Math grad studies risky business

Posted by
On April 12, 2004

How technical professionals who work in hazardous environments communicate the risks of their work can be a risky business, according to research by a recent UMR mathematics graduate.

Using theories about selective and biased reporting and statistical models, Sarah Klein of Tulsa, Okla., a December 2003 graduate, explored how the scientific and technical community communicates risks, how the media reports them, and how these risks are perceived by the public. Her work, presented in August at the Joint Statistical Meetings in San Francisco, details how professionals — and everyone else — often work from limited or biased information when it comes to evaluating risks.

"Individuals may make irrational or incorrect judgments regarding risk probabilities if basing their assessment on face-value information rather than full information," says Klein. "Our research raises questions about how risks are perceived. We found that society thinks about risks differently than technical experts … who need to be careful about how they communicate those risks to people.

"Even though there is a really low risk of something occurring such as a nuclear accident, people are much more worried about that than say, car accidents," she adds. Individuals who judge risks associated with "hazardous engineering systems such as automobiles, airplanes, and nuclear power plants … often have limited amounts of information from which to draw risk estimates."

Klein’s research is detailed in a paper titled "Counter-Intuitive Probability in Risk Assessment." Her work was funded in part by UMR’s Opportunities for Undergraduate Research Experience Program, which allows undergraduate students to work with faculty on joint research projects. Klein’s project began from discussions with a nuclear engineering professor and his students (also supported via OURE) about how the public perceives risk associated with hazardous technology.

"My advisor, Dr. Gary Gadbury, approached me about doing this research. We talked with professors and students in the nuclear engineering department, and being nuclear engineers, they were interested in how people perceive the risk associated with nuclear technology or other complex technologies. They wanted us to research it from a statistical perspective," says Klein.

According to Klein, individuals often receive selectively reported information, with a tendency to hear about studies that find significant, or positive results, over those that produce mundane results.

"We found in our studies, for example, that there will be 30 studies done about causes of cancer and if two of them show significant results, those are the only two that are published, where there may be 28 studies that didn’t show results."

Klein’s research has echoes of psychology, sociology and economics, among other disciplines. She also conducted a detailed analysis of a well-known statistical problem known as "The Prisoner Problem," in which the survival probability of three prisoners is analyzed based on the reported knowledge that only two of them will be executed. "We had not seen the details ever actually published or discussed … or the math worked out," says Klein.

Klein says the experience of conducting her research and presenting it has been a great opportunity. "The whole experience of doing this research was really interesting. I presented at the Joint Statistical Meetings in San Francisco in August, in front of statisticians and Ph.D.s from Harvard, the U.S. Census Bureau and the Department of Labor Statistics, so that was really neat to be able to present with them.

"The experience of being an undergraduate at UMR and getting to do something like that was pretty amazing."

The Joint Statistical Meetings involved members of the American Statistical Association, the Institute of Mathematical Statistics, The Statistical Society of Canada, and the International Biometric Society and is the largest gathering of statisticians in North America.

Advisor Gadbury, an assistant professor of mathematics and statistics, emphasizes the significance of Sarah’s efforts as an undergraduate in the presentation of this research. "Sarah’s presentation was not in a student session, but was in a regular session with established researchers. While it is not uncommon for a graduate student to present at these meetings, it is relatively uncommon for an undergraduate to do so," says Gadbury.

Klein plans to continue this research, focusing more directly on investment risk.

Share this page

Posted by

On April 12, 2004. Posted in Research