How do violent extremists use social media platforms to spread ideologies? And how do individuals then embrace those ideologies and become radicalized online to take violent action?
These are questions that two Penn State College of Information Sciences and Technology researchers are exploring through a three-year, $250,000 project funded by the National Science Foundation.
Through the project, associate professor Anna Squicciarini and associate teaching professor Peter Forster, in collaboration with researchers at Texas A&M University, will analyze a data set of 17 million tweets attributed to Islamic State extremists since 2015. They aim to build mathematical and data-driven models to understand the dynamics of this extremist group at scale and the patterns of its influence.
"We are focused on online deviance, de facto peer pressure and influence," said Squicciarini, co-principal investigator on the project. "We are looking at what dimensions could affect a user's change of behavior, and how a user adopts a certain choice to relay a message, post a picture or support an ideology."
The researchers, in collaboration with Christopher Griffin, associate research professor at the Applied Research Laboratory at Penn State, aim to develop a predictive computing model to drive short-term forecasting. They will then use this model to analyze the types of conversations that these accounts facilitate, as well as how widely the ideologies are spread.
The results could be used to inform law enforcement and government agencies and to guide future interventions, according to Forster, who has spent his career researching terrorism, international relations and national and homeland security.
"When we look at deviant behaviors online, we can begin to identify indicators and warnings to provide us with an opportunity to intervene," he said. "Are there things that we can begin to identify in what individuals are talking about and with whom they're communicating that can give us early warnings?"
Forster added that while similar studies have been done to explore social media posts, his group's study is unique in terms of the interdisciplinary nature of the research as well as the size and time frame of the data set and that many of the tweets were collected in Arabic, which offers additional insight.
"Most previous studies don't have the algorithmic or computer skills, or the computer science background, to really begin to parse this data," he said. "So the fact that we're going to translate it and then we're going to be able to look at such a huge amount of data is really unique."
