Scientist at Work: Karen Kafadar
Determining the effectiveness of cancer screening programs, exposing flawed techniques for matching bullets used in crimes, understanding particle decay rates in high energy physics and reviewing FBI methods used to investigate the 2001 anthrax mailings. All of a sudden it's getting pretty cool to be a statistician, one might think, when looking at a day-in-the-life agenda of Indiana University Rudy Professor of Statistics Karen Kafadar.
Indiana University Rudy Professor of Statistics Karen Kafadar, with her Stanford and Princeton intellectual roots tied to two of the nation's greatest statisticians, jumped at the chance to move to IU in 2007 just a year after the university created its Department of Statistics.
The work being done by Kafadar, who also holds an adjunct post in another IU College of Arts & Sciences branch, the Department of Physics, exemplifies a clarion call to statisticians around the world heard from information technologists, geneticists, physicists and, yes, government leaders, that when you are working within massive sets of data you need every tool in the box to identify what Kafadar describes as "that tiny signal in a vast sea of noise."
And statisticians, it seems, are becoming the "go to" tool in the box as massive sets of data needing exploration have become available through super-computing and cloud computing applications. Hal Varian, the chief economist for Google, succinctly projected on the vocation in a recent interview with The New York Times: "I keep saying that the sexy job in the next 10 years will be statisticians . . . and I'm not kidding."
Kafadar arrived at IU only last year after accepting a professorship in 2007 but first completing a year-long sabbatical at the University of California Berkeley's Department of Physics that was hosted by UCB Professor Robert Jacobsen, a lead researcher at the Stanford Linear Accelerator Center.
Having just been created in 2006, an upstart Department of Statistics like the one confirmed at Indiana University was a chance Kafadar could not pass up.
"It was just a great opportunity to develop degree programs tailored to the types of problems that statisticians today will have to encounter," she said of leaving a professorship at University of Colorado-Denver. "A new department, to me, meant an opportunity to develop a modern statistics program that would include computer science, informatics and quantitative courses in many other disciplines."
Traditional statistics programs tend to rely heavily on the emphasis of mathematical proofs, for one example, which are important but can belie the need for broadening the scope of endeavor to include huge data sets, complex problems and other work best addressed by sometimes large interdisciplinary teams. The scope of that range of work reaches directly into the data sets generated by researchers in genomics, complex information technology networks and high energy physics.
"One of the goals of a statistician is to find those specific target events of interest amidst millions of 'uninteresting' events," Kafadar said.
Kafadar has studied with and been Influenced by two of the nation's most regarded statisticians, Bradely Efron and Stanley Tukey.
While at Stanford University earning a bachelor's degree in math and a master's degree in statistics she worked with Efron, recipient of the 2007 National Medal of Science, the United States' highest scientific honor. Efron is credited with proposing the bootstrap technique, one of the first computer-intensive statistical techniques.
Kafadar went to Princeton to earn a Ph.D. in statistics with Tukey, another National Medal of Science winner, as her adviser. Tukey, who passed away in 2000, is credited with introducing concepts central to the creation of today's telecommunications technologies and with introducing the terms "bit" (short for binary digit) and "software."
Kafadar then went to work at the National institute of Standards and Technology, moved on to a staff position in Hewlett Packard's radio frequency/microwave research and development division and then served as a fellow in the Division of Cancer Prevention at the National Cancer Institute before moving to UC-Denver and then IU.
She remains connected to cancer research, having published several research papers on biases in cancer screening studies with Philip Prorok, an award-winning researcher in the Division of Cancer Prevention and Control at the National Cancer Institute.
In 2004, she was a member on a National Academy of Sciences' National Research Council committee that issued a report discrediting FBI techniques for matching chemical signatures of lead in bullets between crime scene bullets and those possessed by suspects. In 2005, the FBI -- referencing the report -- ended the use of comparative bullet lead analysis.
"That was quite an impressive outcome for an NRC report," Kafadar reflected.
That involvement led to her participation in a Congressionally mandated NAS committee that in February released another much-publicized report, this one calling for major reforms in the nation's forensic science system.
"With respect to those findings, things are moving along," Kafadar said. "The U.S. Senate is still working on addressing issues in that report."
Now Kafadar has found herself on another NAS committee involving the FBI, this one to investigate and conduct an independent review of the science used to investigate the 2001 anthrax mailings. The committee met in June and September, and a report will be produced as a result of the ongoing study.
"Yes, this is a fascinating time to be a statistician," she said. "From forensic analysis to genomics and proteomics, and continuing my work on evaluating the benefits of cancer screenings, which has become even more timely in view of the recent revised recommendations from the U.S. task force on breast cancer screening. This is a fascinating time to be doing the work that we do."
