Eight years ago, right before Republicans gathered in New York City from Aug. 30 to Sept. 2 to renominate President George W. Bush for a second term, a newly released NBC News/Wall Street Journal poll showed Bush vulnerable and in a dogfight against his challenger, a Massachusetts Democrat.
In many respects, that survey bears a striking similarity to the most recent NBC News/Wall Street Journal survey of 1,000 registered voters, conducted Aug. 16-20 of this year and released less than a week before the GOP convention. This time, it is President Barack Obama who is vulnerable and in a dogfight with his opponent, a Massachusetts Republican.
Both the August 2004 NBC News/Wall Street Journal survey of 806 registered voters and the most recent one were conducted jointly by Democrat Peter Hart and Republican Bill McInturff, two highly regarded professionals.
While some of the survey results in 2004 and 2012 are remarkably similar, there are at least a couple of dramatic differences. And they demonstrate why the outcome this November is so unpredictable.
The 2004 poll found Bush's job rating at 47 percent approve/48 percent disapprove, eerily similar to Obama's 48 percent approve/49 percent disapprove in this month's survey.
The public's assessment of Bush's handling of the economy back then, 43 percent approve/52 percent disapprove, was almost identical to their assessment of Obama's now, 44 percent approve/54 percent disapprove.
Those polled also had very similar feelings about Bush in 2004 and Obama now.
Bush was rated positively by 49 percent of those responding and negatively by 43 percent in the August 2004 survey. Obama's personal ratings were 48 percent positive and 43 percent negative in the August 2012 poll.
Perhaps not surprisingly given the nearly identical job and personal ratings, the August 2004 and August 2012 presidential ballot tests in the two surveys were, from a statistical point of view, identical, with Bush leading Sen. John Kerry 47 percent to 45 percent and Obama leading Mitt Romney 48 percent to 44 percent.
Of course, Bush went on to beat Kerry by just 2.1 points in November 2004 (50.4 percent to 48.3 percent) in a very close contest.
So where are the obvious differences between the 2004 and 2012 surveys?
First, Bush's foreign policy rating then was significantly worse than Obama's is now.
This month's Obama foreign policy job rating stood at 54 percent approve/40 percent disapprove, while Bush's was 44 percent approve/52 percent disapprove in August 2004.
The public still viewed Bush's policies on terrorism favorably, but they had become more skeptical about his policies regarding the war in Iraq. As a result, his foreign policy job approval had plummeted from a high of 81 percent after the attacks of Sept. 11, 2001, to just 44 percent before the 2004 GOP convention.
But because voters did not have great confidence in Kerry as commander in chief (35 percent said they had a great deal/quite a bit of confidence in him, while 54 percent said just some/very little confidence), the issue did not seem to hurt Bush.
Second, Kerry had somewhat higher personal ratings (44 percent positive/40 percent negative) before the 2004 GOP convention than Romney has now (38 percent positive/44 percent negative). And while the Democratic Party's rating isn't much different now from then, the Republican Party's is much worse: 44 percent positive/38 percent negative in 2004 compared with 36 percent positive/45 percent negative now.
All of those figures suggest Romney is in a more difficult position than Kerry was at roughly the same point eight years ago, though it is important to note that the 2004 Democratic convention, which took place in Boston from July 26-29, had already occurred. (The summer Olympics that year were held Aug. 13-29 in Athens, so Democrats needed to hold their convention well before it.)
Unlike Romney, Kerry had already introduced himself to the electorate, though according to CNN/USA Today/Gallup polling immediately before and after the '04 Democratic convention, he received no public opinion bounce from the event.
The August 2012 NBC News/Wall Street Journal survey, of course, was conducted before either convention occurred, though the amount of early television advertising this year might dilute the effect of both conventions.
The other difference between the 2004 survey and the 2012 poll may be even more important in understanding the September-October dynamic this year.
While the general mood of voters in both August 2004 and August 2012 can only be described as pessimistic, the current mood is clearly worse.
In the August 2004 survey, 36 percent of respondents said the country was headed in the right direction, while 50 percent said it was off on the wrong track. In this month's survey, only 32 percent said the country was heading in the right direction, while 61 percent said it was off on the wrong track.
The wrong track number - 11 points higher now than it was in the August 2004 survey - is an extremely dangerous number for Obama, and it suggests that, to win a second term, he must make the November election into a referendum on Romney's weaknesses and liabilities rather than on his own performance, which is exactly what the president has been doing.
On the other hand, if Romney can make himself more personally appealing or even acceptable to swing voters, he will have a much better chance of avoiding the same fate that befell Kerry in November 2004.
Of course, the past certainly doesn't predict the future, and both the presidential debates and the unemployment and new job numbers to be released in early September, October and November could move the presidential contest noticeably before Nov. 6.
Stuart Rothenberg is editor of the Rothenberg Political Report.