Remember when it appeared that Sen. John Kerry might make George W. Bush a one-term president just like his father? Can't remember that? Let's do a memory refresh from October 16, 2004. Leftblogger Jerome Armstrong of mydd.com looked deep into a Washington Post poll and found encouragement in the internals. The WaPo survey reported that Kerry "continued to claim a large lead in key battleground states. In these 13 states, Kerry held a 53 percent to 43 percent advantage among likely voters." Armstrong gushed:
That's a 10% lead by Kerry in the battleground states-- Wow!
The WaPost poll showing that Kerry leads in the battleground states is confirmed by Rasmussen, which shows Bush leading Kerry nationally by 2 percent, 48-46, yet Kerry leading Bush in the battleground states by a 48-46 percent margin-- a 4 percent swing.
Looking at the battleground states then, there is a 4-9 percent swing from the national polling. Kerry's lead over Bush is firming.
High expectations for Kerry had started building in the spring of 2004. In a piece which appeared in the May edition of the liberal Washington Monthly, Chuck Todd, then the editor-in-chief of National Journal's Hotline and now ABC's political director, predicted a Kerry landslide:
If you look at key indicators beyond the neck-and-neck support for the two candidates in the polls--such as high turnout in the early Democratic primaries and the likelihood of a high turnout in November--it seems improbable that Bush will win big. More likely, it's going to be Kerry in a rout.
On October 14, CNN was out with the results of a poll which indicated that Kerry had won the previous night's presidential debate hands down:
Sen. John Kerry appeared to gain more momentum heading toward November 2, easily beating President Bush in the third and final debate, a poll taken late Wednesday night suggests.
A CNN/USA Today/Gallup snap poll taken immediately after the presidential debate found that respondents gave a significant edge to Kerry over Bush, 52 percent to 39 percent.
Indeed, throughout 2004 the liberal media's spinning of the poll numbers were ramping up expectations that Kerry would become the 44th president. No one seemed to have learned the lessons from the previous presidential election. In mid-September of 2000, Slate published a William Saletan piece which explained Why Bush is Toast:
In the latest polls, Gore leads by an average of five points. It's fashionable at this stage to caution that "anything can happen," that Bush is "retooling," and that the numbers can turn in Bush's favor just as easily as they turned against him. But they can't. The numbers are moving toward Gore because fundamental dynamics tilt the election in his favor. The only question has been how far those dynamics would carry him. Now that he has passed Bush, the race is over.
Yes, in principle, Bush could win. The stock market could crash. Gore could be caught shagging an intern. Bush could electrify the country with the greatest performance in the history of presidential debates. But barring such a grossly unlikely event, there is no reason to think Bush will recover. Ultimately, reasons drive elections. For months, pundits yapped about Bush's lead in the polls without scrutinizing the basis of that lead. Now they're doing the same to Gore. But look closely at the trends beneath the horse-race numbers, and you'll realize why it's practically impossible to turn those numbers around. Gore doesn't just have the lead. On each underlying factor, he has the upside as well.
All of this was long since forgotten in October '04. On the 21st, the nutrooters hanging out at Democratic Underground could feel the tingle going up their legs. They found hope for change in an article by focus group guru Frank Luntz :
"Its time for Bush to get worried" "Frank Luntz, the top Republican pollster wrote in the Financial Times: "Step by step, debate-by-debate, John Kerry has addressed and removed many remaining doubts among uncommitted voters. My own polling research after each debate suggests a rather bleak outlook for the Bush candidacy: many who still claim to be 'undecided' are in fact leaning to Mr. Kerry and are about ready to commit." In a world where the spinmeisters constantly claim that their horse can not only walk on water, but also trot and gallop on it, having a top GOP strategist with access to real data say his horse is sinking fast is ominous for the Bush campaign."
On November 1, the eve of the election, a Fox News/Opinion Dynamics poll showed Kerry up by three, 48% to 45%, among registered voters.
On election night, exit polling indicated a Kerry win. The National Election Pool, made up of the television networks and the Associated Press, were reporting that Kerry had won the popular vote by a 50% to 48% margin and had taken the battleground states of Iowa, Nevada, New Mexico and Ohio.
Zogby was saying that Nevada and Colorado were too close to call, but nevertheless predicted an electoral vote runaway for Kerry of 311 to 213. Those who took the exit polls at face value were in for a rude awakening. As the NY Daily News reported:
Republicans panicked, Democrats gloated and the stock market tumbled after numbers meant solely for the eyes of news editors were leaked to Internet blogs and other Web sites.
The official results of the 2004 presidential election showed that the exit polls had the popular vote backwards. Bush won it by a margin of 50.73% to 48.27%. In the electoral college, Bush won by 35 votes, 286 to 251. What had gone wrong?
Many people who should have known better had forgotten the small print at the bottom of every poll's results: margin of error. Most of the polls had a three-point margin of error, so what appeared to be a 50 to 48 victory for Kerry was actually a 51 to 48 win for the president.
They had also forgotten Santayana's famous admonition from his classic work, The Life of Reason:
Those who cannot remember the past are condemned to repeat it.
The same kind of poll results which were taken at face value in 2000 were taken in the same manner in 2004. Now it's 2008 and just two weeks before election day, and I see few signs that the lessons of the previous two presidential elections are being taken to heart. To repeat the old Yogi Berra quip, "It's deja vu all over again."
The media is busy pursuing a hard sell for Barack Obama, and he's leading in the polls. Just how much that lead actually is depends on which poll you read. Two polls, CBS/NY Times and Diageo/Hotline have him up by double digits, 14 and 10 points respectively. But three polls - Battleground, Zogby and Rassmusen - each show Obama's lead to be 4 points. Gallup's poll of likely voters shows only a two-point lead for the Democrat.
Now bear in mind that these polls have margins of error ranging from plus or minus two points to as much as... well, we don't know. Some polling outfits don't provide that information. A three-point MOE tends to be common among the more accurate polls, so I wouldn't be surprised if some of the lesser reliable surveys have errors of 5 or 6 points.
Which polling organizations are the most reliable? That would be Battleground, Zogby, Rasmussen and Gallup, not coincidentally the four polls which indicate the race is tighter.
Another grain of salt these polls must be taken with is how they weight their samples according to the party identification of the respondents. The CBS/NY Times poll, for example, sampled 371 Democrats, 206 Republicans and 306 independents. Then they weighted the samples to show 338 Democrats, 263 Republicans and 301 indies. If you're thinking, "No wonder Obama's lead is highest in this poll than in any of the others," don't feel like the Lone Ranger.
According to Zogby, his competitors oversample Democrats because he has found that Dems are more likely to respond to polls than Republicans. He doesn't believe that the practice of oversampling Democrats is overt bias but rather just a sampling issue. I look at that CBS/NYT poll, factor in the NYT's very biased election reporting, and I tend to disagree with Mr. Zogby on its supposed neutrality. The CBS/NYT's weighted sample of 41% Democrats, 29% Republicans and 33% independents is just not credible. It's not even on the same planet with Zogby's more reasonable weighting of 39% Democrat, 35% Republican and 26% Independent.
As a final caveat, let's not forget Zogby's debunking of the two most common myths about political polling:
Myth 1 -- Polls predict the winners and losers. Actually, a poll is only a snapshot of a moment in time. It can point to trends, but things can change on election day, when a lot of undecided voters make up their minds. We do try to ask "projective questions" – i.e. to see how people will react to situations and messages, but a poll can only measure a fixed moment in time.
Myth 2 -- Polls determine the outcome of an election (therefore, why bother to vote?). Polls generally only confirm what professional observers (and many voters themselves) already know – whether a race is close or not. Early polls can have an effect on a candidate’s ability to raise funds, but they do not shape how an election will turn.
The more reliable polls are telling us that the race is close. Four points is just one point outside the margin of error, so what we have here is a statistical dead heat.
Don't believe any talk about an Obama landslide. Pay no attention to the stories about how the Obama campaign is so confidant of victory that it has a transition team already measuring the oval office for drapes. Don't give any credibility to countless media stories about how Obama leads in key battleground states. Don't heed the pundits who are trying to convince you that it's over except for Obama's inauguration ceremony. Yogi had another quip which we would all do well to take to heart: "It ain't over till it's over."
Remember that the media is in the tank for Obama. Remember that their left-leaning bias is as plain as the nose on 83% of the faces of American voters. That bias is all about discouraging those favorably inclined to vote for John McCain and make them stay home on November 4, believing that their vote won't make any difference.
Remember 2000. Remember 2004. Remember Santayana. Remember Berra.
And don't forget to vote!