http://www.tarrance.com/battleground/9786BGCharts.pdfThis poll is a classic goodnews/bad news poll.
In the "unaided ballot" question Kerry has gained 2% and now trails by just 2%, however in the Assisted ballot Bush gained 1% and is now tied.
Bush Job approval is up 1% to 51% versus last Battleground poll.
Strong improvement on economic matters for Bush, mild decline on National security/Iraq related matters.
Dems retain strong advantage on questions related to income redistribution, Kerry has grown a bit is personal positives, and has a modestly more positive set of favorability ratings than Bush.
Bush continues to have a substantial advantage on "character" issues.
Generically speaking, the "Democratic party" is viewed modestly more favorabably than the "Republican Party"
The $90 million in ads Bush has run attacking Kerry seems to have modestly done more damage to Kerry, than the $110 million Kerry and his 527 friends spent have done damage to Bush.
We are in the exact opposite phase from where the campaign was a few months ago.
In April when Bush was getting tons of bad news, Democrats were saying "Why the %#@! are Bush's numbers not droppping more?"
We are now getting very close to July and the economic perceptions of the population have
clearly turned positive and the Republicans are saying "Why the #$%@! arr Bush's numbers not rising more?"
I think this poll is luke warm good news for Bush when taken it it's totality.
A FEW TECHNICAL NOTES:
This poll is one of the few that does an absolute hard weight by party ID.
I disagree with doing a hard weight like this, but again it's a 55/45 thing - you can make good arguments either way
(Joe Montana threw a lot of TDs on passes he "should not have thrown" also)
When you get the breakout right, your polls are deadly accurate, when you get it wrong, you
structurally miss the mark.
In 1992 this poll was just dead on - they got the result exactly right.
In 1996 they missed by just 1%
In 2000 their turnout model was wong and the missed by 5%.
An AVERAGE error of just 2% over the last 3 races is excellent.
An error of 5% in 2000 is not so good...
More good news / bad news...
That warning aside,
This is from two very, very deeply respected and admired pollsters, one a Republican, and one a Democrat.
Read this poll, and the companion notes from the Republican (Ed Goaes) and the Democrat (Celinda Lake) VERY carefully. There is more REAL information here than in 10 "media" polls.