The Simple Tech Tool That Can Predict the Winner of House Elections

BRISTOL, ENGLAND – JULY 23: In this photo illustration a man looks at Twitter on his iPhone and tweets from Clarence House on July 23, 2013 in Bristol, England. Credit: Getty Images

Forget polling data, just check Twitter.

A study by Indiana University may have found the best method to pick winners for the 2014 midterm elections and beyond.

The study found the candidate that gets the most tweets during a campaign – positive or negative – will win be election, based on 2010 and 2012 data.

So maybe there’s no such thing as a bad tweet.

The study, titled “More Tweets, More Votes: Social Media as a Quantitative Indicator of Political Behavior,” looked only at campaigns for U.S. House of Representatives.

“Think of this as a measurement of buzz,” Fabio Rojas, an associate professor of sociology at Indiana University Bloomington, said in university press release. “We call this the ‘all publicity is good publicity’ finding. Even if you don’t like somebody, you would only talk about them if they’re important.”

The lead author of the study is sociology doctoral student Joseph DiGrazia, who presented the findings Monday 108th annual meeting of the American Sociological Association, held Aug. 10 to 13 in New York.

“Our findings show there is massive, untapped reliable data out there that can give insights into public opinion,” DiGrazia said.

The school’s Department of Sociology and School of Informatics and Computing analyzed a sample of 537 million tweets for the study.

The study took into account variables such as incumbency, party make of the electorate, media coverage and socio-demographic makeup of the electorate, according to the IU press release. It drew from an extensive Twitter database compiled by the Center for Complex Networks and Systems Research at Indiana University’s School of Informatics and Computing.

Indiana’s Twitter database contains the largest sample of tweets in the world that is available to academic researchers, according to IU.