A 61-Million-Person Political Experiment on Facebook

If you saw one of above two ads pop up on your Facebook page on November 2, 2010, you may have silently participated in an huge behavioral study. Researchers from UCSD worked with Facebook to decipher what kind of messages were more effective in changing people’s behavior. 

One group of people were shown a voting/polling place reminder with their Facebook friends’ pictures, and some were shown the same message without social connections. A third group saw nothing. By cross-referencing those names with actual voting data, a small but significant number of people were convinced to vote by viewing the reminder along with their friends’ pictures.

The 0.39 percent increase means that 238,000 extra votes were cast … quite enough to turn an election in a swing state. They don’t know how this may translate to other behaviors and ads, but it shows that on the scale of large social networks, the subtle boost we receive by seeing a message connected to our social circle can change behaviors in meaningful ways.

Want more? You won’t believe how big the effect was when they started to dig into how people acted when close friends were featured in their message. Check out more details, and some of the skepticism, at Not Exactly Rocket Science.

(image from the Nature paper describing the research, subscription req’d)

One of the most powerful things about large-scale social media is that even the smallest of influence changes (I mean, .39% is pretty small) can have huge effects. 

Not just for advertising purposes, but in cases like this where the difference could be the leader of the free world..that’s incredible stuff. 

In a continuation of this study, I’d like to see how reliable the answers are. Did they really vote, and were they more or less likely to lie if their friends were being shown.