Online Radicalisation and Social Psychology

Online Radicalisation and Social Psychology

In this week’s post, PaCCS Communications Officer Kate McNeil sits down with Dr Laura Smith, a Senior Lecturer in the Department of Psychology at the University of Bath. Dr Smith is also a co-investigator at the Centre for Research and Evidence on Security Threats, and in 2015 she was awarded an Innovation Award under the PaCCS Conflict Theme for a project on predicting online radicalisation. This interview was an opportunity for PaCCS to catch up with Dr Smith to find out how her work has progressed since she introduced the project to PaCCS readers in a 2016 blog post.

Kate: Thank you for taking the time to speak with me today. 

Dr Smith: First, I would like to offer my sincere condolences to the families and loved ones of Jack Merritt and Saskia Jones, who lost their lives in Friday’s terror attack on London Bridge.

The flags are flying at half-mast for them here in Cambridge, and I know that those based here at the University are deeply grateful for the words of support that have been coming in from around the world over the past few days. Saskia and Jack’s work with Learning Together is emblematic of the efforts of all of those who have been working to turn research into applied practice to make the world a safer place. 

Your work also has the potential to help make the world safer, and I’m wondering if you would mind telling our readers a little bit about how you ended up working on the subject of radicalisation? 

I’m a social psychologist by training, and throughout the last 10-15 years I have focused on the role of social interaction in societies. When we get together with like-minded others in groups, we tend to develop shared understandings of the world, which in turn impact our understanding of what is appropriate to do. My work focuses on the processes of developing those shared norms and behaviours.

I’ve studied social interaction and behaviour in lots of different contexts. For example, I’ve examined workplaces and how new staff settle into their jobs. I’ve also looked at groups of young people and how their attitudes and stereotypes can change through discussion, and the formation of social movements including Occupy and the Arab Spring.

Over the past decade while I’ve been looking at social interaction and how it impacts upon behaviour, social media has transformed everyone’s lives. We’ve gone from the very early days of Facebook, to now having multiple social media platforms with which lots of different people engage. As social media has become a new location in which people talk to each other, online and network interactions have become natural places to apply what I’ve learnt about face to face interactions.

While working in this research area, there was an increase in terror attacks, and I encountered an accompanying narrative in traditional media about whether social media interactions are implicated in the rise of terror attacks. This raised a research question for me: can radicalisation happen online, as so many politicians suggest? There was no empirical evidence connecting social media interactions with an increase in psychological commitment to extremist organisations. That brought me to my current sent of projects, including the work on online radicalisation I did through the PaCCS Conflict Theme Innovation Award.

You and your colleagues wrote an article for the PaCCS blog on this project in the spring of 2016, when the project was in its early stages. What’s happened with the project since then?

We’ve now completed all of our empirical work, and we’re working on writing up papers at the moment. As papers become available, you’ll be able to follow along with research output from this project on Research Gateway. At this point, I’ve already written a few pieces sharing some of the findings resulting from the use of our new methods, including a paper on how social media responses to Aylan Kurdi’s death predicted expressions of solidarity with refugees, and a chapter in a  book on ISIS propaganda which was published this year.

What were some of the project’s biggest successes and challenges along the way, and what have the key outcomes been?

Throughout the project, we had to design a completely new method to look at large volumes of textual social media data and the associated metrics you get with, for example, a post on twitter. We needed a way of looking at text, likes, numbers of friends and followers, etc. To do this, we had to create entirely new methods to analyse those types of data. This included creating a new program that uses natural language processing and which breaks up language data into its linguistic components. It’s not just the words that people post that matter. We needed to incorporate both what people are saying and how they’re saying it into our analysis. We care about vernacular, but we’re also interested in the linguistic, or grammatical, style of tweets and posts. Both have psychological meaning.

Based on our work using these methods, we’ve been able to identify evidence of psychological change over time in social media posts. Looking at a large volume of social media accounts over time, we can identify change in linguistic styles which would suggest that people can develop commitment to extremist groups through the interactions they have online.

This has raised new questions. The research we’ve conducted thus far has looked at extremist groups and online radicalisation in particular, but this is just one small part of how social interaction changes people. It’s part of a much broader process of socialisation. Here, ultimately, through interacting with others people learn to socialise into new groups. We learn the norms of those social circles, and how to behave around different groups. This means you can apply the same methods that we’ve developed to lots of different contexts in which people interact and start developing attachments to groups.

How broadly can these methods be applied? Elements of this project focused on Islamic extremism – are you saying that the results are applicable to other kinds of extremism, such as far right extremism, or are the findings even more widely applicable than that?

These findings are applicable not just to extremism, but to any group – in health contexts, sports, and so on – that might connect people with like-minded others. It’s an incredibly broad tool, and now we have to choose where to go next. It’s quite exciting that now have this new method we can apply to lots of different groups in different contexts.

We also can use these methods to look at the de-socialisation process. While the methods were designed to focus on how people develop attachments to groups, they also work for examining detachment. We’ve recently been invited to present our ideas on detachment to the United States Institute of Peace, which looks at de-radicalisation programs. So, I’ll be going to Washington in January to work on that.

Do you think the things you’ve learned about online social interaction have taught you things about offline socialisation as well?

There are differences in the communication medium that can change the nature of interactions in specific ways, however, the same general socialisation processes apply offline and online. The same psychological process that underlies social interaction online and offline is the process of connecting with like-minded others and getting social validation. 

How have you managed the ethical dilemmas posed by this line of research thus far?

We discuss ethics at every team meeting, as it’s a very sensitive context. We don’t want our tool to be used inappropriately. For example, we don’t want it to be used to identify people who have absolutely no insidious aims. Consequently, what we’ve done in our publications is to keep the content of our tools private so that they can’t get into the wrong hands. We know that it’s really important that they are used appropriately, and that means we’re being very careful about how we design the tools and their usage at the moment. We ultimately want to be confident that what we’re doing is being used to prevent harm from coming to others rather than to discriminate.

Where do you plan on going with this line of research next?

I’m interested in how groups and people in groups change and de-radicalise through social interaction and socialisation processes. It’s quite common that a group that has previously advocated for radical or extremist actions decides to change their strategy, and we can apply our tools and conception frameworks to examine whether we can model and understand those changes, so that we can inform policy and interventions.

We’ve also been working with partners in government to translate our tools and methods so that they work with Arabic social media data. We’re at the stage where we’ve translated the tools, but we still need to validate them. Obviously, there are many other languages that our tools could be used with, but we’d have to decide on the context we’d want to study first, before we know the languages it would be most useful to translate the tool into.

One last question before we finish up. What would you want people working in policymaking to know about this line of inquiry? Have you written anything accessible to broader audiences that you would recommend to policymakers in our audience?

We just published a paper in Perspectives on Psychological Science which details our conceptual model, and in which we make specific policy recommendations. The paper, available on my University of Bath webpage, reviews the ways in which three different country contexts have designed their de-radicalisation policies, including counter extremism policies that focus on anti-radicalisation and counter-extremism interventions.

Throughout this work, we explain how people develop understandings of shared grievances through social interaction, and how policymakers should shift their attentions to group contexts and the group interaction site of radicalisation rather than targeting individuals. Individual targeting tends to discriminate and can even lead to radicalisation itself. Based on these findings, we’d like to engage with policymakers and people who run counter-extremism interventions to discuss how group-based radicalisation works and why it matters. We also want to communicate that people should be permitted space to voice their social grievances, and that governments should engage with people in a way that acknowledges and recognises those grievances (without agreeing with them, necessarily) and work with them to come up with peaceful solutions.

Photo credit: DWRL U. Texas –