By Laurel White
Teaching people how to spot and outsmart a particularly tricky form of misinformation online is at the heart of new UW-Madison research recently funded by the National Science Foundation.
The research, led by Martina Rau, will focus on visual misinformation — specifically, misinformation presented online in graphs. Rau, an associate professor with the Department of Educational Psychology, has studied visual learning for more than a decade. She says visual misinformation in graphs can be particularly powerful because it can make false information seem more believable — and it reaches viewers very quickly.
“When you’re browsing the web, you encounter so many visual representations that aren’t good,” Rau said. “Every time we scroll through media, whether it’s social media or news media, we’re attracted to visuals. We’re more drawn to visual representations than text, and we don’t question those visuals as much.”
Rau said misleading graphs include three-dimensional pie charts that angle a certain section of the chart toward the viewer, in order to make that section seem bigger. A line graph with an axis that doesn’t start at zero is another culprit. Those graphs can cause viewers to accept falsehoods and make ill-informed decisions about things ranging from politics to medical care and product purchasing.
Rau’s project, which was funded for three years with roughly $850,000 from the National Science Foundation, will include the development of a tool that people can install on their web browsers. The tool will provide a short, two-minute training on how to spot misleading graphs — and point out those graphs as they show up during browsing. The goal is to “inoculate” people against misinformation in graphs, and then provide “boosters” as needed, Rau said.
Rau was drawn to the project after keying into the hot-button issue of widespread misinformation during a recent sabbatical. She sees the study as a way to use her deep knowledge of visual learning to help people in the world of education — and beyond.
“I wanted to do something that had a big societal impact,” she said. “If we could learn to not be fooled any more, that would be really what I would hope for.”
Xiaojin Zhu, professor in the Department of Computer Sciences, is co-investigator on the study.