Assessing equifinality in Islamic extremist pathways using fsQCA

ABSTRACT

This study uses a fuzzy-set qualitative comparative analysis (fsQCA) to examine the radicalization pathways of Islamic extremists in the United States from 1980 to 2018. This study draws on the Profiles of Individual Radicalization in the United States (PIRUS) database to examine the role of individual-, experiential-, contextual-, and familial-level conditions – or push and pull factors- on the radicalization process. More specifically, this study looks at radicalization to violent behaviour. Using fsQCA, this study found that the radicalization pathway for Islamic extremists is extremely difficult to reduce to a simple conditional pattern, and that the combination of conditions for the presence of radical violent behaviour is very similar for the absence of radical violent behaviour. More specifically, being married and holding a deep commitment to radical beliefs were individually and collectively sufficient conditions for explaining the presence and absence of radical violence, as was group membership. Having radical friends, being unemployed, and having a college degree each has less consistency, but predictable relationships with these outcomes, based on their conjoinment with other conditions. Implications and areas for future inquiry are discussed within.

KEYWORDS:

• Radicalization
• terrorism
• extremism
• Islam
• fsQCA

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. They also rounded categorical measures back to their original scaling.

2. Some contend that rounding for categorical variables can introduce bias (see Enders, 2010; Horton, Lipsitz, & Parzen, 2003), the amount of missing data was usually pretty modest (see Table 1), and Xia and Yang (2016) argue that there is very little difference in terms of coverage between rounding methods when values are not substantial. More importantly, this technique (QCA) does not rely on standard errors like traditional regression modelling does. As such, the introduction of some level of bias via this technique is not likely to outweigh the benefits gained, in comparison to using listwise deletion instead.

3. The central outcome in question- radicalization to violence- was not imputed, and four cases had missing values on this measure.

4. When examining descriptive statistics, many Islamic extremists do in fact have some level of criminal history, although it is mostly pretty minimal, and there is not a large amount of overlap between this factor and actual violence, given the data.

Additional information

Notes on contributors

Wesley S. McCann

Wesley S. McCann is a Research Social Scientist at RTI International. His research interests center on extremist pathways, P/CVE program evaluation, terrorists’ pursuit of CBRN weapons and interdiction efforts, immigration and border security, and criminal law. He has taught at half a dozen different universities and has published over 30 articles, books, and chapters on many of the foregoing topics.