While hostile tweets are already an excellent predictor of someone having a sports team’s logo as their background image, it turns out they may also be useful in predicting a higher risk of heart disease. A new study conducted by researchers at the University of Pennsylvania found that “negative” tweets—those full of profanity, containing the word “hate” and the like, or made in reply to @TheAVClub, say—correlate with a higher prevalence of deaths from heart disease in their county of origin. Furthermore, analyzing those tweets proved to be a better predictor of heart-disease rates than traditional factors like smoking, obesity, diabetes, and income and education—combined. In fact, the study suggests that the only downside of using Twitter as a data-mining tool may be that it requires looking at Twitter.
The study was conducted as part of a larger investigation into using Twitter as a research tool, following similar studies that found that looking at people’s tweets could effectively track the spread of people who are infected by the flu, in addition to fake stories about celebrities dying. Much as laborious, time-consuming journalism can now be replaced by looking at Twitter, researchers hope that analyzing social media can also replace things like taking representative samples and setting up surveys. This will leave more time for tweeting.
For this particular study, researchers made a random culling of tweets from across 1,300 counties (counties that had enough tweets to measure, anyway, with vast swaths of the Midwest apparently still preferring the old-fashioned lifestyle of keeping their thoughts to themselves). They then sorted those tweets using filters that looked for things like “hate, hostility, and boredom.” Then they made a handy little map of where the shittiest, angriest people live. That map was then placed side-by-side with another map tracking fatal coronaries, as collected by the Centers for Disease Control. Perhaps not surprisingly, it seems people who spend their days grousing on Twitter are all but marked for death.
Still, the study acknowledges that, due to the median age of Twitter users being around 31 years old, those people specifically aren’t going to die anytime soon. (You’ll just have to keep blocking them.) It also admits that the data may be slightly biased, given that many people tweet negative things as a way of cultivating a particular online persona, all in an effort to become that coolest of things: a Twitter badass.
Nevertheless, the researchers feel secure in saying that negative tweets are indicators that the community those tweeters live in is a stressful, grumpy one, where negative energy is spread from person to person like a “psychological contagion.” (The Washington Post, borrowing an analogy we’ve used ourselves from time to time, compares this bubbling Internet negativity to the “mood slime” in Ghostbusters 2.) And that knowing which towns have the most pissed-off people in them provides a fairly good indicator of the towns where they’re going to drop dead of a heart attack, if they keep it up like that.
“These people are the canaries of the psychological profile of their communities,” said the study’s lead author, Johannes Eichstaedt, referencing the birds that coal miners used to take down into the tunnels with them, which would then just sit and tweet whatever came into their heads all damn day, until eventually someone flooded the tunnels with methane to end their collective misery. All told, it seems Twitter could be a useful way of measuring how miserable our own working-in-a-coalmine lives are in our respective communities, and thus just how much longer we have to wait until death frees us all.
Next, those same researchers plan to look at what data regarding preventative health measures can be mined from positive, optimistic tweets, provided they can find any.