Can statistics keep your online information safer? Or predict if you’re likely to be readmitted to the hospital after an initial visit?
Yes, according to the research that Klaire Roggeman, a statistics major in the Kenneth P. Dietrich School of Arts and Sciences, has conducted as an undergraduate.
“In general, statistics plays an incredibly important role in research,” says Roggeman. “In any field, statistics allows for significant, valid conclusions to be made and for trustworthy predictions of future trends.”
Applying Statistics Across Disciplines: From Health Care to Online Security
Roggeman's initial exposure to research involved applying statistical analysis to the health care field—a field that she is interested in pursuing professionally. In a data science class taught by Assistant Professor Lucas Mentch, Roggeman and a group of classmates were charged with predicting whether an individual would be readmitted to the hospital after an initial visit. The group found a data set online and created different statistical models to try and decide which variables were important in prediction, and to see how accurately they could predict readmission rate.
“That’s the beauty of statistics,” says Roggeman. “It can apply to virtually anything you want it to.”
For her next research endeavor, Roggeman focused on forensic applications of statistical analysis during a two-month Summer Research Project through Carnegie Mellon University. Roggeman was one of only 12 people nationwide to be selected for the program.
"The project I was placed on dealt with typing and handwriting,” explains Roggeman. “We had data that consisted of people’s holding and transition times of different keyboard keys (letters/numbers/symbols) while typing a predetermined ‘password’. As a four-person group we created a statistical model using this data that was able to look at a random person’s typing pattern to decipher whether the pattern was typed by a pre-specified ‘Person A’ or ‘not Person A.’
“Research like this could be applied to online password security,” Roggeman continues. “If a Web site’s login page could learn a user’s typing pattern, they could use this typing pattern as an additional layer of authentication to protect from hackers.”
Research in Action: Statistical Consulting
Roggeman's undergraduate research, though, isn’t limited to hypotheticals. She’s already putting it into practice in the real world: Klaire is one of several undergraduate interns who work directly with clients from across the University when they need statistical assistance with their research, a program offered through Professor Allan Sampson’s statistical consulting class.
“The clients we work with come from all different fields—from pharmacy to education to engineering—but all of them need statistics to be able to make any sort of conclusions about what they find,” says Roggeman. “I’m so thankful to have had this opportunity because it has really showed me how statistics can be applied to any field.”
Stats' Surprising Connection to Communication
One field that Roggeman was surprised to find so closely intertwined with statistics is communication. “Communication is such a huge component of statistics, and the statistics classes I’ve taken have emphasized this,” she says. “I’ve had to write papers for a statistical audience, and papers that should be easy to read to people without a stats background. I’ve had to give presentations to people with varying degrees of statistical knowledge. I’ve collaborated with clients who knew more than me, less than me, and about the same as me in terms of statistical background. All of my classes require communication skills. Without effective communication skills, statisticians would be unable to share any of the exciting methods, models, and predictions they create.”
Roggeman is completing a five-year combined bachelor’s/master’s degree in statistics.
“I wasn’t entirely sure what I wanted to do when I came to college,” says Klaire. “It took me until the end of my first year in college to realize what a great field statistics is, and that it’s what I wanted to pursue. My academic advisor, Carl Bodenschatz, was incredibly helpful during my transition into the statistics department and in helping me pursue the combined bachelor’s/master’s program.”
Roggeman says that working with the statistics department is one of the most rewarding experiences she has had at Pitt, not just because of the support and guidance she has received from the department’s faculty members, but also because of all of the opportunities this dynamic department offers to help students get involved at Pitt. She is particularly excited to be a founding member of DATAs, a new statistics and data science club launching this fall.
“Try out lots of things,” Roggeman encourages her classmates. “But stick to what you are passionate about—put your time and energy into the things you like.”