Whew, it’s almost over. Don’t forget to vote, and would you like to know who’s going to win?

So would we all.

There is a new poll. The survey by Rasmussen Reports shows Democrat Peter Shumlin ahead of Republican Brian Dubie by 50 to 45 percent, with three percent undecided and one percent supporting minor candidates.

The poll was taken of a randomly selected sample of 750 likely voters and has a margin of error of plus or minus four percentage points.

Do not despair, Dubie-ites. Just taking the poll at face value, it does not “prove” that Shumlin is going to win. Remember, another poll taken two weeks earlier, by anther firm, gave Dubie a one-point lead. Only the count of actual ballots tomorrow night will reveal the winner (and if the race is as close as was the Democratic Primary, more counting may be needed).

What is often forgotten is that a poll’s margin of error applies not to the “spread,” but to each candidate’s percentage of support. All this poll says is that had all 250,000 or so likely voters been polled, between 54 and 46 percent would have chosen Shumlin, between 49 and 41 percent would have opted for Dubie.

So if on Tuesday night Dubie ends up with 47 percent of the vote to Shumlin’s 46 percent, the poll would not have been wrong. The result would have been within the poll’s margin of error.

(And would require the State Legislature to elect a governor, a matter to be discussed here only if it happens, which it probably won’t).

But don’t be of too much good cheer, either, Dubie-ites. A poll’s reported results represent the midpoint of its findings, which is far more likely to reflect actual public opinion than either extreme. It’s as likely that Shumlin is really ahead 53-42 as that Dubie is ahead 47-46.

Now we come to the question of whether this poll should be taken at face value. The answer will offer no comfort to Dubie and the Republicans, but the question should be explored anyway.

Rasmussen polls have a slight Republican bias. This is not because proprietor Scott Rasmussen has a slight Republican bias. Rasmussen in fact has a colossal Republican bias. But he’s also, on the basis of the evidence at hand, a perceptive businessman (he helped found ESPN), and for the sake of his business,  a pollster tries to make his polls accurate so that people take them seriously and continue to hire his firm. A series of inaccurate polls can put a pollster out of business.

But when a voter answers his or her phone because the Rasmussen folks are calling, an actual folk is not at the other end of the line. Instead, a recorded voice asks the respondent a question to be answered by pushing ‘one’ for candidate A, ‘two’ for candidate B, etc.

This ‘robocall’ technique is frowned on by the polling establishment at the American Association of Public Opinion Research . and scorned by several quality news organizations including the Washington Post and the New York Times. In fact, Rutgers University political scientist Cliff Zukin, in a paper he wrote last year for the AAPOR, argued that “these types of surveys have little claim of scientific validity and probably should not be reported.”

One reason for doubts about robocall polls is that they can not include random sampling within the household. A person on the phone can ask to speak to a woman or a man, or the person with the most recent birthday, or use some other method to try to get a representative sample.

But as Zukin noted, “If interviewers simply spoke with whoever answered the phone, the resulting samples would be older and more female than the population as a whole.”

That could be why robocall polls have been found to tilt about two percentage points more Republican than other surveys, or than the final election results.

On the other hand, Rasmussen polls don’t have a bad track record for accuracy. In the Massachusetts Senate race won by Republican Scott Brown earlier this year, Rasmussen polls picked up Brown’s growing strength earlier than other surveys.

With all the caveats, then, it’s fairly likely that Shumlin is ahead.

Which is not the same thing as predicting he will win. It might rain (though right now the Weather Channel is predicting pretty nice weather for both Burlington and Rutland on election day) and bad weather usually keeps more Democrats than Republicans away from the polls. The Republicans might have a better get-out-the-vote operation this year (though they usually don’t). And the poll could just be wrong. Sometimes, a poll is simply wrong. Polls, after all, are based on the laws of probability, not the laws of certainty.

Rasmussen is not the only outfit polling the race. So are both campaigns. “The campaign has maintained a consistent lead in our internal polling and as the campaigns (sic) momentum grows with each day we expect to win on Tuesday night,” said Dubie campaign manager Corry Bliss after the Rasmussen poll came out.

In the same statement, Bliss tried to argue that the Rasmussen poll showing Dubie behind by five points actually predicted an eight-point victory, because polls in 2002 had understated Vermont Republican strength by 13 points.

As responses go, this one was more inventive than persuasive.

Neither Bliss nor the Shumlin campaign revealed any numbers from their internal polling, requiring the curious to engage in tea-leaf reading to try to figure out what the polls might show.

Until the weekend, there were some signs that Dubie might be ahead. Conventional political strategy is to shift, in the final week or two, from television advertisements attacking the opposition to more positive ads, often featuring the candidate talking directly to the camera.

Last week the Shumlin campaign and its allies were still running anti-Dubie ads, while the Dubie campaign was featuring a spot called, “I’m your man” in which a blue-shirted, smiling, Dubie looked directly at the camera and promised to focus on jobs.

It’s a very effective ad even though its main point – “Vermont’s economy is in crisis” – is wrong. By all the usual criteria – economic growth (or, these days, shrinkage), unemployment rate, poverty rate, foreclosure rate – Vermont’s  economy is less in crisis than that of most other states.

This would be the downside of a Dubie governorship—the constant (and, worse, inaccurate)  griping about Vermont’s economy (there would be upsides, to be discussed later regardless of the outcome).

By the weekend, however, the Dubie campaign was back on the attack, including one ad – falsely accusing Shumlin of wanting to let prisoners out of jail – that has been widely criticized for being not just inaccurate but dishonest.

Considering that even many Republicans have been speculating that it is ads like this that seem to have put Dubie behind in the race, the decision to keep running it could be a sign of some desparation.

Maybe the Dubie campaign’s internal polls weren’t that different from the Rasmussen numbers.

On Wednesday afternoon, the Vermont Democratic Party decided that, on second thought, it would allow reporters to cover its Sunday evening fundraiser starring Rep. Barney Frank of Massachusetts.

A wise move. When a powerful House committee chairman speaks in a quasi-public setting, he might commit what some reporters call a “new” (or perhaps a gnu?). That is, he might make news. News ought to be covered by folks whose business it is to cover the news, not because they have such a wonderful time doing so (it’s often a drag) but because it’s better to have an informed public that an uninformed public.

Such, at least, is the theory behind democracy, or, as it’s sometimes known, a republic, meaning both Democrats and Republicans should facilitate rather than encumber the process. For the nonce, the Democrats seem to have figured this out. Some of us will be watching all political parties for future compliance.

Speaking of politics, let’s deal with that poll about the governor’s race that came out the other day, and with the reaction to said poll.

According to a Rasmussen Reports survey of 500 randomly selected Vermonters, Lt. Gov. Brian Dubie, unopposed for the Republican nomination, is leading all five Democrats running in the Aug. 24 primary to oppose him.

Dubie’s lead over Secretary of State Deb Markowitz is a modest one, 47 percent to 40 percent, within the poll’s plus-or-minus 4.5 percentage point margin of error.

Against the other four, Dubie gets more than 50 percent, though just barely (51) against Sen. Doug Racine. But the lieutenant governor has a 15-point margin over Racine, and bigger leads against Sens. Peter Shumlin and Susan Bartlett and former Sen. Matt Dunne.

Bartlett fared worst, trailing Dubie by a 55-to-29 percent margin, so perhaps it was not surprising that her campaign manager, John Bauer, issued a statement questioning the validity of the poll. Bauer quoted the recent comment of Middlebury emeritus political science professor Eric Davis that 
”Rasmussen polls should be interpreted cautiously(because)  Rasmussen uses automated prompts with touch-tone responses rather than live operators to collect its data.”

That’s true, and Rasmussen’s polls are controversial, partly because Scott Rasmussen makes no effort to hide his own very pro-Republican sympathies.

But that doesn’t mean the polls are inaccurate. Even Nate Silver of  538.com, a critic of Rasmussen’s methods, concedes that the company’s polls have a higher than average record for calling races right.

Some of the poll’s findings at least raise the possibility that the survey over-sampled Republicans. For instance, it found Vermonters almost evenly split on supporting off-shore oil drilling.

Maybe they are, but it would be risky to run for office here on a pro-drilling platform.

Still, there’s not much reason to doubt that the Rasmussen poll is an accurate snapshot of the campaign right now.

But that’s all. The dynamic of the race will change on the night of August 24, when the primary results come in. (Rasmussen, confused, used the old date of September 14, but that doesn’t have any impact on the poll’s accuracy).

The more pertinent question right now, then, is not whether Dubie is ahead of the Democrats, but which Democrat is ahead of the others. That question can’t really be answered until next month, when Research 2000 releases its poll on the Democratic race. But it can sort of be answered now, as long as everyone understands that the answer is tentative and even a touch shaky, because two of the campaigns, Markowitz’s and Shumlin’s have hired respected pollsters to check out the race.

The results have not been made public. They haven’t even been quasi-officially leaked. But some people who have seen the results have…well, let’s just say they’ve chatted. In pubs, on the phone, in meetings. Hints have been dropped. With that shakiness caveat firmly in mind, the race seems to be shaping up this way:

Markowitz is ahead. Racine is second, though just how distant a second is hard to determine. Shumlin is an apparently not-very-close third. Dunne and Bartlett remain in single-digit percentage figures.

These results do not justify a prediction that Markowitz will win, or even justify an assessment concluding that she’s the front-runner.  Nor do they mean that Dunne and Bartlett have no hope. To begin with, this is  very hard race to poll. It all depends on turnout, which is likely to be quite low, and while the pollsters are probably “screening” for likely voters, that’s a most imprecise exercise under the circumstances.

All the pollsters can do is ask respondents whether they plan to vote in the primary. Ok, they can push a bit: “Are you sure you’re going to vote?” Or, “are you enthusiastic about the race?” Or some such. But for a couple of reasons, many folks may say they really, really plan to vote, but end up staying home.

Or, in this case, being away. August 24 is in the middle of vacation season. Some of those respondents who said they were going to vote may have forgotten that they have reservations at a resort or plans to visit their Aunt Louise in California on that date.

Yes, there’s early voting. It starts July 12, and all the campaigns have early-vote operations to urge their supporters to fill out their ballots before taking off for the beach or Aunt Louise’s. But one of the peculiarities of this campaign raises the likelihood that many voters may not bother.

The peculiarity is that all five of these Democrats are just fine with the vast majority of Democratic voters. This is not an ideological battle. There are some policy differences, but all the contenders are mainstream, center-left Democrats. Nor is it one of those contest between “regulars” and “reformers,” whatever that means. So far, at least, nothing resembling a “character problem” has arisen about any of candidates.

The result is that while few rank-and-file Democrats are all that enthusiastic about any of the five, neither do they harbor any hostility toward one or two of them. Harboring hostility is  great turnout builder, absent in this primary.

But there’s one more thing to keep in mind. The dynamic of the primary is likely to shift next month, when that Research 2000 poll comes out and when the candidates report their fund-raising results on July 15. The candidates who bring up the rear in both measurements (and whoever trails in one is likely to be trailing in the other) are going to be in trouble. Both press and public will conclude that they are out of it.

That may not be fair. But it’s what almost always happens. At that point, the race will take on a new shape. Don’t expect anyone to drop out officially. But at least two of those Democrats will effectively spend the last month or so going through the motions, with most attention centered on the two or three at the top.

That’s when it’ll get interesting.

And Don’t Forget: The News Guy is on Vermont This Week tonight (repeated Sunday).