In This Story
Inside the atrium of the Nguyen Engineering Building at George Mason University, a pile of 100 pillows stopped people in their tracks this spring. Some students thought they were sandbags, others wanted to jump into them, and employees asked if they could take a few home. The installation’s purpose was far from anything fluffy, however: Each pillow represented a single person entering homelessness in the Fairfax area over a 10-day period, transforming a statistical model into a hard-to-miss reality.
The project was created by graduating applied statistics major Aleah Talley with guidance from assistant professor Jonathan Auerbach. It was funded through George Mason’s Office of Student Creative Activities and Research (OSCAR).
“So much of what we do centers on measurement,” Talley said. “The way we choose to measure things can determine how certain groups of people are statistically represented.” Talley, who also minored in sociology, said her interest in homelessness began years ago while studying addiction, mental illness, and social systems through documentaries and ethnographies. “I’ve never been interested in doing math or statistics just for the sake of math,” she said.
Using homelessness data from the U.S. Department of Housing and Urban Development (HUD), Talley examined the difference between what statisticians refer to as prevalence and incidence. HUD’s widely used “point-in-time” counts provide a snapshot of how many people are experiencing homelessness on a single night each year. But Talley believed that approach risked missing the larger flow of people cycling in and out of homelessness over time.
“Homelessness is dynamic,” she said. “People enter and exit it all the time. It’s messy and not necessarily something that you may be able to fully accurately capture within a single snapshot number.”
Auerbach said this distinction is important but often overlooked. “For example, the incidence of car crashes is high because they happen frequently,” he said. “But the prevalence is low because active crashes are often resolved quickly, so few are observable at any given time. It is important to remember that just because we don’t see them doesn’t mean they aren’t happening.”
To explore homelessness differently, Talley used a queuing theory framework known as the Erlang A model, often used to study systems where people wait for services or resources. In this case, the model represented people waiting for shelter beds or housing opportunities to become available.
The resulting estimates suggested that while roughly 100 people might appear in a traditional unsheltered count on a given night in Fairfax, the total number of people experiencing homelessness over time was substantially higher. “So she was able to figure out how many people were missing and invisible in the official count,” Auerbach said.
The pillow installation itself evolved through trial and error. Early ideas included a light board and large stacked columns of pillows resembling a bar chart. Eventually, practical limitations led to the sprawling arrangement displayed in the engineering building. “It did stop a lot of people,” Talley said. For Auerbach, that reaction became part of the project’s success. “Few people would take the time to read a poster,” he said. “But it engaged a lot of people who first asked, ‘What are all these pillows doing there?’”
The exhibit also reflected Talley’s interest in blending statistics with art and social science. “It was validating to see that something artistic could be so intuitively integrated into the field. Most people kind of recoilwhen I tell them that I’m studying statistics,” she said.
Auerbach hopes the project also changes perceptions about the discipline itself. “Statistics is more than just mathematical theorems,” he said. “How we choose to count something changes how we think about it. The field of statistics teaches how to balance the competing considerations of data collection and analysis.”
After graduation, Talley plans to continue her studies, with the long-term goal of earning a PhD. But she said the broader lesson from the project extends beyond statistics. “No one is immune to this condition,” she said. “We’re all so much closer to being homeless, unfortunately, than we are to being billionaires.”