Tag Archives: elections

Yes, I know, a lot of people are still recovering from the 2012 Presidential election. People who watch television tell me that it’s really nice to be able to do so without relentless political advertisements.

Me?

I’m thinking about our local and school board elections, set for the beginning of April (just 4 months from now!), and about how, especially in these smaller races that don’t receive nearly the same media attention, the ways in which we communicate about the issues, and the candidates, and the importance of voting are even more critical.

And that got me thinking back to a study in the journal Nature (which always makes me think about the time my friend Tim had a paper published in Science, and told me that all the best journals have just one name, I guess kind of like Madonna?), about the impact of social media posts on people’s political activities and even their opinions. The big-time science-y types who get published in Nature did a study that included everyone who visited Facebook in the U.S., ages 18 and older, on Election Day 2010 (61 million adults). They found that political messages in a social context influenced not just users but also other friends who also saw them. Critically, the effect of the social transmission–the fact that the messages were delivered through a social network–mattered more than the content of the messages themselves. If we see those patterns hold up in future elections, you just may be saved some of those political television ads in the future.

For methodology types, here’s a little more detail on how it worked.

Most Facebook users that day saw a “social message”, encouraging them to vote. It gave them a link to local polling places, and clickable button that said “I voted”. They could see how many people had clicked the button on a counter, and which of their friends had done so. But the remaining 2 percent saw something different. Half of them saw everything the same except WITHOUT the pictures of their friends–the information, but without the ‘social’. The other half saw nothing. When they compared the three groups, in such a large sample size, the scientists found that the messages mobilized people to express their desire to vote by clicking the button, and the social ones even spurred some to vote. These effects rippled through the network, affecting not just friends, but friends of friends. (Best part alert): By linking the accounts to actual voting records, they estimated that tens of thousands of votes eventually cast during the election were generated by this single Facebook message. It was an increase of 0.39% in voting probability, just by seeing the social message. As the analysis of the study cited, “Facts only mattered when paired with social pressure.” Furthermore, when they crunched the ‘friends’ into more precise types–close friends, with whom Facebook users interact frequently, versus the more ‘regular’ connections with whom one might not have much (or any) face-to-face interaction, they found that the size of effects varied as one might expect. The more distant ‘friends’ influenced the odds that someone clicked the “I voted” button, but not the likelihood that a user investigated his/her polling place or went to vote.

Without the institutional subscription that I enjoy, you won’t be able to read the whole article, so here are the pretty cool points:

• Nearly all the transmission occurred between ‘close friends’ who were more likely to have a face-to-face relationship, and the effects were strongest there. It makes sense–I may get annoyed when my neighbor or my cousin post political content that I don’t agree with, but I don’t/can’t walk away from them. And if someone I respect points me towards information of which I am skeptical, it makes me think twice.
• The effects weren’t just on expression–what people posted themselves–but also on information-gathering (who goes to look for what information) and actual voter turnout. Those latter effects were more modest, but, still, with some of the razor-thin election margins we’ve seen recently, even small effects matter.
• The messenger matters–we know that it’s not just the quality of one’s information, but also the trustworthiness and relational power of the person(s) delivering it.
• Scale matters, A LOT. The messages themselves and the friends who shared their activity, collectively, accounted for about 0.14% of all the votes cast during the 2010 election. That’s more than 280,000 votes, from one Facebook message.
• One of the coolest things, to me, about this study, is how ‘real’ it is. People didn’t know that they were part of an experiment. They were just doing what they do every day–spending some time on Facebook–and, in the process, shaping their own (and their friends’, and even their friends’ friends’) political behavior. The implications are significant.

And, again, this was for a mid-term congressional election that was, after all, a pretty big deal. Most people, arguably, knew it was happening. There were many other messages in the arena, about the same election.

What about those smaller elections, where, if we knew what our friends were doing and knew that they would know what we, in turn, were doing (or not), we could see, maybe a few dozen votes in an area that we normally wouldn’t, in elections with historically very poor turnout?

Maybe we need some experiments of our own, four months from now.