Newsweek Poll

A recent Newsweek poll shows John Kerrt with a slight lead over President Bush, 47% to 45%.

While some democrats are gloating, the poll is within the margin of error and there are some porblems with the polling sample.

The sample used by Newsweek has 5% fewer Respublicans than other recent polls and only includes voters on the west coast.

Voters in battlegorund states and the south were not included.

More below.

THE NEWSWEEK POLL

The Kerry Spot's Kerry-backing readers are very excited about the Newsweek poll, showing Kerry up 3 among registered voters. (The lead is 47-45 with Nader thrown in.)

In further good news for Kerry, the Los Angeles Times conducts its own poll of those who watched the debate:

The voters who watched the debate were slightly more favorable to Kerry than the overall electorate even before the encounter began, the poll showed.

For instance, in last week's Times poll, Kerry trailed Bush among all registered voters by 49% to 45%. But the voters who watched the matchup preferred Kerry by 48% to 47% for Bush before the debate. After the debate, viewers divided nearly the same way, with 49% favoring Kerry, 47% Bush.


(Interestingly, as Drudge noticed, Bush's favorability rating among debate viewers actually improved slightly, although within the survey's margin of error. Before debate, 51% of watchers viewed Bush favorably, 49% unfavorably; after, numbers were 52% and 47%.)

Is this a real shift? Can a nearly-double digit Bush lead be blown away by one strong performance by Kerry, in which he touts a "global test"? Or are these samples skewed?

We will see in the next round of polls, and it will be particularly interesting to see if Kerry gets a similar bounce in the state polls. But for now, it seems safe to say that predictions of a George Bush landslide were at least a bit premature.

(For what it is worth, Rasumussen's robotic-polls are showing Bush continuing to hold about a three point lead - but their numbers do shift about a point a day.)

UPDATE: The blog Political Vice Squad says the Newsweek poll is skewed from its earlier one. They decreased Republican sampling by 5 percentage points and increased Democratic sampling by 6 percentage points. Furthermore, on the first of the three nights, the poll was limited to the "Pacific and Mountain time zones." In other words, registered voters from the following states completely were excluded: Texas, Florida, Ohio, Pennsylvania, Missouri, Louisiana, Arkansas, Kentucky, Indiana, and the entire old South.