Wednesday, October 20, 2004
2002 is not evidence of a weakening Incumbent Rule.
Never before have the voter preference polls been more scrutinized than this year, thanks to the convergence of three factors: the explosion of political analysis on the internet, the increased interest in this year's election, and of course, the very close nature of this year's race.
One of the "rules" of polling that has received a lot of attention lately is the "Incumbent Rule". Put simply, because undecided voters historically break in large numbers away from the incumbent, the polling number to look at is not the margin between the incumbent and the challenger, but rather the percentage of the vote the incumbent is receiving in the poll. There's a logical sense to this, in that voters are familiar with the incumbent (especially an incumbent President), and have strong opinions of him (or her). On the other hand, the challenger is usually more of an unknown, so there's a natural hesitancy to strongly declare support for him. This results in incumbents typically ending up within a range of +1 to -3 in their final vote percentage relative to their final polling number. To really fully grasp the dynamic here, I urge you to read this piece by "The Mystery Pollster", a must-read blog for anyone who follows the polls.
Put in the context of this race, polls that show Bush ahead of Kerry, say 47-44, are actually bad news for Bush since he probably wouldn't end up with much more than 48%. And currently (with the exception of the flawed Gallup poll and a couple of others with questionable "likely voter" screens) Bush is polling in the 46-48 range despite being level or slightly ahead of Kerry.
This well-established polling rule has caused quite a few Republicans heartburn lately, and they've yet to refute the theory convincingly. Their only hope seems to be Bush's favorite mantra, that "9/11 changed everything". And as evidence of this change, they point to the unexpected (by the polls) Republican success in the 2002 mid-term elections. Here is one example and here is another.
But there are several reasons why 2002 doesn't have predictive value for 2004:
1. 9/11 was barely a year before the 2002 elections. It's been over three years now.
2. The election took place during the run-up to removing Saddam. The Republicans ran on this issue, as well as 9/11, and gee, who wouldn't want Saddam out of power? Of course, now we know full well the cost of removing Saddam, in terms of American lives lost, billions spent, and America's diminished prestige in the world.
3. Related to #2, the Democratic turnout was depressed, as the base was demoralized by the Democrats in Congress who were so eager to get the issue of Iraq "off the table" in the final weeks of the election that they gave Bush a blank check for going to war without much dissent. The theory was that getting an Iraq resolution out of the way in early October that year would allow the Dems to run on economic issues. It did, but their base was demoralized by the Iraq cave-in, and stayed home. And it was this turn of events that led to the rise of Howard Dean in 2003, whose success led to the Democratic Party being more aggressive on the issues, and which led to the energization of the Democratic grass roots for 2004.
4. Mid-term elections always draw a disproportionately conservative electorate relative to presidential elections, which of course favors Republicans.
I don't doubt Republicans were more on their game in GOTV (Get out the vote) efforts in 2002 after they were whipsawed in 2000. But simply attributing the disparity in Republican vs Democratic turnout that year to their GOTV efforts and ignoring the factors set forth above is ludicrous.
The bottom line is that 2002 is no template for 2004. 2000 may not be either, but it's more of an apples-to-apples comparison than 2002 is. And until the Incumbent Rule is proven wrong, it remains a valid way of looking at the 2004 election.
|
One of the "rules" of polling that has received a lot of attention lately is the "Incumbent Rule". Put simply, because undecided voters historically break in large numbers away from the incumbent, the polling number to look at is not the margin between the incumbent and the challenger, but rather the percentage of the vote the incumbent is receiving in the poll. There's a logical sense to this, in that voters are familiar with the incumbent (especially an incumbent President), and have strong opinions of him (or her). On the other hand, the challenger is usually more of an unknown, so there's a natural hesitancy to strongly declare support for him. This results in incumbents typically ending up within a range of +1 to -3 in their final vote percentage relative to their final polling number. To really fully grasp the dynamic here, I urge you to read this piece by "The Mystery Pollster", a must-read blog for anyone who follows the polls.
Put in the context of this race, polls that show Bush ahead of Kerry, say 47-44, are actually bad news for Bush since he probably wouldn't end up with much more than 48%. And currently (with the exception of the flawed Gallup poll and a couple of others with questionable "likely voter" screens) Bush is polling in the 46-48 range despite being level or slightly ahead of Kerry.
This well-established polling rule has caused quite a few Republicans heartburn lately, and they've yet to refute the theory convincingly. Their only hope seems to be Bush's favorite mantra, that "9/11 changed everything". And as evidence of this change, they point to the unexpected (by the polls) Republican success in the 2002 mid-term elections. Here is one example and here is another.
But there are several reasons why 2002 doesn't have predictive value for 2004:
1. 9/11 was barely a year before the 2002 elections. It's been over three years now.
2. The election took place during the run-up to removing Saddam. The Republicans ran on this issue, as well as 9/11, and gee, who wouldn't want Saddam out of power? Of course, now we know full well the cost of removing Saddam, in terms of American lives lost, billions spent, and America's diminished prestige in the world.
3. Related to #2, the Democratic turnout was depressed, as the base was demoralized by the Democrats in Congress who were so eager to get the issue of Iraq "off the table" in the final weeks of the election that they gave Bush a blank check for going to war without much dissent. The theory was that getting an Iraq resolution out of the way in early October that year would allow the Dems to run on economic issues. It did, but their base was demoralized by the Iraq cave-in, and stayed home. And it was this turn of events that led to the rise of Howard Dean in 2003, whose success led to the Democratic Party being more aggressive on the issues, and which led to the energization of the Democratic grass roots for 2004.
4. Mid-term elections always draw a disproportionately conservative electorate relative to presidential elections, which of course favors Republicans.
I don't doubt Republicans were more on their game in GOTV (Get out the vote) efforts in 2002 after they were whipsawed in 2000. But simply attributing the disparity in Republican vs Democratic turnout that year to their GOTV efforts and ignoring the factors set forth above is ludicrous.
The bottom line is that 2002 is no template for 2004. 2000 may not be either, but it's more of an apples-to-apples comparison than 2002 is. And until the Incumbent Rule is proven wrong, it remains a valid way of looking at the 2004 election.
Comments:
Post a Comment
