A new grant from the National Science Foundation will advance a project in the UO’s School of Journalism and Communication that is exploring how good people are with numbers, which in turn affects their decision-making.

Ellen Peters, Philip H. Knight Chair and director of the Center for Science Communications Research, received the $634,975 grant to further her research in objective numeracy and numeric confidence.

Peters, along with senior research associate Brittany Shoots-Reinhard and assistant professor David Markowitz, will research how people rely on their numeric ability and confidence as they make choices when numbers are involved.

“We’re looking at two ways that people can be good or bad with numbers,” Peters said. “And those two ways end up making a difference in how you perceive risks in the world and how you make decisions.”

Peters said the first way to tell if someone is good or bad with numbers is simply their math ability. She gives people a math test and sees how well they score. The second way is to see how confident they are in their numeric abilities, also known as their numeric confidence.

“Ideally, if you’re good with numbers, you would know it, and if you were bad with numbers, you would know it,” but that’s not always the case, Peters said. “So, you get these mismatches of how you think you are with numbers versus how you actually are. And it turns out that it makes a difference to what happens in your life.”

Peters said those who are good with numbers and confident in their numeric abilities end up enjoying better financial and health outcomes. On the other hand, people who are bad with numbers but have high numeric confidence end up making the worst decisions.

“Can we manipulate people’s confidence in their numeric abilities?” Peters asks. “We don’t want to do that out in the real world because we don’t want to actually affect someone’s life outcomes. But we can do it in the lab in a very simple way that doesn’t last very long.”

One way of manipulating numeric confidence is by giving very complicated math problems to people who have too much confidence in their abilities. If the problems are hard enough, they may walk away from the test thinking they should probably ask for help if they want to make the right choices.

Similarly, if you give easy problems to someone who is good with numbers but doesn’t feel confident, it may provide the confidence boost they need to try harder and take a little longer when it comes to making the right decisions to succeed.

The three-year grant proposal, initially submitted in January, was recommended for funding earlier this summer. However, due to pandemic-related budget cuts, the grant will not be funded until November, and then the tests will begin.

“What we normally do is a combination of experiments that we conduct online and in person,” Peters said. “We will just be putting off in-person experiments until COVID allows and everybody feels safe.”

The grant is not COVID-19 based, but Peters does have an ongoing coronavirus research project she started several months ago. She and her colleagues have been looking at the same group of Americans over time and seeing how their risk perceptions and emotional reactions have changed over the course of the pandemic.

Part of that is looking at how much they’ve exposed themselves to the media, including the numbers given to them.

“The problem with COVID is that the data are constantly changing,” Peters said. “The numbers are hard to understand, and we have inadequate data. In some ways, COVID-19 is the perfect exemplar of science. We learn from this, and then we realize we weren’t quite right. So, we have to backtrack.”

Peters said the reason some people don’t take protective measures from the virus may be because  they don’t know anybody who actually had it. Because of this, COVID-19 does not seem real to them.

“We rely most on our feeling minds, which are based on experience more than statistics,” Peters said. “We also have this huge political divide in the country where people are being told it’s a hoax by some radio announcers or government officials.”

She notes policymakers often throw numbers out at people because they think people can understand them. In reality, approximately 29 percent of the U.S. adult population can only do very simple processes with numbers, like counting, sorting numbers from biggest to smallest, and using simple arithmetic and percentages like 50 percent.

“People who are college-educated tend to be more numerate,” Peters said. “But we lose track of all the people who aren’t. So, that’s why we study what we study.”

—By Joanna Mann, School of Journalism and Communication