Measuring candidate quality

My wife bought me Malcolm Gladwell's new book What the Dog Saw, which is really just a compendium of Gladwell's more interesting stories in the New Yorker, to which I do not subscribe because I only end up reading the cartoons and then feeling guilty about the long, excellent pieces of journalism lying unread.

Anyway, the first essay in the book is an October 2000 piece about Ron Popeil. It's fascinating. But there's one anecdote in there that struck me as particularly interesting. As a young man, Ron was convinced he was the best pitchman out there. His relative Arnold Morris felt the same way about himself, as did a man named Frosty Wishon. So the three of them decided to have a shoot-out at the Eastern States Exposition in West Springfield, MA. They alternated selling the same knife set in different shifts during a ten-day stint at the show. Popeil just edged out Morris, and both of them buried Wishon.

This strikes me as a promising example for political scientists who study campaigns. We often come up with rough measures of candidate quality -- whether they've held office previously, whether they've run for office before, how moderate/extreme they are, etc. But what we really want to know is how good they are. We just knew that Barack Obama was a better candidate than John McCain last year, but how do we measure that? An actual election has too many moving parts to separate out candidate quality. We need a shoot-out. Maybe try the Glengarry Glen Ross approach: put a bunch of candidates in the same room, give them a randomized list of voters and a telephone, and see how much money each of them raises. I don't know. There must be some way to do this.