Statisticians listen and give a voice to data

Adam Weyhaupt, Chair of the Department of Mathematics and Statistics at SIUE writes, Data can be used in many different ways and to answer many different questions. One statistician believes that in order to successfully use data, you must first be willing to “listen” to it."

 Dr. Andy Bartlett, SIUE Photo

"I'm going to listen to the data," said Andy Bartlett, assistant professor in the department of mathematics and statistics at Southern Illinois University Edwardsville. "I want to sit back, look at the data, do the analysis and see what it says. You've got to put your ear close to it and see what it tells you."

Bartlett's objective view of statistics is especially relevant since his research works to find answers to questions that are frequently subject to bias: Do global warming and climate change exist? He studies statistical models that focus on extreme values, not on traditional descriptive statistics like the mean or median that describe a typical experience. 

"We all recognize these extremes,” he said. “However, when you do analysis most people just want to use the average as an estimate. But you don't feel a little shift in the average from one month to the next. What you detect is extreme events – spikes if you will – like those days we all talk about when the temperature is 120 or negative 20 degrees."

He said that he tries to see if you can detect a change in the climate based on how frequently these extreme events occur. This is a relatively new approach since most techniques rely on so-called central estimators like the average temperatures.

"We set a threshold of what is extreme and see how many times that happens, and then use those observations to come up with estimates for whether there has been as significant change," he explained. Because of this innovative approach, Bartlett was recently awarded the Henry and Annette Baich Award for his project, "Detecting Change-Points Using Extreme Values in a Changing Climate."

Bartlett, who has a Ph.D. from the University of Georgia, said he was drawn to statistics in college because he was frustrated by “cookbook” style problems shown to him in his college statistics class. He continued to study mathematics and later was attracted to the ability to apply statistics to real-life problems. He said that he views one key role of a statistician as separating meaningful observations from noise in a large mess of data.

Read more...

Source: The Edwardsville Intelligencer