Skip to main content

Radicalization Trajectories: An Evidence-Based Computational Approach to Dynamic Risk Assessment of “Homegrown” Jihadists


Authors: Jytte Klausen, Rosanne Libretti, Benjamin W. K. Hung & Anura P. Jayasumana

Date of Publication: 1 October 2018

Journal / Publisher: Studies in Conflict and Terrorism


Purpose of Study

Key Questions Addressed

Examines mobilisation to violence


Design of Study


Quasi-Experimental. Empirically test the use of four phase model for assessing radicalization trajectories of US homegrown terrorism offenders using coders trained on a detailed codebook which enumerated specific cues to various behaviors linked to progressive radicalization linked to the phases of the NYPD four phase model.  The behavioural indicators were used to assess increasing radicalization. In addition, the research effort sought to identify specific behaviors associated with Salafi-jihadist radicalization and identify common sequential segment pairs of behaviors that reliably anticipate criminality.

Number of Participants

335 records from which 135 forensic biographies were developed.



Type of Participant

Islamic terrorism          


Understanding mobilisation to violence/radicalization pathways, mobilisation to conduct attacks on home soil, identifying behaviors that reliably anticipate criminality for Salafi-jihadist actors.

Key Findings

The evidence strongly supports the idea that terrorist learning follows a limited number of highly scripted pathways, although only a handful of sets of sequenced behavioral changes proved to anticipate terrorism-related criminality. The findings do not support the idea that radicalization is the product of social alienation. Modeling the trajectories as a stochastic dynamic process and “pruning” transitions less than 10% informs an understanding of what behaviors may immediately come next (for example, issuing a threat is likely to be followed by some form of criminal action). Some variables were salient for women and not for men (e.g., marrying a jihadist). The scripts for “what to do” vary greatly between the jihadists and other brands of violent extremism (e.g., the White supremacists or antifederalists). 

Key Recommendations

Further research is required to assess variations in the process and to specify more precisely the behaviors associated with different subgroups: women, youths, older  , and former criminal offenders, for example. Little can be learned that is of practical use for crafting intervention programs addressing homegrown terrorism, if the ideological and religious components are downplayed or ignored.

Reviewer Notes

Recommendations for law enforcement or intelligence practitioners.

53.5% of the subjects who took discernible steps toward violence subsequently followed up with some form of criminal action that was not averted through preventive arrest.

27.8% of those who immersed themselves with likeminded peers subsequently communicated some desire for violent action.

27.5% of those who undertook da’wah (proselytizing) online subsequently communicated some desire for violent action.

24.2% of those who communicated some desire for action subsequently took steps toward violence.

You might also like: