A Virginia student just took home a huge prize in a national science competition

After a national search, a local student was named one of the winners in a major competition for STEM-inclined high schoolers.

Falls Church, Virginia-based Emily Ocasio took home second place in the annual Society for Science’s Regeneron Science Talent Search for high school seniors. After being named to the top 40 in February, Ocasio took the runner-up spot and won $175,000 for her project, “Demographic Correlates of Humanizing Language in Media Coverage of Crime.” Vienna, Virginia’s Ethan Zhou took seventh place, winning $70,000; Max Misterka of Harrisonburg, Virginia came in fourth, winning $100,000.

For her project, Ocasio, who attends the New School of Northern Virginia, examined how race, gender, age and other demographics impact how news outlets cover homicide victims. Using GPT-3, she tracked how the Boston Globe used humanizing traits when reporting about homicides in every issue between 1976 and 1984. Ocasio said she wanted to take a look at newspapers because she was interested in textual analysis and wanted to use data readily available.

But before all that, Ocasio told Technical.ly, she had a longtime interest in nonprofit work and the social sciences. She first got into tech after participating in a GirlsComputingLeague-hosted workshop to get young girls into STEM. After volunteering for a few years, she’s now the CEO of the organization.

“Meeting a bunch of other students only a couple years older than me who were doing such amazing, interesting projects was really what made me realize: ‘Wow, this is something that I can do. This isn’t something that only adults do,'” Ocasio said. “I think it was what opened my eyes to the tech world and how it could be applied to social science.”

In homicide reporting, humanizing coverage is when articles mention personal details about a victim beyond their death. In other words, it details how they lived, not just how they died, and while it is often positive, it’s not always. In one case, Ocasio said a newspaper mentioned that a victim was a “misfit” in school.

With her project, Ocasio found that Black male youth under the age of 18 were 30 percentage points less likely to receive humanizing coverage than their white counterparts. Over the age of 18, they were 15% less likely to receive that coverage. Black women between the ages of 19 and 34 were 23% less likely to receive humanizing coverage than white women, though there was no significant difference for other age groups of women.

“That’s beyond being statistically significant, just a very large difference that actually is meaningful in our lives,” Ocasio said.

Since coming in second place, Ocasio said students from around the country and the world have reached out to talk about her project. She said she’s already spoken to several younger students that were interested in social science but didn’t realize how AI and other research methods could complement it.

“I’m so appreciative to have had that [chance to] talk to students and inspire others about my work,” Ocasio said. “So I’m really grateful, I’m super happy.”

Emily Ocasio presents her project: “Demographic Correlates of Humanizing Language in Media.” (Courtesy photo)

Moving forward, she’d like to apply her research process to more modern data for updated research on humanized reporting. She’d also like to use a different location.

Beyond the competition, the senior will be putting the prize money towards her college tuition — she’s already been accepted to Harvard University, though she’s still deciding where she’ll be next year.

As she progresses into her next chapter, she’s also eager to see how GPT-3 grows as use becomes more widespread. Since many still haven’t tried it out, she thinks there will be a bit of a learning curve and humans will quickly confront questions about how much we want to incorporate AI into our world. She predicts a period where human work is undervalued and then a revival when AI’s flaws are revealed.

“These AIs aren’t always perfect in the same way that people aren’t always perfect,” Ocasio said. “So, I think that there’s going to be a learning curve, and I think that the tide is going to go back and forth a little bit on it before we end up in a good spot.”