The 2010 elections: Twitter isn’t a very reliable prediction tool


This article was originally on a blog post platform and may be missing photos, graphics or links. See About archive blog posts.

There’s nothing quite as much fun as watching the election returns on CNN, whose newsroom, packed with all sorts of goofy push-button video monitors and high-tech gadgetry, looks like it could double as the set for the next ‘Star Trek’ movie. I guess all that technology didn’t entirely go to waste, as CNN proved during its coverage of the high-stakes Nevada Senate battle between Senate Majority Leader Harry Reid and ‘tea party’ insurgent Sharron Angle.

It was an ugly race, with each side slinging mud left and right. In the end, Reid won surprisingly handily, perhaps thanks to a host of Angle whoppers, including an interview she did in June in which she said that if she were counseling a 13-year-old rape victim who was considering an abortion, she would tell her that ‘two wrongs don’t make a right.’ But CNN came up with the most fascinating social media take on the race, illustrating the fluctuating opinions that voters had of each candidate through the number of tweets sent that registered support of or opposition to each candidate.


As you can see from this interactive chart, CNN found nearly 572,000 tweets involving the two candidates. As it turned out, Reid was hardly a Twitter favorite; 54% of the tweets were anti-Reid while only 12% were anti-Angle, while Angle edged out Reid among favorable tweets, 11% to 8%. You can play around with the chart yourself to get an overview how many tweets were sent during each week of the campaign.

But one big question remains, which CNN didn’t seem prepared to answer: Why did Reid win the most votes if he got so thoroughly trounced in the Twitter race? Does that mean that candidates such as Angle, who was clearly propelled by passionate ‘tea party’ supporters, were natural magnets for Twitter-friendly voters? Were Reid voters more comfortable with old-media habits? Did ‘tea party’ supporters simply send out far more tweets, drowning Reid in negative messages? Or, ahem, is Twitter just not an especially reliable indicator of what people are going to do once they enter the voting booth?

It would be interesting to see the Twitter traffic for a separate race, such as the Jerry Brown versus Meg Whitman gubernatorial contest in California, to see if the patterns of support were any different. But my own takeaway here is that Twitter may be a great way for a community of like-minded people to stay in touch -- like Justin Bieber fans -- but so far it doesn’t seem to be a particularly useful barometer for more complex, important decisions, such as political voting patterns.

It’s a conclusion that Malcolm Gladwell arrived at in a recent piece he did in the New Yorker, which demolished a lot of myths about Twitter’s supposed impact on social revolutions in Moldova and Iran. It turns out that what got Harry Reid re-elected was party discipline and get-out-the-vote strategy, two things that existed long, long before anyone was making inflated claims about the so-called Twitter Revolution.