"if no one under the age of 30 had voted, Obama would have won every state he carried with the exception of two: Indiana and North Carolina," Chuck Todd and Sheldon Gawiser wrote in "How Barack Obama Won." Despite Mr. Obama's much-hyped "youth quake," the young social-media army often credited with carrying him into the White House, the 2008 election was won by galvanizing an older population of Democrats and independents, many of whom had never used social media.

Even more recently, the wild inconsistency between Republican primary wins and social-media prowess has made the relative unimportance of social media in political campaigns all the more obvious (i.e. internet heavyweight Ron Paul hasn't won a single primary).

Ultimately, Facebook, Twitter and YouTube fans consist of reliable supporters, the growing demographic of nonvoting 20-somethings or opposition spectators. Despite the increasing importance of social media in business, there is no solid evidence that it matters in politics.

If the Republican primary were a social-media popularity contest, Mr. Paul would be on Easy Street to the nomination. According to social-media analytics firm Crimson Hexagon, he has vastly outperformed all of his conservative counterparts as measured by total volume of Twitter mentions. Chatter about Mr. Paul accounted for 26% of the total discussion leading up to the New Hampshire primary, where the actual winner, Mitt Romney, had 22 %.

In raw social-media count, Mr. Paul has six times as many Facebook followers as breakout success Rick Santorum, who only a month ago had a meager Facebook presence of 42,147 followers. Mr. Romney (1.5 million Facebook fans), the once-undisputed front-runner, has had a stronger social-media presence than both Messrs. Paul (869,362) and Santorum (149,610) combined yet lost the last three primaries—in Colorado, Minnesota and Missouri.

In statistical parlance, the social-media universe is a biased population: Active online users hold views unrepresentative of the larger voting public. This is further corroborated by Google's slick new election toolkit, which measures total search volume and news mentions. News volume is a much more representative look at the wider population, as many more citizens get their news online than follow candidates on Facebook or Twitter. Using this tool, both Mr. Romney and Mr. Santorum had roughly equal news volume, with Mr. Paul trailing both.

The Facebook politics team likes to think their website can predict election results. In 2010, the Facebook Politics blog boasted that "an early sample of some of the hottest House and Senate races bodes well for the world's largest social-networking site." It noted that 81% of the Senate candidates with more Facebook fans had won their race. Yet 87% of incumbents in the House won re-election that cycle. Incumbents are simply more popular (and they have many opposition followers, who need to follow the updates of a sitting representative).

In a key contest of 2010, Tea Party favorite Sharron Angle vs. Senate Majority leader Harry Reid, Facebook was wildly off. In late October, Ms. Angle had a staggering seven times as many Facebook fans as Mr. Reid (104,000 to 14,000), yet he snagged a comfortable victory, 50% to 45%.

Thus, while it is tempting to believe that the size of a candidate's online fan-base is significantly helpful in getting elected, the numbers just don't add up.

On the internet, no one knows you didn't vote

So, why doesn't a social-media following translate into votes? The internet is like a giant trade show: a cacophony of aggressive marketers pushing strangers to read their pamphlets. People, however, are far more responsive to personal messages and face-to-face peer pressure.

In an experimental field study, Professors Alan Gerber and Donald Green randomly chose 30,000 potential voters to receive personal canvassing, direct mail or phone calls. Face-to-face canvassing "substantially" increased turnout, while the impersonal messaging of mail and calls had a much less successful effect. The researchers concluded, "A certain segment of the electorate tends not to vote unless encouraged to do so through face-to-face contact."

Social networks, as a sea of unblinking avatars, do not exert the same social pressure as direct human contact.

Eight years later, to further investigate their findings, Gerber and his colleagues conducted a peer-pressure-enhanced version of the direct-mail experiment by threatening to reveal to recipients' neighbors whether or not they had voted. Direct mailings with social pressure had a dramatic impact (an 8.1% increase) "The difference between our intervention and mail used in previous experiments is that ours harnesses one of the most-formidable forces in social psychology, pressure to conform to social norms," the study said.

On Facebook, the closest thing to social pressure is a vote counter appended to the top of the newsfeed on Election Day. While this most likely has some positive effect, it certainly doesn't pressure the legions of uninterested voters who can float by anonymously without ever feeling the scrutiny for shirking their civic duties.

Moreover, internet-specific evidence on the importance of personalized messages was observed by New York University Professor Sinan Aral, in a rare experimental study of Facebook-app adoption. Aral found that personalized messages to friends asking them to try out an app increased user adoption 98% vs. a generic broadcast messaging through the newsfeed, and increased engagement by 7%. Because Mr. Obama cannot personally message his 25 million fans, his influence is limited to the faceless, soapbox shouting of a newsfeed post.

Campaigners have yet to crack the code of translating digital enthusiasm to offline action. For now, the influence of Facebook and Twitter are confined to where they live: online.

Voting is so Generation X

Voting is not a social activity; privately casting a numerically insignificant ballot in a sealed booth is not appealing to a generation of hyperconnected content creators. "If you ask people how could you affect politics and have an impact, voting is only one of a whole variety of things, and it's fairly limited," says University of California at Irvine Professor Russell Dalton, who found an uptick in a young demographic of nonvoting citizens who were more interested in boycotts than in voting or jury duty. "Engaged" citizens, as he calls them, desire to be part of the political process and eschew traditional notions of civic duty.

In other words, many of the most active young social-media fans are the least likely to vote. "The Engaged Citizen is more likely or as likely to do all kinds of participation, except voting," said Mr. Dalton (though he tries to tell his students that not voting is a mistake). Instead, his Engaged Citizens are pitching tents in Occupy protests, sharing "I've Got a Crush on Obama" and volunteering in their neighborhoods.

While this may be good for creating buzz around a new candidate, which Mr. Obama was at the time, it doesn't translate into the ultimate measure of a candidate's popularity: votes.

The internet is a v-ery important tool. Mr. Obama won not because he had more Facebook fans but because he ceded unprecedented control to his online followers. His team opened up the voter file so that volunteers could make calls from home, randomly selected low-dollar donors for a small dinner and created tools for offline organizing. Despite the digital trailblazing in 2008, little innovation has taken place since, and most candidates, especially the current crop of Republicans, merely recycle messages from stump speeches into updates of less than 140 characters. The potential of the internet lies in not who follows a candidate but in how those followers are empowered.