When it comes to predicting the outcome of the UK Election 2010, I want to know what Dan Berman thinks. Who is he?
A couple of years ago, Berman was my son Brendan's debate partner at Bates College. After debate tournaments, they often stopped by our home for pizza on the way back to Lewiston, Maine.
Last year, Berman leaped into the global headlines as a graduate student at the University of St. Andrews in Scotland. That's when he co-wrote an article that pointed out "data-driven evidence of widespread election irregularities" in Iran.
Yep, he's that Daniel Berman. He's what we call in New England "wicked smart."
As The Boston Globe reported in an article entitled, "Statistics ace raises doubt, fans anger on Iran's vote," Berman was burrowing into election-return data that had been posted online by the Iranian government two days after the June 12, 2009, election. The self-described elections junkie thought he saw suspicious patterns that backed up the opposition's claims of fraud.
"Before dawn the next morning, the self-described elections junkie zipped off an e-mail containing the fresh data and outlining his suspicions to Nate Silver, an elections guru known for his incisive work during the 2008 US presidential race. Silver, who knows little about Iran but plenty about statistics, thought Berman was onto something," wrote James F. Smith of the Globe.
Silver then posted an item on his website, fivethirtyeight.com, that highlighted Berman's data -- and the rest is history.
Where is Berman today? He's working with Silver and Renard Sexton on The FiveThirtyEight UK forecasting model. That's why I want to know what he thinks.
I've read Silver's post, "How Our UK Forecast Model Works." At one point, Silver says, "The math on this gets modestly complicated," but that's because the model tries to customize vote shifts in the UK's 650 individual constituencies. For those who speak English, the model tries to take into account what is normally called tactical voting.
It's a relatively elegant model that embodies what Silver believes to be sound logic about voter behavior, and there's something to be said for that. But, I know Berman and I've learned to trust his analysis. So, here's what he thinks:
If the election were held yesterday (and it won't be held until May 6), then Conservatives would have won 33.5 percent of the vote, Liberal Democrats would have won 28.7 percent, and Labour would have won 26.3 percent. Because of the British first-past-the-post electoral system, these percentages would not be reflected in the Parliament.
That's where The FiveThirtyEight UK forecasting model comes in. It currently projects that Conservatives would win 299 seats, short of a majority. Labour would win 199 seats, despite coming in third. And the LibDems would win 120. The remaining seats would be divided among three other regional and nationalist parties in Scotland, Wales and Northern Ireland.
Now, before Americans get on their high horse, remember that Al Gore won the popular vote in 2000, but George W. Bush got more electoral votes -- following a 5-4 vote by the Supreme Court that ended the Florida recount. So, the electoral systems on both sides of the pond should be reformed.
Plus, a lot can change in the next five days. But if and when it does, I plan to ignore the ubiquitous "swingometers" that purport to forecast the number of seats in the House of Commons that will change hands. They are based on a questionable assumption: a uniform national swing, i.e., if Labour finishes 7 points below their standing from the previous election in 2005, then their share of the vote will drop by 7 points in each individual constituency.
Instead, I'm going to head over to at fivethirtyeight.com to find out what does Silver, Sexton and Berman think? I'll bet you a pizza that they get it right.
Twitter Canada MD Kirstine Stewart to Keynote Toronto
ClickZ Live Toronto (May 14-16) is a new event addressing the rapidly changing landscape that digital marketers face. The agenda focuses on customer engagement and attaining maximum ROI through online marketing efforts across paid, owned & earned media. Register now and save!