RandomICEation

It’s hardly random that Kari Lock skated her way into Duke's department of statistical sciences.

The newly appointed assistant professor of the practice of statistics, Lock spent seven years as a professional ice skater before developing a new theoretical framework for randomization in scientific experiments.

“[Randomization] is kind of the backbone of scientific research,” Lock said.

As a graduate student at Harvard University, Lock focused her research on improving randomization in randomized experiments—experiments designed to determine causation between two or more factors.

“If you want to see whether something causes something else, you randomly determine who gets which level of the variable, and see whether there’s an effect,” said Lock.

In a randomized experiment, it is possible that experimental groups will not be diverse enough after initial randomization efforts, she said.

“My research is looking at a way of checking [experimental group compositions] before [one does] the experiment, and [re-randomizing] if you get a bad randomization,” said Lock.

Lock was a competitive figure skater until she graduated high school, when she began performing professionally. For two years she toured full-time with a show called Holiday on Ice, traveling around Europe, Asia and Central America.

Lock said her experiences as a skater provided her with important insights for her academic pursuits.

“Just because something doesn’t work the first time doesn’t mean it’s never going to work,” said Lock. “You always fall the first 100 times you’re doing a jump [on ice], but if you keep working at it, you perfect it and it becomes second nature.”

Although Lock plans to continue her research, her focus at Duke is on undergraduate education.

“A lot of my time, and where my passion lies now, is in education,” she said. “I think a lot of my efforts are going to be towards making undergraduate statistics classes fun, informative and useful for everyone who takes them.”

Lock believes her work can help Duke faculty improve their research.

“I’m hoping to work with people in other departments because there are many different disciplines that work with randomization,” she said.

Discussion

Share and discuss “RandomICEation” on social media.