The public policy honors thesis program draws students with a diverse array of passions, all interested in pursuing original research.
This academic year, the program attracted 30 students, a 10-year high, said Ken Rogerson, associate professor and thesis program director for the Sanford School of Public Policy
To accommodate the large number, both Rogerson and Adam Hollowell, an adjunct professor, taught a section of the thesis course. The course guides students through the research and writing process and provides a community for collaboration and support. Each students also seeks out a faculty advisor.
“The value of the program for the student is not the end product so much as the process of figuring out their own research project,” Rogerson said. “It makes them a better student, employee and potential graduate student.”
Sanford’s thesis program differs from other undergraduate thesis programs at Duke because it starts in the spring of a student’s junior year and concludes in the fall of senior year. This gives students the summer to work on their projects. Each student also receives $200 to support data gathering.
Rogerson encouraged students to evaluate their interest in writing a thesis by considering if there is a policy topic they might want to study more deeply. The honors thesis is a big commitment, he said.
“You have to ask yourself, ‘are you willing sacrifice blood, sweat and tears to take ownership of the project?’ If so, the outcome will be more rewarding than you could ever imagine,” he said.
Here are the stories of three of this year’s honors students.
As a cadet in the Army ROTC program, senior Amy Kramer knew she wanted to do research related to gender and the military.
“Based on anecdotal evidence and my own experience under female leadership here at Duke, I asked the question, ‘What impact does the gender of the leader have on cadets? Particularly if there’s a female leader, what positive impact might that have on female cadets in ROTC programs nationally?’” Kramer said.
Originally, she was worried that this question would be taboo, but she was pleasantly surprised when the Army leadership told her they were curious, too. In March 2017, she requested their permission to administer a survey to cadets in North Carolina. ‘No’ was the reply from leaders, citing the small sample size. Instead, she’d have access to perform a national cadet survey.
With a 54 percent response rate, her survey received more than 3,000 replies. During the summer, she interned at Cadet Command (ROTC Headquarters) at Fort Knox, Kentucky where she completed much of the data analysis and worked on other projects related to representation in the military. When she finished her report, she had the opportunity to brief senior civilian and military leaders of Cadet Command.
She was surprised to find that ROTC programs headed by women have no statistically significant positive impact on female cadets. However, there was a negative relationship between female cadets and male ROTC leaders. In these programs, female cadets did not think their leaders had an interest in their success and were significantly less confident about their own capabilities.
"Gender-based perceptions matter," Kramer said. "They can cause attrition." She also found that when leaders rank all the cadets—a process that already accounts for physical differences between genders—male leaders still rank female cadets lower on average than their male peers. This has implications for their jobs in the Army after graduation.
"But women play a critical role in the Armed Forces," Kramer said, "especially as greater segments of the U.S. population are unable or uninterested in serving."
“The United States military, as is, cannot meet emerging threats and cannot continue to maintain readiness without utilizing female talent,” she said.
When she presented her results to Cadet Command leadership, Kramer said the officers asked questions and took the results seriously. They now plan to continue looking at the issue for several more years as part of a mandatory annual survey.
Kramer, a Robertson Scholar, was also recently named a Schwarzman Scholar, which provides recipients an opportunity to study at Tsinghua University in Beijing. She said her thesis project was instrumental to her plans.
“I’ve learned so much and I want to keep working on it,” Kramer said, praising her advisor Kyle Beardsley, associate professor of political science, for his help.
Her biggest piece of advice for other students in the thesis process? Don’t be afraid to tackle big questions.
As a public policy major in the pre-med track, senior Christopher Hill’s previous research experience was in microbiology. But in the wake of the 2016 elections, he decided he wanted to study fake news.
Although quantifying its impact on vote totals is likely impossible, Hill focused on analyzing the content of fake news itself to see how it influenced the public.
“The ultimate question I started to ask was were there differences in the fake news that liberals had been consuming versus the fake news that conservatives had been consuming?” Hill said.
He took both a quantitative and qualitative approach to answer the question. First, he used a database of fake news stories compiled by Stanford University researchers after the election to categorize how “fake” each was. With categories ranging from “distortion” to “absurd,” he found that liberal and conservative fake outlets both had the same distributions among the various types of fakeness.
Yet his analysis also confirmed previous studies suggesting that there were at least three times more pro-Trump fake news stories than pro-Clinton stories. In only about one percent of the stories was the bias difficult to find.
Hill—advised by Professor Judith Kelley—undertook a qualitative analysis as well to study subtle trends within his data.
“We found that the pro-Trump fake news on average had been written with a higher level of professionalism and with a greater amount of supporting ‘evidence,’ if you want to call it that,” Hill said. “It had been presented in a much more authentic and deceptive manner.”
Some of the most common themes that recurred in the data centered on Islamophobia, and the finances and personal health of the two candidates. A large swath of the stories continued to falsely suggest that Hillary Clinton had Parkinson’s disease.
Hill said he saw his thesis project as an opportunity to apply what he had been learning in class and drive his own work. Although the experience might not be directly relevant to his post-graduation interests, Hill said it had opened his eyes to how exciting good social science can be.
In this case, his research reflects a problem the United States must deal with critically.
“There’s not much to be said about who was telling more lies during the election. The point that I was trying to make is that people nowadays are consuming media tailored to their own preferences. This creates dangerous echo chambers and environments ripe for fake news.”
Senior Kelsey Sicard knew she wanted to combine her interest in health policy with a quantitative-based thesis.
Her research focused on how students engage with health care systems. Young adulthood is a formative time for making health habits that last for life. In recent years, public health research has helped define a “health engagement” model that focuses on the factors critical to maintaining health. Although people often remember to think of health in terms of physical wellness or healthy behaviors, they may forget that engaging with providers and understanding resources are just as important.
“Health engagement is more of an active promotion view of health,” she said. “Instead of waiting for health to fail and then fixing it, it’s about keeping it up all the time.”
Much of the research on young adult health engagement has tended to focus only on risk behaviors, such as drinking and smoking, at four-year colleges. Though this research is valuable, it lacks a wider focus on other engagement behaviors and ignores a large part of the student population—those at other types of schools who tend to be students of minority and lower socioeconomic status.
Sicard, who was advised by Rogerson, used survey results from the Duke Center for Research on Personalized Health Care and focus groups from students at Duke, North Carolina Central University and Durham Technical Community College to gauge how they understand health and what their barriers to accessing resources were. Students at each school gave different responses.
At Duke, students overwhelmingly knew that resources were available but had complaints about accessing mental health services. At NCCU, students felt that only some people cared about maintaining health. And at Durham Tech—the school with the highest average health engagement score—students complained about a lack of healthy food or access to exercise facilities. Although Durham Tech students were most engaged with their health, Duke students reported the best overall health, suggesting that Tech students’ scores came from their fight to access resources.
“A lot of students talked about Medicaid and so they had to know how to interact with that system because it wasn’t something where they just checked a box and got insurance,” Sicard said.
Students at all schools voiced concern that mental health wasn’t as emphasized in minority communities. At each school, students expressed confusion with the insurance system and saw it as a barrier to good health.
Overall, her results showed a correlation between health engagement scores and better physical health. More broadly, Sicard associated socioeconomic and minority status with lower health engagement scores.
Conducting focus groups at other schools was a challenge, as being a Duke student limited her understanding of the problems faced by those students. And from some focus groups she couldn’t glean enough useful information to make a conclusion. The experience taught her a lot about qualitative research.
“When you’re working with people, you never know what you’re going to find,” Sicard said. “You go in with one question but then you get hit with all different points of view and you end up going down a different avenue than you thought you would, which is definitely a positive thing.”
Going forward, Sicard said she wants to keep looking for ways to use numbers and statistics to improve people’s health outcomes, but her project has taught her to value the human side of public policy.
“Numbers can give you a concrete organization strategy and provide a starting point, but change can be driven by understanding how those numbers come about,” she said.