Statistics and Counterterrorism: Workshop Report

by David Banks and Ed Melnick

On November 20, 2004, the Stern School of Business at New York University hosted a workshop on Statistics and Counterterrorism. The workshop was jointly sponsored by the ASA Section on Risk Analysis, the ASA Section on Statistics in Defense and National Security, and the National Institute of Statistical Sciences. The co-organizers were Ed Melnick and David Banks.

The workshop explored ways in which statisticians could contribute to counterterrorism. It was a follow-on effort to the conference funded by the ASA in Washington, DC in May 29-30, 2003, and is a prelude to the year-long program in this area that will be started by SAMSI (the Statistical and Applied Mathematical Sciences Institute) in September, 2005.

The keynote address of the Nov. 20 workshop was given by Nancy Spruill, Office of the Under Secretary of Defense for Acquisition, Technology and Logistics and chair of the ASA Section on Statistics in Defense and National Security. Her topic was "Giving the Warfighter the Tools for Counterterrorism;" it laid out the kinds of statistical issues that arise throughout the development and supply pipelines, as well as strategies for making systems tests more accurate and acquisition more flexible.

Seven subsequent sessions focused upon risk analysis, sensor analysis, biosurveillance, data mining, and related topics. The other invited talks were:

Many of these talks are available at the National Institute of Statistical Sciences website. Just go to and follow the links.

The workshop closed with a panel discussion led by Jon Kettenring (former ASA President), Ed Melnick (NYU), Fred Roberts (Director of DIMACS), Cliff Spiegelman (Texas A&M), and Nancy Spruill (DOD). The broad conclusions of the panel were that (1) the Department of Homeland Security does not understand statistics, and could do their mission better if they did; (2) the syndromic surveillance field is rapidly becoming mature, but other aspects of statistical counterterrorism are ripe with research opportunity; (3) new collaborations between statisticians and operations research experts are needed.