This past Thursday I met with Evans Smalley, close friend and student in the Countering Organised Crime and Terrorism MSc course at UCL over coffee to discuss her upcoming research, conducted as part of her dissertation. The conversation went as follows:

Hello Evans! First things first, could you explain what your current project is on?

Hi Alex, I would be delighted to explain a bit about my project. So basically, I am looking at the radicalization process from a social-ecological perspective, which simply means examining causal mechanisms of radicalization through the social relationships, environment, and larger systemic factors that contribute to or facilitate radicalization. Inherent in this approach is the assumption that radicalization is caused by interactions between factors on multiple levels of analysis, and a de-emphasis on individual or personal factors. Specifically, I am focusing on instances of radicalization in Minneapolis MN, USA. This city is interesting as, despite being the 46th most populous city in the country it has the third highest count of radicalized individuals according to the PIRUS database.

So this is already interesting, why are so many people being radicalized here, specifically? Digging a bit deeper I found that most explanations for the high number of radicalization simply cite the fact that Minneapolis has the largest Somali community in the country, and leave it at that. But that answer is not helpful, for many reasons… More interestingly, the city with the second largest Somali community in the country, Columbus, OH, does not evidence disproportional rates of radicalization. Even more interestingly a recent in-depth examination of the Somali communities in the two cities (Chambers, 2017) found that the community in Minneapolis is better integrated socially and politically and doing better economically than the Somali community in Columbus. So by most hypothesized models of radicalization focused on individual risk factors, Columbus should theoretically have more radicalized individuals than Minneapolis. But this is not the case. Something else must be accounting for these differences, and I thought it was unlikely to be factors on the individual level.

I decided to focus on a number of different variables using open source methodology, and compiled a dataset of radicalization instances from both cities. I was most interested in where and in what type of setting these instances of radicalization took place. What I found was that radicalization instances in Minneapolis were more likely to occur in large, public spaces compared to Columbus. I also found that radicalization significantly clustered in Minneapolis. I used measures of social control, population heterogeneity and residential instability, as outlined by social disorganization theory to try to make sense of this spatial clustering and found that high percentages of individuals speaking English as a second language increased the likelihood of instances of radicalization occurring in that area, but also that increased language diversity in an area decreased the likelihood of radicalization occurring there. No other measures of population heterogeneity and residential instability were significant.

Very interesting stuff! Why do you think other social disorganization measures did not have any relationship with instances of radicalization?

I think the answer is simply that these measures may not be a good fit for explaining causal mechanisms in the radicalization process. Which, I mean, may not be that surprising as we are dealing with moderator variables here. The measures used were actually measuring side effects or indicators of social control. Decreased social control can theoretically affect monitoring levels of radicalization settings, which then impacts the radicalization process. So there were at least a few mechanistic layers between the moderator and the actual effect on radicalization. I think social control may still be playing an important role in this context, a qualitative analysis of open source data indicated themes of monitoring failures at radicalization sites, however it might simply be that population heterogeneity and residential instability are not great indicators for how social control affects radicalization specifically. Which is not necessarily a new idea, there have been many conflicting results in the literature on how population heterogeneity affects crime, I think I read a recent meta-analysis that found that immigration in an area has generally no affect and even actually may decrease crime rates (Ousey & Kubrin, 2018). So, I think the takeaway message here is that moderator and mediator variables are tricky, and we need to be careful about the conclusions we draw.

Does the research suggest that radicalization mostly occurs in public spaces across the United States or is Minneapolis an exception?

This is an interesting question, in short, I don’t know. Of course, my research can only speak for Minneapolis and Columbus, but there has been a trend expressed in the literature of radicalization transitioning from public places to more private places. This is thought to be due to increased formal and informal monitoring in public places because of both increased community awareness and programs and policing specifically designed to target radicalization. First contact may still happen in public, but increasingly it seems that radicalization is taking place in private spaces with little to no monitoring. What is interesting here is that there are many formal programs, mostly community-based CVE, in Minneapolis designed to police radicalization, and these programs have been active for years. So it’s interesting that this pattern of radicalization in public is persisting. I think the key here may lie in the lack of trust in federal authorities that exists in the Somali community in this context. Some researchers have found that trust in police or authorities running these CVE programs is absolutely vital for the success of the programs. Community trust in formal authorities and how it impacts radicalization clustering would be a great next step for this project.

And finally, could you name two references that significantly contributed to your literature review?

I will give you three: Lafree and Bersani, 2014 investigated how population heterogeneity and residential instability correlated with terrorist attacks at the county level in the United States, I was hoping to replicate this finding with radicalization as well.  And Southers and Hienz 2015 have done some really interesting work examining the social structure and attitudes of the Somali community in Minneapolis.  I also learned a lot about the ecological approach to radicalization through Genkin and Gutfraind 2011.