Former Skater Puts New Spin on Randomness

New statistics professor, Kari Lock, is excited to help Duke scientists and students improve their research skills.

Kari Lock, a new professor in statistics, interacts with her students during class in Love Auditorium. Credit: Megan Morr, Duke Photography.

After seven years as a professional ice skater, Kari Lock understands that when you fall down, you have to get back up if you're going to perfect a move, or an idea.A new professor of the practice in statistics, Lock, 29, plans to collaborate with scientists around campus to help them use a theoretical framework she developed to "improve the way we do experimental research," she says.

The framework gives researchers a way to check that the groups they will compare are actually comparable before an experiment begins. Lock says she is also looking to improve the way students learn statistics. To do that, she has co-authored a book with her parents and brothers -- all of whom are in statistics or math -- called Statistics: Unlocking the Power of Data."Kari is exceptionally dynamic and an exceptional communicator," says Alan Gelfand, chair of the Department of Statistical Science. He says Lock will likely collaborate with researchers in the social sciences and definitely add energy to Duke's introductory statistics classes.Both of Lock's parents are professors in math and statistics at St. Lawrence University and her two brothers are also studying statistics at the graduate level. Growing up in a family that spoke math almost as a second language, Lock became interested in numbers at an early age.Her first formal training in the field began during her undergraduate education at Williams College, where she majored in mathematics and statistics. She also skated professionally during school breaks. After graduating in 2004, Lock spent two years touring internationally with other skaters and performing in a show called Holiday on Ice.Upon returning to the U.S., she attended Harvard, where she earned her master's and doctoral degrees in statistics and developed her theoretical framework for randomization in scientific experiments.Randomization is the "gold standard" for judging and being able to identify causality, or whether one variable truly influences another, Lock says.Scientists randomize their samples at the beginning of an experiment to prevent bias in the results. But even in an arbitrary division of two groups, there's a chance that the groups will be unbalanced. In people, for example, a majority of males could end up in one group and a majority of females in the other group.To correct the imbalance, scientists can do complex statistics after the experiment. But Lock and her advisers wondered if there was a better way to use statistics in these experiments. Lock's idea, which became the subject of her dissertation, was to develop the theoretical framework for scientists to check that the groups were randomized before the experiment. She then developed the statistical techniques scientists could use re-randomize their groups, if the initial randomization did not provide adequate balance.Getting scientists to understand why and how they should run pre-experiment checks on randomization won't happen overnight, Lock says. But, just like in skating and teaching, she is prepared to be persistent as she helps scientists understand and use this approach.Helping people is as natural to Lock as math and statistics. At 14, she began coaching other ice skaters. "Teaching them to master new skills and transferring some of my love for the sport to them was really rewarding, and teaching skating naturally transitioned into teaching statistics, " she says.Sharing interests in teaching and mathematics with her family eventually led the Locks to conclude that statistics could be taught in a way that "we believe makes it easier for students to understand the concepts," she says. Lock's family feels so strongly about a need for change in the teaching process that they have collaborated to write a new introductory statistics textbook based on simulations rather than the standard plug-and-solve method.Both methods give the same answer, but simulations are "intrinsically connected to the fundamental idea of the method," giving the students a better conceptual understanding of what they are doing, Lock says.She adds that writing the book with her parents and brothers has been "fun," but the Locks do have different statistical backgrounds, creating disagreements. "Each section usually has a few heated arguments," she says, "but I think ultimately these differences in opinion help to make the book even better."The publisher Wiley & Sons plans to print the first official edition of the Lock family textbook in 2012.By then, Lock will have enjoyed one Durham's warmer winters, which she says she is looking forward to after years in Massachusetts.Lock is also excited to play ultimate Frisbee with her husband Eugene Morgan, a professor in the civil and environmental engineering department, and with their new female black lab puppy, Charlie.