By Chris Carroll | Illustrations by Jason A. Keisling
Precise, deliberate, thorough—good descriptions for the population-based surveys that public health research often depends on, but a hilariously off-target summary of the daily discourse and diatribe on Twitter.
But that freewheeling, broad-based nature of the platform has its advantages, and now a University of Maryland public health scholar has found that our 280-character pearls of wisdom (or rants) translate to real-world health outcomes that can be mapped nationwide.
In a series of studies over several years, Assistant Professor Quynh Nguyen in the Department of Epidemiology and Biostatistics has shown that how the estimated 22% of American adults on Twitter discuss topics like food and exercise correlates to measures like obesity and mortality in various locations; a town awash in triathlon chatter, for instance, probably has lower blood pressure than one where barbecue discussion holds sway.
Although she researches Twitter primarily, she has also cross-checked Yelp and health data: “If you look at an area with a lot of hamburger reviews—high-caloric foods—you also find higher prevalence of diabetes and obesity.”
More recently, Nguyen has focused on social media expressions of racial sentiment. In a paper early this year, she and co-authors showed that areas with plenty of tweets dripping with racial animus had worse cardio health among minority and majority populations alike compared to more harmonious locales. A forthcoming study shows a similar negative effect on birth outcomes.
Twitter is far from a perfect cross section of America, Nguyen admits, but still, this platform for spontaneous sharing—and venting—provides “access to things that are very difficult or impossible to find in other data.”
