Nate Silver Thinks Gabriel Gomez Is Losing Bigtime

Forget the polls, says the political prognosticator.


Nate Silver feels your pain, Gabriel Gomez. Credit: nikolai36 on Flickr

By focusing only on how this Senate race compares to the 2010 special election that put Scott Brown in office, we risk ignoring all the useful examples set by hundreds of other Senate races that could help us predict the outcome when voters choose between Democrat Ed Markey and Republican Gabriel Gomez. Rumored witch Nate Silver, he of the Five Thirty Eight blog that accurately predicted many an election outcome last November, weighed in on the Gomez-Markey contest Monday night. His message to those who see Gabriel Gomez as the Second Coming (of Scott Brown): Not so fast.

“The 2010 Senate race in Massachusetts is one point of comparison,” Silver writes in his column, “but so are the 2012 elections, and so is the evidence from the dozens and dozens of Senate races across the country in recent years.” Silver notes that his own model for predicting U.S. Senate elections doesn’t give much weight to the polls early in the race. (After all, at this point, Scott Brown was behind Martha Coakley by like… 30 points.) So while Public Policy Polling put Gomez just four percentage points behind Markey, Silver says his model shows that Gomez will likely lose by 15.

To do so, he looks at three of the race’s “fundamentals.” He considers the candidates’ ideological positioning relative to the state’s voters, their fundraising, and their history of election to public office. On all three counts, Markey has the advantage (though to varying degrees, and Silver goes quite in depth on all three.)

Meanwhile, a bit closer to home on the Mass Politics Profs blog, UMass Boston political scientist Maurice Cunningham similarly warns that we should look outside the single 2010 race for evidence. He considers another Senate race model: one built in 2002 to predict the Maine Senate contest between Susan Collins and Chellie Pingree. Political scientists Kedron Bardwell and Michael Lewis-Beck’s model considered national economic and political conditions, the quality of the candidates, and the ratio of incumbent to challenger spending. The model came quite close, predicting the eventual result in Maine better than polls until fairly close to the election. Cunningham doesn’t run the numbers, but he suggests that all three of those factors favor Markey. Obama was popular enough to get Warren elected over Brown, Gomez is a fairly amateur candidate, and Markey isn’t likely to get radically outspent.

Looking at Scott Brown’s historic 2010 win gives Gomez (and the media hungering for something interesting to happen) a glimmer of hope that we might have an exciting upset on our hands. But putting that win in the context of the many other Senate races in our nation paints a more realistic picture: Gomez has a long road ahead if he’s going to give us an upset.