A research team led by Ali Shojaie, assistant professor of biostatistics, is investigating the connection between people’s social network and their physical health — specifically, their weight.
Shojaie’s proposal, which was put forward jointly with Hedwig Lee, assistant professor of sociology, and Tyler McCormick, assistant professor of statistics, is funded by one of five Emerging Challenges pilot grants the UW School of Public Health awarded in September.
Shojaie said the team’s goal is to make it possible for researchers to use social media as a tool for identifying groups of people who may be at risk for obesity and other health challenges. Ultimately, he added, this would give health organizations a clearer target and make helping those at-risk groups easier.
“Social-media networks can help us in so many ways, by providing information and helping us lead better lives in general,” Shojaie said.
Shojaie’s pilot grant was one of five awarded during the first steps of the School of Public Health Strategic Plan’s implementation. It addresses two of the emerging health challenges the plan identifies and combats — the social determinants of health and obesity.
“The emerging challenge areas the school is addressing are based on public-health significance, the ability to build on our formidable existing strengths at the school, and the impact on the health of future generations,” said Mark Oberle, associate dean for public-health practice, in the Sept. 28 press release awarding the grants. “These specific grant proposals were chosen based on their potential to lead us in future research directions to improve health and well-being locally and globally.”
These pilot grants differ from typical research grants in three key areas, explained Jeff Hodson, editor for the School of Public Health Office of the Dean. The grants are worth relatively less (up to $40,000); generally don’t recur; and are specifically designed to expedite innovative, experimental, and cross-disciplinary investigation into new avenues of dealing with emerging health threats.
“They’re meant to spur an idea and give [researchers] a head start,” Hodson said.
Shojaie’s project brings together a diverse combination of academic perspectives — statistics, biostatistics, and sociology — to help pioneer new methods of investigating the relationship between health and social media.
He said his team is one of the first to incorporate the perspectives of multiple disciplines into the analysis of social-networking data, and he hopes their research will lead to other interdepartmental studies on the subjects in the future.
“Our attempt is to see whether people’s social networks are associated with risk of obesity,” Shojaie said. “We’re hoping that this will trigger research and help from different groups across campus.”
Even though the researchers are trying to determine whether having friends who are overweight is associated with an increased chance of obesity, Shojaie made it clear that current evidence does not indicate that having obese friends or followers increases the chance of a person becoming obese themselves.
Shojaie said his team chose to focus on Twitter because it’s both widely accessible and capable of streaming huge volumes of information in a variety of forms at once.
Twitter’s accessibility is especially important, Shojaie said. Traditional surveys often can’t accurately account for a population’s diversity because the methods they use underrepresent many groups. Most notable among those are minorities, the demographic that Shojaie’s team is especially interested in investigating.
“At this point we are not limiting [this study] to any particular demographic groups,” Shojaie said, “but we are paying special attention to minority groups in particular in terms of obesity, since rates are higher in those groups.”
The fact that Twitter streams information constantly and in varying forms also makes it a good research tool, Shojaie explained.
“[It] has a wealth of information,” Shojaie said. “The text, the pictures, … all this could come together to give us a better idea of how factors relate with each other.”
One challenge for all social-media researchers is discerning fact from fiction in information shared through Twitter and other social networks. Because Shojaie’s team plans to develop a general, formalized tool for other social-media researchers, they needed to find a way to fix that, too.
The first step in this process is getting an idea of what percent of the information posted to Twitter is factual and what percent is embellished or inaccurate, Shojaie said. The team can then work to develop a statistical model that accounts for these factors.
Shojaie’s team also wants to follow up on previous studies that investigated whether a person who is obese is likely to follow or be followed by other people who are obese, and ultimately determine why that is the case.
Reach contributing writer Cara Skalisky at development@dailyuw.com. Twitter: @caraskalisky
Please read our Comment policy.
