any particular considerations for ethics and consent?

Dr Anne-Nicole Casey:

Well, honestly my experience with data mining is somewhat limited. I’ve used data from existing databases in order to look at the associations between social networks and other variables, specifically participant level data from ongoing cohort studies. Sorry. And if someone’s planning to repurpose similar existing data to answer new questions, then there are always ethical considerations around consent and privacy and confidentiality, as well as practical considerations around the quality and the validity of the data. So, if you weren’t involved in the data collection yourself, it’s important to understand who provided the data and how, and if the data’s been properly de-identified and cleaned and checked in any way. And as a researcher, you are required to understand and adhere to any existing data use agreements and data mining guidelines, as well as performing your own due diligence to establish how informed consent was obtained in the first instance, and if the consent was actually given for data sharing and reuse.

Dr Anne-Nicole Casey:

And of course, if you’re accessing sensitive personal data, then you’ll be expected to expand in great detail how you’ll securely manage and store the data and any forms of the data that you produce. And something, as a reviewer I’d like to put out there that it’s important in any publication or translation of your research findings that you should clearly and transparently explain the steps that you’ve taken to confirm informed consent, as well as ensure the privacy and the confidentiality of the data.

Leah Fullegar:

Brilliant. Informed consent and confidentiality are something we can’t ever forget about, isn’t it? Are there any other methods that someone might use alongside social network analysis?

Dr Anne-Nicole Casey:

Well, it’s my opinion that it’s hard to think of any method that you wouldn’t use alongside social networking analysis. But as always, it depends on your research question and the population you wish to research, the people that you want to reach, and the available resources or constraints of your study, and social network analysis can be used alongside everything from evaluations of training and education programs, to assessments of biological or psychosocial interventions, to neuroimaging results, to data from twin study, to comprehensive neuropsychological assessment, survey measures, qualitative interviews, behavioural observations, all of the above. I would definitely recommend people use social network analysis.

Leah Fullegar:

So, to sum up, the answer to every question is, “It depends.” What about software or graphs and things like that? Is there any particular software that’s used in social network analysis?

Dr Anne-Nicole Casey:

There are several. I throw that question over to Dave if he doesn’t mind answering it.

Dr David Griffiths:

Okay. So, the one that I advise people to start with is a software called Kayak, which is a really simple piece of software that I always tell my students it can be done by a 10-year-old. And basically, they laugh at me and say that that’s not possible, but then when we come to actually in the module, they’re quite convinced that it can be done by a 10-year-old. It’s basically, you can just produce a network within Excel, where all you need is a row and column being sort of the first name, and second column being the second person’s name. So, if you want to sort of match people to people that they talk to, you can just do that really simply in Excel.

Dr David Griffiths:

There’s a little converter that you can get that into Kayak data. And then all Kayak is, is a little package with some dropdown windows where you can get it very easily into being able to draw the sociogram and come up with some quite nice ways of sort of visualizing that, and if you want to come up with some of the measures like centrality or any of the more statistical elements, you can do that relatively easy. So, I would always say Kayak is a very good one to start off with.

Dr David Griffiths:

The other one that is easy is UCINET, which does a few little things that are more advanced than Kayak. It’s a slightly more advanced version of it that can do some more functions once you start to get to use more data, and once you become more familiar with it. But again, it’s very, very simple, and both of those packages have little sample datasets that you can play about with that are easy enough to download from their website. So, I would advise people to start off with Kayak or UCINET.

Dr David Griffiths:

There’s also a lot of work gets done in the [Dar-Package 00:34:21] now, so that quite a lot of network analysis is getting done using [Dar 00:34:26], which is more of a general social statistics package. And there’s quite a lot of more specialist packages as well for some of the more specialist things you need to do. So, there are quite a lot of different packages out there, but UCINET and Kayak are the two easy ones to get started with if you just want to have a little bit of a play about and create your own networks and try and just basically draw them visualize them.

Leah Fullegar:

I do have to point out though that you said the resulting diagram is called a sociogram, and I think that whoever came up with all these terms in social network analysis did it on purpose to make them sound like social media platforms.

Dr David Griffiths:

All the social network analysis stuff started before social media, but there is an incredible number of terms that we can call the links between people. Nodes, or we can call them edges and arcs, and we can call them all sorts of different things. But basically, it’s a really simple idea that we just have some nodes or some actors that are connected, and we can produce some pictures that we can call sociograms, or graphs, or as I often call them to my students, pretty pictures. The actual names don’t really matter too much.

Leah Fullegar:

Well, fantastic. Thank you so much. This is great, and I could really see how this particular method would’ve added to my own PhD research. So, what have we learned so far? We’ve learned that social network analysis is about mapping the sort of interactions and relationships, and the quality of those, and the directionality, and the purpose of those relationships in a defined group of people, and how those relationships might impact every aspect of someone’s life really. Which it sounds like a very, to say it succinctly in a couple of sentences, doesn’t explain how much work probably goes into that.

Leah Fullegar:

So, in this final part of the show, we’re going to discuss the common pitfalls, challenges, and how to avoid them. Interdisciplinary research on social networks is experiencing unprecedented growth fuelled by the consolidation of the field of social network analysis and the increasing availability of data from digital networking platforms, so maybe all the problems have been solved, but I doubt it. So, Anne-Nicole, can you tell us what challenges did you come across in delivering your research, and what might you do differently if you were to do it again?

Dr Anne-Nicole Casey:

As just was pointed out by the previous conversation, social network analysis methods come with their own unique terminology and way of looking at the data. And although the methods have gained tremendous ground in use and popularity over the past few years, the terminology and the applications can still seem esoteric if you’re not accustomed to them. And I found it challenging in the early days to translate methods and study outcomes in ways that were accessible and meaningful to the target audiences that I wanted to reach. And as well just generally, in doing the research there were challenges to collecting self-report social network data from people who are experiencing complex health conditions with dementia. And working with people in the community and residential care takes time. It’s important to establish rapport, with everyone you’re going to interact with, and to be prepared to explain exactly what you’re doing and why in clear language without the jargon, and to flexible and understand that your research is not their first priority.

Dr Anne-Nicole Casey:

And when you’re dealing with someone who is experiencing challenges of their own with their health and their cognition, it’s important to be with them in the moment wherever they are. Don’t force them into your space and time and be prepared to come back later for a chat if it’s not a good time to talk with them. Basically, conduct person centred research when you can, and try to make it inclusive for people who have sensory deficits if someone has impaired sight or hearing, or mobility if they can’t… if they have difficulty expressing themselves, difficulty with their speech. And thinking of social network data in particular, it’s typically more difficult for a person to recall information than to recognize it. So, consider using rosters of names or photos, or other visual aids that might trigger someone’s memory or something that they can recognize.

Leah Fullegar:

And David, what do you think are the common pitfalls, and how would you avoid them?

Dr David Griffiths:

Yeah, I think one thing many people do when they start to use network analysis is basically trying to do everything. So there’s quite a lot of different ways we can apply network analysis, going from basically just a continuation of more qualitative interview type things where we’re just adding the network as sort of an add on, so we’re not moving too far away from adding in a few theories like the way through to trying to understand the structure of networks and producing algorithms in order to understand a lot more about the different types of configurations that can exist within a network.

Dr David Griffiths:

So, one thing that I often find people struggle a bit when they first get involved with networks is trying to think that they need to have centrality measures in there, and they feel that they need to have some form of an imitation test, just because all the textbooks talk about these particular things. But what we’re trying to do with network analysis is basically use methods which aren’t that different to what we’ve used before within our disciplines, but just take an extra step to understand a bit more about the way that the interconnected structure would operate. So, the most important thing to remember is that what we need to do at the end of our projects is go back into our disciplines and communicate in the way that people who know our discipline but don’t understand network analysis can basically understand what we’re talking about. So, we need to be really aware that we’re selecting the best methods for the thing that we’re trying to do, rather than trying to do everything.

Dr David Griffiths:

I think following on from what Anne-Nicole said, we also need to be aware that no matter how easy we make it for respondents, they’re always going to try and trip us up. So, if we’re thinking about asking people questions about who’s important to them in terms of people, then we’re going to get people saying, “Right, my friends at the tennis club,” or people on an internet forum, and organizations with them forward which flip in to being the people. And one important thing is to be trying to anticipate what sort of things respondents might be saying to trip you up, because you can guarantee when you get into the field that you’re going to get some answers that you just never expected, but which do seemingly make sense. So, try and anticipate entirely what people would say, and definitely do pilot studies with people you know, and also people you know are going to try and annoy you because you need to be ready for the unexpected to come up.

Leah Fullegar:

I know full well, if anyone said, “Who’s important to you,” I’d told them one of them’ s sat on my feet and is a whippet, and the other one’s asleep on the sofa is a labradoodle. So, one question to both of you. If you could give a bit of advice to anyone who was planning to use this research method for the first time, what would that be? And if I put Anne-Nicole on the spot first.

Dr Anne-Nicole Casey:

Certainly. Well, I guess thinking about the type of information that’s needed in the literature generally, I’d encourage people when they’re designing the research, if they can aim for longer follow-up times, a greater focus on the diversity, the quality, and the meaningfulness of relationships. And aim for precise information that can then enable specific targeted network interventions that can promote, in the case of the area of my interest, interventions that promote healthy brain aging for at risk members of the community.

Leah Fullegar:

Thank you. And David, what would your piece of advice be?

Dr David Griffiths:

I would advise people to embrace the fact that it’s interdisciplinary and read lots of stuff from outside of your discipline. So, one of the things that I love about going to the big social network conferences is that there are literally people there from every academic discipline, so you can suddenly find that you go into a session where with people from agriculture, and biology, and then a study from health sciences, and then another one from geography, which are all fields I don’t really work in at all. But just seeing the way it’s getting applied and seeing it from an outside perspective is really nice. So, I would advise people to not just look at the network studies within their area. Look across all sorts of different disciplines.

Dr David Griffiths:

And a lot of the stuff that I’m interested in, in terms of human sort of social connections and the way we might measure that, I get quite good insights from looking at the way people measure the friendship patterns of whales. Sociologically, that doesn’t make any sense, but it’s a really nice way of being able to see the way people are thinking about the methods and people are taking it forward. So, the main advice I would give to people is to embrace the interdisciplinary priority element of it.

Dr David Griffiths:

And the second bit of advice is that there’s always going to be studies and things that you’re interested in. So, when I’ve seen papers called the social networks of Brit pop, or the social networks of Formula 1 drivers, I suddenly just drop everything and read them. Not from an academic perspective, but just based on my interest. So, you can often find some really fascinating studies that you wish you’d written. And just embracing the interdisciplinary element of it rather than reading the stuff related to your work is a really nice way to move forward.

Leah Fullegar:

I can see in my head, thinking of doing a social network analysis of my dogs’ doggie friends. That would be quite fun. This has been really helpful. I mean, in dementia, we know that social isolation and loneliness can be major factors in contributing to rapid deterioration, and to people getting the disease in the first place. But more research on this would be great, as Anne-Nicole pointed out earlier, and this method is clearly essential.

Leah Fullegar:

So, in this final segment, I’m going to give our expert, David, one minute to tell our listeners what they should go away and read to further their knowledge on this method. David, over to you. I’m starting the clock now.

Dr David Griffiths:

Okay, so the first book I would advise is the book called Social Network Analysis by John Scott, which is an overview of how network analysis exists and the way you apply all the different methods. So, it’s a really nice introduction to the concept. If you’re thinking about doing a study using network analysis, then there’s a book called Doing Social Network Research by Garry Robins, which assumes that you know the basics of network analysis but gets more into the practical things that you’re going to encounter, such as all the ethical considerations and all the sort of pitfalls that we’ve talked about already. So that’s a fascinating book to read if you’re thinking of doing a project.

Dr David Griffiths:

And if you want to just get an overview about how network analysis works and how you can apply it, then the book Connected by James Fowler and Nicholas Christakis is a fascinating book about the way that people’s friendships have an impact upon their health and gives a really nice example of the way that we can use network analysis to understand more about health.

Leah Fullegar:

Fantastic. Thank you so much. I’m afraid that’s all we’ve got time for today, so let me say a huge thank you to our wonderful guests. In the expert corner, from the University of Sterling, it’s Dr. David Griffiths, and the brilliant Dr. Anne-Nicole Casey. Thank you so much, both of you, for coming and chatting to me, and making this seem like something I might understand. And to you dear listeners, join me tomorrow for some more romp through research methodology. But for now, I’m off to go do a critical analysis of my very few Twitter followers. We will be back tomorrow continuing our journey, as part of the Research Methods Festival, where I’ll be joined by Professor Andrew Clark and welcoming back Dr. Kahryn Hughes discussing qualitative longitudinal methods. Bye.

END