Thursday, October 19, 2017

Using SNA to Evaluate the Effectiveness of Youth Engagement CVE efforts

(Not taking the second module) 

Background: SNA has been used for years by law enforcement and intelligence agencies to identify and understand the nature and structure of existing terrorist networks. As the focus of counter-terrorism efforts have broadened to also include prevention and disruption of the radicalization process for vulnerable youth, international NGOs and local civil society organizations are beginning to address this problem directly through programs that counter extremist efforts. However, many organizations have trouble measuring whether their programs have truly prevented violent extremism and recruitment, and SNA may be able to help in quantifying and measuring the success.

A 2017 study by Search for Common Ground, an NGO addressing VE and recruitment to al Shabaab in East Africa, applied Social Network Analysis to design a CVE intervention in Kenya and Tanzania. After administering a social network survey at the start of the program to youth and community leaders in four geographic areas, the NGO identified who in the community were “influencers”, what physical spaces men (versus women) went to receive information, and what social media platforms are most popular amongst youth.  

The information from the SNA revealed that youth are most influenced by their peers, and their peers are the most influential source of sharing extremist ideologues. Some of the SNA maps used in the study are presented below. These, along with others, helped inform the design a program in which Search for Common Ground specifically targeted youth networks, and provide positive messages, empowerment, and peer-to-peer support to urge young men away from extremist ideologies and recruitment.



(Graph 10. Network visualization of influencers advocating violent solutions)

Research Question:
1)    How effective has the CVE intervention been in disrupting the communication channels used by extremists for recruitment and radicalization?
2)    What can SNA help improve the effectiveness of CVE interventions?

In the above example, SNA was employed as a diagnostic tool before the program to understand the nature of networks that are vulnerable to extremist recruitment. I propose to apply SNA as an evaluation tool after the implementation of a CVE intervention to measure the change in the shape and density of networks over time, which could serve as a metric for evaluating the success of a CVE intervention. Because the radicalization process takes place within networks, SNA is particularly appropriate as an evaluation tool to understand whether a CVE intervention was successful disrupting negative relationships over time, or amplifying the influence of those people who support peaceful solutions to problems (as opposed to violent ideologies). Many organizations have trouble quantifying the impact of their youth engagement or social media messaging campaigns, but SNA can help quantify the extent to which these programs change the nature of relationships.

Methodology and Data Collection: The network for the purposes of this SNA would include all youth who participated in the CVE youth engagement activities. Primary data would be collected through an SNA survey for participants at three points: 1) before the project starts (baseline); 2) mid-way through the project (mid-line); 3) at the end of the project/follow-up (endline).  This would be supplemented by qualitative focus group discussions to better understand the social network map. Potential SNA survey questions might include:

1)    What individuals in your community do you go to for information about identity, religion, and politics?
2)    What individuals in your community espouse extremist ideologies?
3)    What individuals espouse positive solutions to your problems?
4)    What social media platforms do you trust most to receive information?

Relevant SNA metrics: Using survey responses at the end of the CVE program, the evaluator would then compare network maps over time to see if there was any change of the network shape, structure, and centrality measures of key participants. SNA metrics could be indicators of whether the amplification of “positive” voices, or the disruption of particularly pernicious connections to recruiters, resulted from the program.

-       Size and shape of egonets of key program participants who participated in the CVE intervention can inform whether the program helped to develop influencers with the ability to amplify positive messaging to counter extremist narratives.
-       Degree Centrality (closeness and betweenness) can indicate the potential power and influence of a member by virtue of their position within the network structure.
-       Directed ties can expose from whom/to whom information is flowing and whether the direction of information has changed since the implementation of the project.
-       If the density of the network of participants increases over time, it may indicate the program is effective in building peer-to-peer networks who are reinforcing positive messaging from the program, rather than rely on negative messages from other peers.

Limitations: There are two primary limitations to this SNA. The first is the difficulty of collecting all the necessary information from participants, since many of the questions around CVE, may be politically sensitive and youth may not be willing to divulge accurate information. The second limitation of SNA as an evaluation tool for CVE is the difficulty of attributing any observed changes in the network measures to the program. In other words, even if a network changes over the length of the program, without more rigorous methods of data collection we cannot definitively conclude that the change was a result of the NGO’s program, or simply a natural evolution of a network from other causes. However, when SNA is complemented by other data collection and qualitative interviews, this limitation may be somewhat mitigated.

1 comment:

Christopher Tunnard said...

A shame you won't be doing this in 217 with us, as it's at the nexus of a whole host of interesting subjects (CVE, Al Shabaab, etc.) which lend themselves really well to SNA. You describe clearly and completely how your post-intervention survey will get you the data you need to do the analysis you propose, and you recognize the limitations implicit in the work. Nice explanation of the SNA measures and the insights they would bring. WOuld have liked to see mor on this, perhaps using combined centrality measures.

If you every want to do this in some shape or form, come see me.