Sunday's New York Times had a fascinating article on the huge difficulties administration officials are having selling the No Child Left Behind (NCLB) Act in rock-ribbed Republican Utah. To say they're getting a skeptical reception would be to considerably understate the case.

In light of this reception, which has been duplicated in numerous states around the country, Democrats may be tempted to run hard against NCLB and say, implicitly or explicitly, that it needs to be gotten rid of. A review of public opinion data on NCLB suggests they should resist that temptation.

Public opinion on public education has consistently shown that the public has a two point program for education reform: more accountability and more resources. The linkage between the two is neatly captured by a result from a 2001 Education Testing Service (ETS) survey. Respondents were asked what the best way to improve the quality of the public education was: more funding; accountability or both accountability and funding. The dominant response was both accountability and funding (48 percent), rather than simply more funding (25 percent) or accountability (23 percent).

In many ways, this was the premise of the bipartisan NCLB. Backers of the bill, who not only included the Bush administrion and its Congressional allies, but also Democratic liberals like Ted Kennedy and George Miller, maintained that the strict standards set by the law would be accompanied by increased funding that would help schools, particularly those with many at-risk students, meet those goals.

The public for its part, was also favorably disposed toward the bill in the period before its passage. The 2001 ETS poll, for example, reported support levels ranging from 58 percent (funds tied to performance) to 93 percent (funding for K-3 reading) for the very general provisions of the bill, with the emphasis on annualized testing scoring about in the middle of that range (76-78 percent). The poll also found, however, that a large chunk of the public–about two-fifths–did not have any real knowledge about the bill prior to being asked about it on the survey. A 2002 Gallup/Phi Delta Kappa (PDK) poll, conducted roughly five months after passage of the bill, found similar support levels for NCLB’s general provisions, combined with the same lack of prior knowledge of NCLB among two-fifths on the public.

Since that period leading up to and right after NCLB’s passage, however, the bipartisan consensus around education policy has broken down. Critics accuse the NCLB of being inflexible and underfunded and constituting, in many ways, an “unfunded mandate” on the states. Recently, we have had the spectacle of Republican-dominated state legislatures in Ohio, Virginia and Utah severely criticizing the law and threatening to opt out of federal funding to avoid being subject to it.

Unfortunately, it is difficult to gauge how much public support for NCLB has actually shifted due to the lack of knowledge problem and due to the lack of consistent time series data (i.e., the lack of questions about NCLB that have been asked in the same way from passage of the bill to the current time period). However, there are certainly indicators that the public has serious problems with many specific aspects of NCLB implementation.

In terms of testing, the public, in a 2003 Gallup/PDK poll, overwhelmingly (84 percent to 14 percent) said that the best way to judge a school’s performance is to see whether students show reasonable improvement from where they started, rather than whether they meet a fixed standard, as specified in the NCLB Act. And, in the same poll, by more than two to one (66 percent to 32 percent), the public thought that a single test, as in the NCLB Act, cannot provide a fair picture of whether or not a public school needs improvement. The public also strongly endorsed the idea (72 percent to 26 percent) that a single test cannot judge a student’s proficiency in English and math accurately. And, in a January, 2004 Greenberg Quinlan Rosner/Tarrance Group/National Education Association (GQR/TG) poll, 67 percent agreed that the law was unfair because it labels schools as “failing”, if one group of students doesn’t do well on a test, even if the vast majority do; 71 percent agreed that some kids should be given more time to pass tests for their grade due to differing ability levels; and 89 percent supported a proposal to allow schools to evaluate students’ progress on a number of criteria in addition to standardized tests, including classroom performance and graduation rates.

In terms of sanctions and funding, in a January, 2004 CBS News/New York Times poll, the public, by 58 points (77 percent to 19 percent), opposed using the results of tests to withhold federal funds from those schools where students perform poorly. And over four-fifths (81 percent) in the GQR/TG poll wanted schools to be given more time before penalties are assessed if funding promised by the NCLB has not been given to these schools. In the same poll, by 60 percent to 38 percent, voters supported increased funding, rather than cuts, for schools that are not able to meet federal testing standards.

On the other hand, there is little evidence that the public rejects the law itself or wants to do away with it. General descriptions of the law, particularly its goals and broad emphasis on accountability, continue to elicit strong public support. But it seems fair to say that the public would be supportive of changes that would make NCLB more flexible and better-funded. That's the sweet spot for Democrats in criticizing NCLB--not opposing the act, but seeking to reform it in line with the public's twin commitments to effective accountability and more funding.

That appears to be the attitude of Bush's political team, given that the president has now announced his support for a constitutional amendment to ban gay marriage. This move has been widely viewed as an effort to satisfy the GOP's conservative base and that judgement seems about right.

Why would the Bushies do this? Because they're very worried about fragmentation of the GOP base, as questions pile up about Bush's free-spending ways and deviations from the conservative agenda (e.g., the immigration bill). They can ill afford such fragmentation in a situation where the Democratic base is consolidating, swing voters are moving away from the GOP and Bush is running behind in trial heats against the probable Democratic challenger. Sure, the GOP spinmeisters talk a good game about how none of the polls really matter at this point; and of course the Democrats are running ahead what with all their free publicity at this point in the campaign cycle; and so on. But they know better.

In fact, it's quite unusual for an incumbent president to be running behind the challenger at this point in the cycle. According to a recent Gallup analysis, every incumbent president back to Harry Truman was leading their eventual opponent (and all other possible opponents) at this point in an election year with the lone exception of Gerald Ford in 1976. And of course, he lost.

So they're right to be worried. And shoring up your base by endorsing a constitutional amendment to ban gay marriage is a natural reaction. But is it smart politics?

Probably not. In fact, it's exactly the sort of thing the Bush strategists wanted to avoid in an election year--having to cater to the conservative base and risk alienating swing voters in the process.

And alienate they will. The latest Newsweek poll shows that, while the public is split about evenly, a strong majority of the most volatile age group, those 18-29, supports either full marriage rights for gays or civil unions (58 percent to 36 percent opposition). Similarly, while a healthy majority (56 percent) of the public supports either full marriage rights for gays (23 percent) or leaving decisions on the issue to the states (33 percent) , rather than a constitutional amendment (39 percent), views among young people are positively lop-sided: 66 percent for full marriage rights or state decisions, compared to 30 percent for a constitutional amendment.

And then there are independent voters, the quintessential swing group. In just-released ABC News/Washington Post data, a question on a constitutional amendment vs. state decisions elicits an even split among the public as a whole, but a 58 percent to 35 percent majority against an amendment among independents. Similarly, an Annenberg Public Policy Center poll question about support or opposition to a constitutional amendment against gay marriage shows an overall 48 percent to 41 percent majority against an amendment, but a larger 52 percent to 37 percent majority against among independents.

So, this attempt to satisfy their base is likely to come at a price among other voters they need to win the November election. Of course, they will immediately seek to rebuild Bush's image as a tolerant, compassionate conservative, as they have done before, through speeches, gestures and other political tools they have at their disposal. But every move like this makes that rebuilding task a little bit harder. And the November election draws ever closer.

Now that Dean has left the race and the Democratic nominee will either be Kerry or Edwards (probably Kerry), the temptation will be great to just forget about Dean's movement. Democrats who, six weeks ago, took Dean very seriously indeed now appear prepared to deny under oath that they ever did any such thing.

That would be a mistake. As a number of observers have pointed out, the fact that Dean couldn't secure the nomination doesn't mean the process of transformation he started within the Democratic party isn't needed. For a useful sampling of opinion along these lines, see the mini-symposium on Dean on the American Prospect website, with contributions by Michael Tomasky, Simon Rosenberg, Garance Franke-Ruta and Nicholas Confessore and the article in Salon.com by Joan Walsh.

A lot of these authors make the same couple of points in different ways. Here's the short course.

1. Dean did not make a real ideological or policy contribution to the party (though his willingness to stand up to Bush played a critical role in reviving the Democrats' fighting spirit). Confessore well-summarizes Dean's lack of ideological distinctiveness:

He put forward the least radical health-care proposal of any of the five major candidates running before New Hampshire. His ideas to expand federal aid for child care and college tuition were not much more than Clinton retreads. His best-known proposal -- repealing even the middle-class tax cuts passed by Congress in 2001 and 2002 -- was notably only for its stupidity, and he likely would have dropped it had he stayed in the race.

2. Dean's real contribution lay in the process by which his campaign operated, especially via the internet: recruiting enthusiastic volunteers; raising huge sums of money from small donors; and generating a "movement" level of energy at the grassroots of Democratic party. By doing so, he showed the party what it was missing and how hollow Democratic party organization had become.

Therefore, if the party is to maximize its chance of winning in 2004 and, especially, build an effective majority party for the future, it will have to internalize and further develop the organizing methods of the Dean campaign. In a sense, Deanism is now a "third force" in the Democratic party, not clearly tied to either the traditional liberals or the orthodox New Democrats of the DLC. Harnessing that third force is key to the Democrats' future; neither liberals nor New Democrats should delude themselves that things can now go back to the way they were. And, for that matter, neither should John Kerry--he will need the third force's help and plenty of it to beat George Bush.

In DR's view, Democrats need to get over the obvious fact that Nader's decision to run for president is morally irresponsible, egotistical, betrays the progressive principles he's based his life on, etc. and look the truth squarely in the eye. This guy is irrelevant. He might as well team up with Ramsey Clark and run on the Workers World Party ticket for all the votes he's going to get. He's not going to get 2.7 percent of the vote, as he did in 2000, or even .7 percent of the vote, as he did in 1996. I think we're talking more like .27 percent of the vote, which will put him firmly in the list of splinter party also-rans whose votes are counted but have no real effect on the election. He will join the distinguished company of the American Independent Party, the Citizens Party, the National Alliance Party and, of course, the Libertarian Party, which has averaged a solid .38 percent of the vote in the last three elections. Indeed, as Mark Schmitt points out in an excellent post in The Decembrist, the Libertarian candidate in this coming election will have a more natural constituency on the right, given the distress with Bush among fiscally conservative and libertarian conservatives, than Nader will have on the left and may well outpoll him.

Why is this? Because the fundamental rationale for his candidacy--no difference between the two major parties, so it doesn't matter which one occupies the White House--has no plausibility on the left to anyone outside of, well, the Workers World Party. That's why the left activists and intellectuals who supported him last time are deserting him en masse, from the Green Party to The Nation. Nader will have no organization, no big-name supporters, no money and precious few volunteers. He's likely to fail to get on the ballot in many states and, even where he does, he's unlikely to get more than a handful of votes.

So: be not afraid. Nader will fail and fail big-time. And tell Ralph that Workers World is waiting for his phone call.

The latest Democracy Corps report is titled "New World: Bush in Peril". While Democracy Corps can sometimes be a bit over-optimistic in their poll interpretations, in this case their optimism seems justified.

In their latest survey, they find some significant evidence that the tide is turning. For example, they find that the Democrats now have a 5 point lead in party ID, a lead that first emerged in their polling five months ago and now seems solid. DR has been arguing for awhile that this was occurring and Democracy Corps is kind enough to credit him with correctly predicting the emergence of this trend.

The poll also finds that likely voters give Democrats a 7 point lead in a generic Congressional ballot and give Kerry a 4 point lead over Bush. In addition, by 13 points, voters say the country is off on the wrong track and by 8 points they say they want the country to go in a "significantly different direction" than the direction Bush is headed in. They also say they want to go in a significantly different direction in a wide variety of specific areas: by 32 points on the federal budget; by 25 points on health care; by 22 points on prescription drugs for seniors (more evidence that the GOP has lost the debate on the prescription drugs bill); by 22 points on jobs in America; by 15 points on the economy; by 11 points on taxes (another highly significant finding); by 10 points on creating more employment opportunities; by 9 points on income and wages; and by 9 points on middle class living standards. (For more detail on these sentiments, see Democracy Corps' recently-released report on focus groups with swing voters.)

Democracy Corps also finds that independents and voters in swing areas are moving rapidly away from Bush. In their Bush-Kerry trial heat, independents favor Kerry by 11 points, voters in swing states favor him by 6 points and voters in swing congressional districts back Kerry by 4 points. And, on the question about whether the country should go in a significantly different direction, independents favor a different direction by an impressive 23 points (60 percent to 37 percent), voters in swing districts favor a new direction by 11 points and voters in swing states want the same by 10 points.

The latest Newsweek poll has more on the increasingly chilly climate for Bush. The poll finds his approval rating at 48 percent, with 52 percent saying they would not like to see him re-elected, compared to just 43 percent who say they would (an all-time low in this poll for Bush). He also receives poor ratings on tax policy (45 percent approval/47 percent disapproval); the situation in Iraq (45 percent/44 percent); the economy (41 percent/52 percent); health care (37 percent/50 percent); and (a new and interesting question) job creation and foreign competition (32 percent/55 percent).

So, if it is morning in America again, it definitely isn't the kind that Ronald Reagan had (or proclaimed) in 1984. It's a new morning, whose results, both for the GOP and for the country as a whole (thank goodness), seem likely to be quite different.

Several recent polls provide a wealth of information about how the political terrain is shifting against Bush and the GOP.

Start with the latest Pew Research Center poll. According to this poll, Bush's favorability rating has dropped from 72 percent last April to 53 percent today. And, when respondents were asked to supply a one word description of Bush, they were evenly split (36-36) between those who supplied negative or positive descriptions. That compares to almost a 2:1 split (52-27) in favor of positive descriptions last May. And the most common negative description today? “Liar”, which nobody even mentioned last May.

Bush's approval rating in the poll has fallen to 48 percent (down 8 points since mid-January), the lowest Pew has ever recorded. His approval rating in the last month has dropped has dropped 9 points among white women, 10 points among those 30-49 years of age, 11 points among women under 50, 11 points among white Catholics (a critical swing group), 12 points among high school graduates, 12 points among white non-evangelical protestants and 16 points among those in rural areas.

The Pew poll has Kerry and Bush tied in a trial heat question (47-47), while other current polls show Kerry ahead, but this is still quite a shift from Pew's mid-January poll, when Bush was ahead by 11 points (52-41) in a matchup with Kerry. Republicans haven't budged in the last month, while Democrats have consolidated behind Kerry in the trial heat question and independents have dramatically shifted away from Bush, going from 52-37 Bush to 51-41 Kerry, a swing of 25 points. Similarly, political moderates have gone from 49-43 Bush to 55-39 Kerry, a shift of 22 points.

Another huge shift has been among white non-evangelical protestants, who have gone from 57-36 Bush to 49-46 Kerry, a swing of 24 points. It's also interesting to note that Bush's current lead over Kerry among white men is 16 points--believe it or not, a sign of weakness. In 2000, Bush beat Gore by 24 points among this group.

Big shifts among independents can also be seen in changing evaluations of the political parties. Last June, independents gave the Democratic party a 55 percent favorable evaulation and the Republican party a 54 percent favorable rating. Now 65 percent of independents rate the Democratic party favorably, compared to just 50 percent favorable for the Republicans. And it is entirely because of this shift among independents that the overall public now rates the Democrats more favorably than the Republicans (58 percent to 52 percent).

It's also intriguing to note that independents rate the state of Massachusetts (76 percent favorable) more highly than the state of Texas (70 percent). Hmmm.

More on the shifting political terrain tomorrow.....

Poor John Kerry. First, he gets beat up for having so many people think he's electable. You see, that's just the "bandwagon effect" or, for those who wish to dress that point up in fancier intellectual clothing, it's an electoral "bubble", just like the run-up of the Nasdaq. To put it in the simplest possible terms, the reason people think he's electable is 'cause other people think he's electable. Of course, this dynamic is always an important part of any frontrunner's success, so criticizing Kerry for benefiting from the bandwagon effect didn't really have a lot of force. Things that are always true are....well, always true.

But, in the aftermath of the Wisconsin primary, the Kerry-doubters have shifted their line of attack to something with a bit more bite: hard evidence (in their view) that Kerry is less electable than the other viable Democratic candidate, John Edwards.

The basic rap couldn't be simpler and DR pointed out the factoid upon which the rap is based on Tuesday night: Edwards, as he has consistently done, performed much better among declared independents than among Democrats in the Wisconsin primary. And he not only performed better, he actually beat Kerry by a solid 12 points among these voters, 40 percent to 28 percent

Conclusion: since you need to reach independents, not just Democrats--who presumably would support either candidate--to win a general election, Edwards is clearly more electable than Kerry.

To which DR says: maybe. But then again, maybe not. Here are a number of objections which, taken together, suggest the Edwards electability thesis is not quite the irrefutable case that some Kerry-doubters seem to think it is.

1. Edwards may have beaten Kerry among independents in Wisconsin, as well as South Carolina, Oklahoma and Tennessee. But Kerry beat Edwards among independents in Iowa, New Hampshire, Arizona, Delaware, Missouri and Virginia.

2. Independents who vote in the Wisconsin Democratic primary are likely to be fairly liberal. Therefore, just as you can argue that Wisconsin Democrats are likely to vote for either Kerry or Edwards, so are these kind of independents, once they are confronted with the choice between Edwards or Kerry and Bush. So Edwards' advantage over Kerry among Wisconsin independent Democratic primary voters does not necessarily imply a real advantage among the much larger universe of general election independents.

3. When it comes to general election independent voters, Edwards will not have the advantage of little press scrutiny and almost all favorable publicity, like he did in Wisconsin. The Wisconsin primary was a poor test of Edwards' ability to hold up under the spotlight, since the spotlight was almost all on his opponent.

4. General election independent voters, unlike independents in the Wisconsin primary, are likely to be quite sensitive to the national security issue and the credentials of the Democratic candidate to be commander-in-chief. Kerry has a huge advantage here over Edwards and it baffles DR that some observers are glossing this over. The ability of Kerry to highlight real combat experience and campaign with war veterans adds a great deal to the Democrats' ability to appear tough enough to defend the nation's security. Not to mention the implicit comparison with Bush's cushy, politically-connected, non-combat gig in the National Guard, a comparison which is unflattering for Bush, no mater what happens with the AWOL story. Independents who don't like Bush, particularly working class independents, are going to be looking for a way to feel safe about voting Democratic in this election. Kerry has a leg up over Edwards in his ability to provide the assurance.

5. Edwards may have done well among independents in Wisconsin, but, judging from the publicly-released exit poll data, it appears to have been mostly among upscale independents. The Democrat in the general election will have to do well among working class independents as well (not to mention keeping working class Democrats in the fold). Kerry actually has a good track record of support from working class, especially traditional blue collar, voters in Massachusetts. As Michael Crowley points out in an excellent article in the new issue of The New Republic, Kerry has repeatedly shown he can connect with those kinds of voters, where his military and combat experience and friendships with veterans loom large, even when other politicians and elites typecast him as aloof and out-of-touch.

6. In the latest CBS News poll, Kerry wins among all voters by 5 points, but among independent voters by 10 points. But isn't it true that voters don't like Kerry and don't cotton to him personally?

In the latest Time/CNN poll, by 68 percent to 28 percent, the public says that "likeable" applies to Bush, which is good....but not quite as good as the 69 percent to 18 percent by which public says that Kerry is likeable.

And look at other characteristics. In the latest Gallup poll, Kerry has a net rating (applies-doesn't apply) of +33 on being a strong and decisive leader, about the same as Bush's +32; a net rating of +38 on being honest and trustworthy, far higher than Bush's +13; a net rating of +20 on shares your values, compared to Bush's +7; and a net rating of +16 on "generally agrees with you on the issues you care about", compared to Bush's -3.

And check out this one: Kerry has a net rating of +57 on "did his duty for the country during Vietnam", while Bush's net rating on doing his duty is +2. +57 vs. +2. Wow.

Gallup also finds that 15 percent say Bush's actions when serving in the Guard make them less likely to vote for him, compared to 4 percent who say it makes them more likely. With Kerry, on the other hand, it's just the reverse: 19 percent say his combat experience in Vietnam makes them more likely to vote for him, compared to 2 percent say it makes them less likely.

So......Don't get me wrong, I like John Edwards. And it's certainly a respectable hypothesis that he is more electable than John Kerry. But, to put it mildly, I don't think the results of the Wisconsin primary put that hypothesis into the realm of established fact. Indeed, the alternative hypothesis, that Kerry is more electable than Edwards, seems just as plausible, if not more so, once you put those results in a broader context.

Gary Langer, ABC News polling director, has a very good article up on the ABC News website about the silly idea that "NASCAR dads" will be the swing vote that decides the 2004 election. Langer points out that the definitions of NASCAR dad vary wildly and--the fatal flaw--tend not to be swing voters, no matter what the definition. Here's what he came up with when he took the typical characteristics commentators have attached to this group and actually ran the numbers (!)

When we run data from our recent polls we find that married, middle- and lower-income white men account for a single-digit share of the national population, and support President Bush in precisely the same proportion as all white men. (Make it rural white men, and it goes down to low single digits.) And white men, particularly Southern white men, are a solidly Republican group, highly unlikely to swing anywhere, anyhow.

For good measure, we checked rural, suburban or small city married white men with children and incomes under $50,000 in the 2000 exit poll. They accounted for 2 percent of all voters, and supported Bush over Gore by 70 percent to 27 percent. You really want to call this a swing voter group?

Not over here in DR-land, Gary, we know better! And just to stick a fork in this one, here's what DR's favorite nonpartisan analyst, Charlie Cook, had to say on the subject:

But this business about the "NASCAR dad" being the swing voter group of the 2004 election, or any other national election, is one of the dumbest ideas I've heard in my 32 years in and around politics. In NASCAR fans, we are talking about an overwhelmingly white, disproportionately male and Southern electorate. It's also disproportionately working- and middle-class, and in the 30-39 year age bracket, the age group where Bush is strongest.

Don't get me wrong -- these are terrific, hard-working, salt-of-the-earth type people. But any group that is disproportionately white, male, Southern, working- and middle-class 30-somethings is not made up of swing voters.

You tell 'em, Charlie! So: who are the real swing voters? Back to Langer, whose simple definition is a thing of beauty.

In our view, a swing voter group ought to fit two basic criteria — its majority vote ought to swing between Democratic and Republican candidates from election to election; and it ought to be big enough to make a difference in the outcome.

Exactly. And one of the groups that fits this definition quite crisply is an old DR favorite: independent voters. Rather than wasting our time trying to figure out how to reach NASCAR dads, let's try to figure out where independent voters are coming from this year.

Fortunately, that isn't so difficult. We can just ask 'em. According to a just-released CBS News poll, independents give Bush a 46 percent overall approval rating, a 45 percent approval rating on foreign policy, a 46 percent approval rating on Iraq and a 40 percent rating on the economy. They think the country's off on the wrong track by 55 percent to 37 percent and they give Bush only a 41 percent favorability rating. By 52 percent to 40 percent, they don't think Bush has the same priorities for the country that they do.

They favor a generic Democratic presidential candidate over Bush by 48 percent to 38 percent and they also prefer John Kerry over Bush by the same margin.

By 53 percent to 42 percent, they're uneasy about whether Bush can make the right decisions on the nation's economy. About three-quarters think Bush administration policies have either decreased the number of jobs or had no effect and three-quarters also think his policies have either made their taxes go up or had no effect on their taxes.

In terms of foreign policy, 51 percent think Bush is "war president" because of the choices he made, not because world events caused him to be one (38 percent). By 52 percent to 39 percent they don't think the result of the war with Iraq has been worth the loss of life and other costs. In addition, 62 percent believe that either that Iraq was a threat the could have been contained or that it was not a threat and 61 percent believe the Bush administration either hid important elements or mostly lied about what they knew about Iraq's WMD.

Sounds promising. They're (a) real swing voters and (b) accessible to the Democrats in multiple and important ways. Who needs fake swing groups like NASCAR dads when you've got these kind of voters to work with?

Sure he does. Not a particularly good one, but a chance nonetheless. With his strong second in Wisconsin--at this point, 40 percent to 34 percent with 99 percent of the precincts reporting--he has defied expectations and damped Kerry's momentum, at least temporarily. The last pre-election poll had Edwards down 27 points to Kerry, so closing that gap to 6 points or so is a pretty amazing accomplishment.

And with Dean's distant third in the primary (at 18 percent), it seems likely he'll either drop out or become a complete non-factor anyway. If Edwards can pick up a disproportionate amount of Dean's erstwhile support, that help him turn this opening into a serious, rather than token, two person race (though note that in Wisconsin voters who said they had supported Dean in the past gave about the same number of votes to Edwards and Kerry).

That said, Kerry has now won 15 of 17 contests and leads Edwards by over 400 delegates, 608-190. And the latest CBS News poll, conducted February 12-15, has him trailing Kerry nationally by 53 percent to 7 percent among likely Democratic primary voters.

Of course, we all know how quickly momentum can shift against the national front-runner this campaign season--though the primary process is now so far along that such a shift is far less likely against Kerry than it was against Dean. Still, it's interesting to note that the Rasmussen national tracking poll of Democrats, for what it's worth, has Edwards narrowing the gap with Kerry from 51-18 to 43-25 in the last three days.

Looking at the exit poll results, there are many intriguing results, some of which make Edwards look good (though more as a general election candidate than a Democratic primary candidate), and some of which do not.

In terms of standard demographics, Kerry and Edwards pulled about the same amount of white support (37 percent), with Edwards doing slightly better among white men and Kerry doing slightly better among white women. But Kerry clobbered Edwards among both black voters (55-15) and latino voters (46-23).

Kerry also carried every income group up to $75,000 with his largest margins among the lowest income voters (50-22 among those with less than $15,000). But Edwards beat Kerry among those between $75,000 and $100,000 (41-34) and tied him among those with $100,000 and over.

Edwards also lost every education group up to some college, but tied among college graduates and carried those with a postgraduate education (36-32).

Consistent with these patterns on education and income, Edwards carried the 21 percent of voters who said their family financial situation had gotten better (37-30), but lost among the 75 percent of voters who said their financial situation had gotten worse (45-32) or stayed the same (39-36).

Looking at partisanship and ideology, Edwards, as he has consistently done, performed much better among declared independents than among Democrats. In fact, he managed to actually beat Kerry among independents, which he has not typically been able to do, by 40 percent to 28 percent. But he lost among Democrats (62 percent of voters) by 48 percent to 21 percent

Edwards also won among conservative voters (37-30) and just barely lost moderate voters (41-39). But liberal voters prefered Kerry, 41 percent to 30 percent.

Intriguingly, Edwards lost handily among those angry (43-28) or dissatisfied (44-37) with the Bush administration (82 percent of voters), but beat Kerry among the small numbers either satisfied (50-22) or enthusiastic (36-10) about the administration. Edwards also fared well among the 16 percent of voters who said the national economy was good (40-28), but lost among the 79 percent of voters who said it was not good (41-36) or poor (45-29).

Interesting! So the more overtly populist candidate got the happy camper vote and lost those most dissatisfied with the way things are going. And then to deepen the puzzle, Edwards carried economy and jobs voters (46-35), the single largest group of issue voters, at 41 percent, but lost by wide margins on all the other top issues: health care, Iraq, education and taxes. So one would infer that the economy and jobs voters tended to be disproportionately well-off and satisfied with the state of the country. Whodda thunk it?

One last thought: some commentators are opining that Edwards did well because, like Iowa, Wisconsin has this large and significant group of rural and small town Democratic primary voters. But Edwards actually lost to Kerry 44-33 in rural areas and 41-34 in small towns. His real attraction apparently was in suburban areas, which he carried 39 percent to 35 percent and in small cities, which he narrowly lost 37 percent to 35 percent.

Much food for thought. And perhaps for a more interesting and competitive race for the next few weeks than many--including DR--had believed possible.

Historically, incumbent presidents have fared well when economic growth numbers were good and not so well when they weren't. That's why, when economic growth in the third quarter of last year came in at a sizzling 8 percent annualized rate, pundits were quick to pronounce Bush's re-election a sure thing.

Since then, economic growth has subsided, but still came in at a solid 4 percent rate in the fourth quarter of last year. Most economists expect this kind of reasonably good growth to continue. By the standards of most economy-based election forecasting models, this should be good enough to give Bush the election.

But will it? DR has previously written on the foibles of these election forecasting models and their spotty track record. And here is a very interesting analysis released by Gallup, based on their latest poll, that goes into some detail about how the jobs issue continues to bedevil Bush, even as the growth rate has improved. The key findings are:

1. The public thinks Bush's economic policies do matter. Nine in ten think the president's policies affect national economic conditions and 8 in ten think his policies affect their own personal economic situations.

2. The number of people mentioning some aspect of the economy as the top national problem has started to increase again. That number has now reached 46 percent, including 20 percent who specifically mention unemployment or jobs. The latter figure is twice as high as a year ago and the highest figure for unemployment/jobs as the top national problem in 10 years.

3. Consumer confidence in the economy is declining again. In the last month, there has been a 10 point decline in the number saying current economic conditions are "good" or "excellent" and a 5 point increase in the number saying they are "poor". There has also been a 13 point decline in those who say economic conditions are getting better and a 13 point increase in those who say they are getting worse.

4. Bush's approval rating on the economy--consistent with other national polls--has slipped to net negative (45 percent approval/52 percent disapproval) after being net positive (54/43) a month ago.

Conclusion: weak job growth is dragging down Bush's image on the economy, despite the pickup in overall economic growth. Therefore, economic growth is unlikely to be enough, by itself, to lift Bush to re-election.

Corollary: the Democratic position on jobs in 2004 will be key. You can check out Edwards' position on jobs here and Kerry's position on jobs here. Both Edwards' and Kerry's approaches seem reasonable enough as far as they go, but a bit scattered and difficult to summarize in a crisp fashion. And both give relatively short shrift to infrastructure spending and put fairly heavy emphasis on tax incentives for manufacturers and getting tough on trade. It seems to DR that reversing the relative emphases here might be desirable given the easily-understood connection between infrastructure spending and job creation, particularly of non-manufacturing jobs. Non-manufacturing jobs, after all, are the overwhelming proportion of jobs in the United States and the kind of jobs most American voters hold. If Democrats hope to capitalize effectively on the jobs issue, they need to be responsive to that reality.

Ruy Teixeira is a Senior Fellow at The Century Foundation and
the Center for American Progress