Statistics Professors Scott Holan and Christopher Wikle have received a $2.85 million grant to develop new research models that will improve census demographic data analysis.
When Holan and Wikle opened the email notice for their award, they couldn’t resist celebrating. The award, after all, represents the largest grant ever funded in the statistics department.
“We both saw those emails simultaneously, and we immediately converged in the hallway and gave high-fives,” Wikle said. “It was pretty exciting.”
The grant, funded by a partnership between the U.S. Census Bureau and the National Science Foundation, focuses on the American Community Survey (ACS), an ongoing survey that replaces the constitutionally mandated decennial census long-form. MU is one of eight universities to receive NSF-Census Research Network’s federal grants this year and will collaborate with The Ohio State University.
In previous decennial censuses, long forms were sent out to collect demographic information on different regions. These forms, Wikle said, were only collected on a decadal basis, whereas the ACS provides a continuous stream of up-to-date information throughout the decade.
The team of statisticians will borrow methodology from spatio-temporal statistics, which looks at the relationship of data over space, location and time. By comparing the data accumulated among contiguous counties, Wikle and Holan say they can generate more accurate estimates of ACS demographic quantities.
“If there is some similarity between neighboring counties, which is common, especially in a rural environment, and if there’s strong dependence between those measurements, it’s like getting more information,” Wikle said. “Instead of sampling as many people in one county, you can sample what you can afford. Because they’re related to what happens nearby, and if you combine [the data] in a clever way, you get more bang for your buck.”
The approach is called “borrowing strength” and will provide cost-effective means with better results.
According the U.S. Census Bureau’s website, ACS data determine the allocation of $400 billion in federal revenue. This money incorporates many important government programs, including school lunch programs, No Child Left Behind and Small Area Income and Poverty Estimates.
Wikle and Holan say the grant money will strengthen their department and give students invaluable paid-research opportunities. In the project’s five-year span, Holan said two postdoctoral students, about four graduate students and up to a dozen undergraduates will work on the project in a variety of positions.
The grant has generated buzz among the statistics majors, Holan said. Many have approached him after class to discuss the research and are excited to apply for positions. Holan said he hopes the grant will build a stronger relationship among statistics majors from the different degree levels.
“It’s a nice way to build a hierarchy,” Holan said. “We can have faculty mentoring the postdocs, graduates and the undergraduates. The postdocs can mentor the graduates, and the graduates can mentor the undergraduates to foster a nice balance.”
Wikle, director of undergraduate studies for statistics, said the research provides rare opportunities for undergraduate statistics majors.
“It’s unusual in statistics to have the opportunity to actually fund undergraduates in research,” Wikle said. “We specifically put it into the grant because one of the goals is to train people to get involved in the federal statistics system.”
The statisticians will work during the next five years and report their findings to the NSF-Census Research Network.
— Trevor Eischen