Statistical Model in Airports Unlikely to Catch Terrorists

By Prof. Alan (Avi) Kirschenbaum*

Engineering researchers at the University of Arkansas have created a statistical model of the daily operations of general aviation airports in order to help show unusual activity that could suggest a security threat, according to a recent report.

The model is based on factors such as annual number of landings and takeoffs; total number of planes based at an airport; whether an airport has a traffic-control tower, and other detailed data.

“We want to understand the variation associated with usual general-aviation activity and operations, so unusual activity can be detected, analyzed and resolved,” said Justin Chimka, associate professor of industrial engineering and a researcher at the rural transportation center.

Catching potential terrorists by monitoring thousands of airport activities and then looking for blips from an artificial “norm” is like discovering a rash on your body (statistically abnormal), then trying to discover its origin, knowing full well that it may have originated from dozens of potential sources.

Anyone familiar with airport operations and the thousands of potential sources for “abnormal activities” will realize that the tunnel vision being promoted by the research supported by the DHS (U.S. Department of Homeland Security) to monitor airport security may look great on colored graphs (or monitors) but is unlikely to catch the terrorists until it’s too late.

The tunnel vision promoted by a stand-alone engineering or technological approach is simply missing the boat.

As airports are complex social organizations where employees like ourselves are trying their best to maintain continuity of operations in the face of potential security threats, the decisions we make – be they facing passengers or checking suppliers of services – are not based on statistical abnormalities but on those good ole-fashioned human qualities gained through training, experience, and to some extent, gut feelings.

Technology is a tool to help make those decisions, but we have enough research evidence to show that trusting technology and bending the rules are a critical part of the security decision-making process: a very human characteristic.

So, perhaps it may be a good time to start refocusing on people and not on a virtual world of statistical devices in catching the bad guys.

* The writer is the initiator and coordinator of BEMOSA (Behavioral Modeling of Security in Airports).