Splintering the cell: learning how to disable terrorist networks

Before Sept. 11, 2001, few people could have imagined that a terrorist cell in the United States could cause so much death and destruction. Located throughout the world, these terrorist cells are social networks tied together by a few people. A University of Missouri-Columbia researcher, through a grant from the Office of Naval Research, is examining these terrorist networks and identifying the major players in each cell in an attempt to infiltrate and fragment them permanently.

"The idea is that if we can identify the key threads to a network and eliminate them, then the network will be split so severely that it won’t be able to optimally function," said Douglas Steinley, assistant professor of Psychological Sciences, who is conducting his research with Stanley Wasserman of Indiana University. "The trick is trying to make inferences about these networks when they’re trying to hide themselves from us."

Through a series of mathematical equations and charts, called dependence graphs, Steinley is developing new techniques for analyzing these networks, even when important data is missing. Dealing with missing data, Steinley said, is common when observing covert networks. Steinley uses two research methods that examine the density but not the size of the terrorist network in order to determine how networks are connected.

For example, Steinley examined the terrorist data that defined the network among the hijackers in the Sept. 11 attacks. The network consisted of 19 hijackers spread among four groups. Through this analysis, Steinley identified the four most "important" terrorists in the network.

Once he removed these key individuals from the network, Steinley found that 64 percent of the ties formed had been severed. Additionally, the original, fully connected network was reduced to six isolated sub-networks.

2004

Additional links:

Douglas Steinley
Department of Psychological Sciences

Stanley Wasserman, Indiana University

<< back to news
<< back to archives