When Alison Field, director of the Centers for Epidemiology and Environmental Health and chair of epidemiology, began working on the Growing Up Today Study in 1996, almost 17,000 children aged 9 to 14 enrolled. But as the study, which looks at the effects of diet and exercise on weight changes throughout a person’s life, progressed, she found that at around age 18, boys began dropping out of the study at higher rates than girls, which hurt the quality of the study’s results. To make matters worse, the researchers could not figure out how to improve the response rate of boys.
And now? “We haven’t gotten any firm answer,” Field said.
Field’s problems are not unique: The struggle to obtain representative population samples continues to plague researchers. “Across the board, it is getting harder and harder to get people to return surveys in academia,” she said.
Those from low-income backgrounds pose an especially difficult challenge. Middle- and upper-class families are more likely to have the time and means to participate in studies, increasing their response rates, Field said. And it’s sometimes difficult to track down low-resource individuals due to high mobility rates.
Minorities can be difficult to reach as well, particularly in randomized trials, said Ira Wilson, professor and chair of health services, policy and practice. Surveys that are meant to reach a wide spectrum of people often do not, he added.
Since researchers have faced increasing difficulties reaching certain populations, they have adopted a variety of methods to account for uneven representation in samples and improve participation rates in research studies.
Answers may be weighted to adjust for lower-than-expected response rates, said Joseph Hogan, professor of biostatistics and public health. As a result, an individual’s answer from a population with a low response rate will be counted more heavily than an answer from a better-represented population.
Additionally, researchers have increasingly turned to community engagement, a new type of research that focuses on participation and involvement from the local population.
In the last decade or so, researchers have looked to the community to discuss what issues should be studied, Wilson said. Researchers ask locals to take part in the research team as well, where they can encourage involvement and ensure that the study is relevant to the community. “If the community doesn’t know why something is being done, they might not be interested. It might be quite hard to enroll people in the study,” he added.
Wilson has been able to increase response rates to his studies through his use of community engagement, meaning that fewer groups are left out of research studies.
Amy Nunn, associate professor of medicine and behavioral and social sciences, strongly advocates for community engagement research. “I’ve always found it to be rewarding and worthwhile. It’s the only way I do my work,” she said. Nunn has used the process while researching and creating programs for the Rhode Island Public Health Institute, including Food on the Move, which provides mobile, discounted food markets in a variety of neighborhoods, and Do One Thing, which addresses issues faced by those with HIV.
But community engagement does not come without problems. “One of the big challenges that researchers have is that they don’t know how to engage people in research studies,” Nunn said. The process is extremely time consuming as well, forcing researchers to deeply understand the community being studied, she added.
Nunn spends a significant amount of time reaching out to locals, listening to their issues and working with them while conducting her research. “Sometimes that’s at the expense of my more academic responsibilities,” she said.
Despite these challenges, Nunn and Wilson believe that community engagement is key to the future of research studies that involve a human component. “Sometimes it’s easy to lose sight of the issues,” Nunn said. “But if you focus on what communities need and tailor your grants accordingly, then you usually don’t go wrong.”
