The psychology and physiology of propaganda: A study of the radicalization of women

With the help of a $716,065 grant from the National Institute of Justice in 2016, professors from the University — in collaboration with those at other colleges and the Federal Bureau of Investigation — combined theory with science in their study of the radicalization of women by ISIL.

During a two-part project, Janet Warren, professor in the department of psychiatry and neurobehavioral sciences, aided by Gregory Saathoff, professor in the departments of public health science and emergency medicine and the FBI’s conflict resolution specialist, developed a framework to describe this phenomenon based on accounts of western women who devoted their lives to committing acts of terrorism.

To test their model, Warren and Saathoff turned to founding director of the School of Data Science Donald Brown for help with gathering and analyzing relevant data. Mojtaba Heidarysafa, a doctoral candidate in the department of systems and information engineering, also joined the research project during the data collection phase, in which the team enrolled women ages 18 to 35 from the University and wider Charlottesville communities who volunteered for the study and were compensated for their involvement. Participants attended one of two sessions in which researchers recorded their emotional and physiological responses to propaganda.

As a result, a general pattern emerged regarding women’s reactions to certain types of violent visual stimuli, confirming the possibility that a model could be created to show how women become radicalized. The researchers hope these findings could assist in preventing women from being recruited into terrorism. 

Each woman completed a consent form reviewed by the University Institutional Review Board. Researchers provided women with the ability to discontinue their involvement in the study at any point and reimbursed them regardless of whether or not they completed the study. Furthermore, each participant had access to a tamer set of images, of which two women took advantage, as well as guidance after the study if she felt deeply affected.

Warren noted that similar studies often focus on men. But her consultant work with the FBI’s counterterrorism efforts inspired her to consider the ramifications for women, Warren immediately considered the ramifications for women.

“I thought it would be really interesting, as we were all together in these meetings talking about risk assessment and terrorism, for the first time to start a study on women,” Warren said. “Most research done on very violent crime is done on men … I have often followed these huge areas of research and asked, ‘How does this apply to women?’”

Saathoff collaborated with Warren to understand the overall trajectory of women’s conversion to terrorism. He stressed the need to characterize the process not only as a means of pursuing criminals but also as a means to provide insights and strategies that could protect women from plots to recruit them.

“In society we often learn through media anecdotes, and that is one way to become aware of these issues, but to truly understand the situation, it’s important to understand how this occurs in a large number of cases,” Saathoff said.

First, Warren, Saathoff and their partners constructed a risk assessment model to outline the radicalization process. To successfully summarize and explain that transformation, the team hoped to find 300 women with sufficiently detailed court and investigation records but doubted the feasibility of that goal. When they uncovered almost five times as many women, researchers chose to analyze the 300 with the most robust court and investigative reports.

Over the next couple of years, Warren, Saathoff and their research team produced a risk assessment model that identifies not only aspects of a person’s life that could increase the likelihood of her radicalization, but also preventative measures to proactively impede radicalization efforts. The hope was to generate a cohesive framework for analysis of potentially dangerous individuals for use in government agencies.

“We were trying to do something that integrated what people in different countries were talking about, what we thought was interesting, what was in our academic research and make something that was more translational and international in terms of its broadness,” Warren said.

The model incorporates the theory that risk and protective factors can be viewed as two ends of a spectrum. The factors in and of themselves are not diametrically opposed aspects of a person’s life. Rather, their intensity and outlets determine someone’s likelihood to adhere to extremist beliefs.

Warren’s three-stage framework pinpointed these elements. Beginning with the propensity for radicalization, individuals can transition to mobilization and eventually action and capacity for terrorist exploits. Key factors that dictate progression from state to state include morality, self-regulation, setting, physical activity and perception of alternatives.

“Many women we have found who embrace this new identity are doing that because of unhappiness with their current situation,” Saathoff said. “Travel to another country is sometimes an escape, and it’s facilitated by those who would specifically tailor the message to the individual person … It’s remarkable how compelling and powerful that seduction can be.”

Second, the team looked to Donald Brown, founding director of the School of Data Science, for help with gathering concrete evidence to support their theory, specifically as it relates to propaganda. Warren cited propaganda as a powerful recruitment tool that can persuade people to completely change their outlook. Though many studies dissecting the conscious reactions to materials disseminated by extremist organizations exist, Warren emphasized the lack of research on physiological indicators of emotional responses.

“We actually know what people are thinking consciously, but we’ve got to try and grab some data about how they’re reacting unconsciously and see what’s more powerful, or we’re never going to get a handle on [propaganda],” Warren said.

According to Mojtaba Heidarysafa, doctoral candidate in the department of systems and information engineering who joined the research team for data collection, the initial phase of the data collection portion of the project consisted of 45 women probing the internet for content they believed could contribute to radicalization campaigns.

With the permission of the University Information Security Department and the Office of the Vice President for Information Technology, researchers enabled women to investigate the dark web, a conglomeration of networks with restricted access that facilitate anonymous, illicit interactions. The U.S. National Security Agency reports that terrorist organizations such as al-Qaeda utilize the dark web to maintain communication across their global network, as well as recruit new individuals to their cause.

“We were also able to [connect] women and allow them to go into the dark web if that was something they were interested in doing,” Warren said. “The University was fantastic in giving us the freedom we needed to pursue this.”

An additional 45 women who identified themselves as Muslim, non-Muslim conservative and non-Muslim liberal attended four sessions in which they viewed a series of images depicting extremist activities. During these sessions, researchers monitored eye gaze, pupil dilation, heart rate and galvanic skin response — which tracks sweat gland activity — to determine the arousal level and emotional state of participants.

While the first session contained neutral pictures of everyday objects to create a baseline for participant responses, subsequent sessions depicted white nationalist campaign posters and violent acts by jihadist, alt-right and alt-left groups.

“We showed them all of these pictures,” Heidarysafa said. “When looking at hangings or beheadings, all of them had a non-neutral arousal response … What we found was that no matter the background, there was a reaction.”

In fact, Warren, Heidarysafa and Brown contributed to a pending publication that suggests pupil dilation and the aspects of images that attracted women’s gaze were largely consistent across participants, regardless of their religious or political affiliations. In other words, the pictures play a larger role in individuals’ emotional response than certain major components of their identity. The authors of the study claim the commonalities point to patterns applicable to multiple cases of radicalization.

As this preliminary research project draws to a close, with a portion of the final papers and the results published in 2018 and 2019 and more to come in 2020, Warren said she and her colleagues already applied for an additional grant to expand their efforts and replicate the study with men. In doing so, Warren aims to take full advantage of the resources at the University to create materials useful for the FBI’s observations of and interventions for at-risk individuals.

“Usually academic researchers don’t work with the FBI, and the FBI doesn’t necessarily want to work with academic researchers,” Warren said. “The most important thing is the relationships exist so that we could do this research.”