According to researchers at Indiana University, the link between tweets and votes in political elections could be stronger than once thought. In a study analyzing 537 million tweets from August 2010 —the largest sample of tweets ever accessible to academic researchers — the percentage of votes for Democratic and Republican candidates in the House of Representatives race correlated with the percentage of tweets that mentioned them.

The interesting part is that it didn't seem to matter if the surveyed tweets were positive or negative. "We call this the 'all publicity is good publicity' finding," said Fabio Rojas, an associate professor at Indiana. "Even if you don't like somebody, you would only talk about them if they're important." The study adds to the mounting evidence that social networks are not ephemeral, spam-infested sources of information as once thought. In fact, Twitter and Facebook might turn out to be the most valid means yet of measuring the American electorate, since the sentiments inside the actual posts don't make a difference.

The findings spell bad news for Gallup and other polling companies, which have traditionally been the go-to sources for information on election day. Perhaps someday soon, determining the popularity of political candidates might only mean a trip to Twitter's search box.