The data science is in: Romney a lock on Super Tuesday

Updated: Oh, Rick Santorum, it was fun while it lasted. Tomorrow is Super Tuesday — when 10 states hold their presidential primaries and 419 delegates are up for grabs — and the election trackers at Yahoo and Fizziology have Mitt Romney as the favorite to win the day. The day probably won’t go as well for Romney’s erstwhile competitor Santorum, who is facing incredible backlash over recent comments on issues such as birth control, college and gay marriage.

The data scientists behind Yahoo’s ”The Signal” blog, who follow the various prediction markets very closely, are projecting Romney will 7 of 10 states tomorrow, with Santorum likely to win two (the conservative bastions of Oklahoma and Tennessee) and Newt Gingrich taking his home state of Georgia.

Most important in terms of winning the Republican nomination, however, is the state of Ohio, which Yahoo’s David Rothschild points out presents an interesting case study in whether polls are still the best method for gauging public opinion. As I explained recently, social media specialists with big data mastery feel they can do a better job of assessing public opinion than can traditional surveys, and Rothschild seems to think prediction markets’ big data prowess puts them ahead of pollsters in figuring out who will win elections. ”Romney has 81.6 percent odds in the crown jewel, Ohio, though polls—which lag compared to markets—have him virtually tied with Santorum,” he wrote.

And while the numbers don’t bode well for Santorum tomorrow, he’s also taking a hammering in public opinion. Here’s what Fizziology, which performs sentiment analysis across Twitter, Facebook and blogs, had to say about recent sentiment on Santorum:

Rick Santorum is currently seeing the highest percentage of negative conversation surrounding a GOP candidate that we’ve ever tracked. People are beginning to express their discontent more organically as seen with neutral conversation falling, and negative and positive conversation rising. Santorum is currently receiving much of his negative sentiment towards his morals and values, as they are seen by many as being outdated (Sample: “you should only vote for rick santorum… if you live in 1963”). A big negative conversation topic towards Santorum was latest statement that if we were to be President, he would nullify all current samesex marriages (Sample: “Santorum backs nullifying existing gay marriages – http://t.co/eBOrcMrH // yeah not in my lifetime Tricky Ricky. #Frothy #bigot #Santorum”). The social conversation suggests that Santorum, missing his window in Michigan, is seeing support erode.

The prediction markets actually agree, as Yahoo’s Rothschild noted. After winning Michigan and Arizona last week, Romney’s chances for taking the nomination have soared to over 80 percent.

One thing to watch for is whether Gingrich can pull a rabbit out of his hat and actually prove the oddsmakers wrong by winning more than one state. Fizziology noted that Gingrich has seen a sharp increase in positive sentiment lately along with a decrease in negative sentiment. Anyone who follows U.S. politics should recognize that’s a big deal, because while negative sentiment can come from anyone — often from the other party — positive sentiment suggests potential primary voters are actually in Gingrich’s camp.

As I said previously when covering another post on “The Signal” about potential vice-presidential picks, I’m willing to accept these numbers and predictions as factually accurate, but that accuracy can be short-lived. Presidential politics can change course with the wind, and one more big Romney gaffe or a new, big-name candidate riding into town on his (or her) white stallion could change the discussion completely.

Update: The sheer volume of tweets about the candidates seems to back up what the prediction markets and Fizziology are seeing. Here’s a graph from Twitter showing Romney leading in volume and Gingrich experiencing a sharp spike in the past couple days.