Daniel Sinker: Why 2008 is Not 2004 Redux: Remember the Bush/Kerry Matchup? Kerry was Polling Strongly, Right? Think Again.
I've heard whisperings the last few days that, despite the polls, people shouldn't be too confident in an Obama victory next week. The reason? Kerry was polling well leading into election 2004, but Bush won the election. While the latter is true--obviously, we've all been suffering under it--the former is a memory that is been clouded by wishful thinking. We wanted the polls to be better, so we remember them as better.
Looking at polling results from the final days of the 2004 campaign paints a very different picture: With only three exceptions, Bush held the lead in 22 out of 25 polls going into election day. Those leads, which averaged 2.23% for Bush, very closely mirrored the final election outcome: a 2.45% margin of victory.
It was going to be a close race--but it was clear who was going to win. The thing about close races, though, is that you can look at a close race and hope that your candidate can leap that final distance (the few polls that had an Kerry victory showed him at 1.33% up). But, with the luxury of time, it's easy to see now that there wasn't really a contest: the numbers were overwhelmingly in Bush's favor, even if the margins were small. Realizing that back then would have saved me a lot of heartache (probably many of you too).
2008 is a different story. In the most recent data available (10/26) Obama leads in every single national poll. His average? Of the seven polls released, he averages a margin of victory of 7%, more than three times what Bush was averaging in 2004. That's an entirely different ballgame.
As a result, this year you can see the "wishing a poll was better" game flipped on its head. The Drudge Report today (in addition to butchering a seven-year-old quote from Obama) cites a single poll, John Zogby's, as evidence that the race is getting closer. Even that poll has Obama up by 4.7%. By cherry-picking the outlier (an outlier that still shows the other guy ahead by a healthy margin), Drudge hopes to show that the race is "tightening." It's not.
This election is not going to swing 7% away from Obama in the next seven days any more than 2004 was going to swing 2% towards Kerry in the last week. Even someone as bad at math as myself can see that 2008's numbers are beginning to look like they add up to victory.
An update on a previous post: In my last blog entry, I compared Google Trends results for Obama and McCain. They very closely mirror "real" polling results, showing the same sorts of fluctuations, preferences, and trends--yet they're based on nothing more than Google searches for a candidates name. A few people in the comments said that it wasn't representative, that it just demonstrated that more people search for Obama and that, if you were to include Sarah Palin in the search criteria, "I bet she out-paces Obama on that graph." You'd guess wrong.
Read entire article at Huffington Post (Blog) (Click here for hot links in this blog.)
Looking at polling results from the final days of the 2004 campaign paints a very different picture: With only three exceptions, Bush held the lead in 22 out of 25 polls going into election day. Those leads, which averaged 2.23% for Bush, very closely mirrored the final election outcome: a 2.45% margin of victory.
It was going to be a close race--but it was clear who was going to win. The thing about close races, though, is that you can look at a close race and hope that your candidate can leap that final distance (the few polls that had an Kerry victory showed him at 1.33% up). But, with the luxury of time, it's easy to see now that there wasn't really a contest: the numbers were overwhelmingly in Bush's favor, even if the margins were small. Realizing that back then would have saved me a lot of heartache (probably many of you too).
2008 is a different story. In the most recent data available (10/26) Obama leads in every single national poll. His average? Of the seven polls released, he averages a margin of victory of 7%, more than three times what Bush was averaging in 2004. That's an entirely different ballgame.
As a result, this year you can see the "wishing a poll was better" game flipped on its head. The Drudge Report today (in addition to butchering a seven-year-old quote from Obama) cites a single poll, John Zogby's, as evidence that the race is getting closer. Even that poll has Obama up by 4.7%. By cherry-picking the outlier (an outlier that still shows the other guy ahead by a healthy margin), Drudge hopes to show that the race is "tightening." It's not.
This election is not going to swing 7% away from Obama in the next seven days any more than 2004 was going to swing 2% towards Kerry in the last week. Even someone as bad at math as myself can see that 2008's numbers are beginning to look like they add up to victory.
An update on a previous post: In my last blog entry, I compared Google Trends results for Obama and McCain. They very closely mirror "real" polling results, showing the same sorts of fluctuations, preferences, and trends--yet they're based on nothing more than Google searches for a candidates name. A few people in the comments said that it wasn't representative, that it just demonstrated that more people search for Obama and that, if you were to include Sarah Palin in the search criteria, "I bet she out-paces Obama on that graph." You'd guess wrong.