Kerry Spot - States Of Haze
This Boston Globe article on Bush's slipping numbers in battleground states caused a bit of a stir among campaign watchers earlier this week.
The headline itself is the type of news that would probably cause Karl Rove to awake suddenly in the middle of the night, his elephant pajamas drenched in a cold sweat: "Bush may need broader campaign, Slips in polling in crucial states."
"Surveys in battleground states confirm the national slide, suggesting that Bush would have a very hard time defeating Kerry if the election were held today," the Globe reports. "State by state, Kerry appears to be doing better than former vice president Al Gore was at this stage of the election in 2000, closing the margin or beating Bush in every state where Gore was competitive, except in Tennessee."
And the article points to what initially looks like a strong pile of supporting evidence: A Quinnipac University poll showing Kerry up 44 to 41 in Pennsylvania. An American Research Group poll shows Kerry winning Ohio, 49 to 42. Then they turn to Iowa — which hasn't garnered that much attention as a swing state — and cite a Research 2000 poll showing Kerry up 46 to 42.
Cue the ominous music! But then the Globe puts out the counterarguments. A more recent Ohio poll of 1,500 registered voters has Bush 47, Kerry 41. Michigan voters surveyed by the Detroit News on May 11-12 gave Bush the edge over Kerry, 44 percent to 40 percent. Kerry's only up by 3 in New Jersey, which was thought to be one of the bluest of blue states. And recent polls show Oregon and Florida as statistical tie.
The article's premise — that Bush is in trouble — is plausible, considering the rough spring he has had. But if you look at all of the polls cited in the article, the overall picture is...messy, with a lot of contradictory data. Kerry can look at all the states he is either leading in or competitive in and tout himself as a challenger that has already proven he can go toe-to-toe with Bush. But he also hasn't nailed down many of Gore's states, and there's scant evidence that he's definitively brought any of the red states into his column.
Some polls ought to make Kerry advisers Mary Beth Cahill and Bob Shrum do cartwheels, and some of them ought to make them kick the cat.
To really say the battleground states are turning towards a candidate, one would need results like John Zogby's latest batch. Zogby recently unveiled a slew of polls that suggested Kerry is working his way toward a 538-0 Electoral College rout. Perhaps that's a mild exaggeration — but almost everything was breaking the Democrat's way, with Kerry winning Bush states Florida, Ohio, New Hampshire, Nevada, and Missouri. With Iowa the only Gore state slated to go Republican in November, Zogby's polls forecast a 320 to 218 rout. In addition, he has several swing states almost out of Bush's reach; in Pennsylvania, Kerry up by 8.2 points on a on 3.8 percent margin of error, in Wisconsin, it's another 8.2 point lead on a 3.4 percent margin of error.
So why is Zogby's poll so out of whack with everyone else's? Well, for starters, it's an Internet-based poll. "Likely voters from each state followed instructions sent by e-mail that led them to a survey on Zogby's secure servers. Slight weightings were applied to ensure that the selection of participants accurately reflects characteristics of the voting population." Is Internet-based polling more accurate than telephone-based polling? We will see. But it's worth noting that in Zogby's May 19-20 telephone poll before the South Dakota special election, he had Democrat Stephanie Herseth leading Republican Larry Diedrich 52 percent to 41 percent. Actual results: Herseth 51, Diedrich 49. Maybe his weightings are overestimating Democratic turnout a bit.
The big, clear conclusion of the latest round of swing state-by-swing state polls is...that there is no big, clear conclusion. Almost every swing state in the last month has shown at least one poll with Bush or Kerry up by a solid margin. A quick review of the comprehensive collection of state-by-state polls gathered by RealClearPolitics.com and Gerry Daley reveals:
Arkansas: Bush up by 4, Bush and Kerry tied.
Florida: Bush up by 1, Kerry up by 2.
Iowa: Bush up by 5, Kerry up by 4.
Michigan: Bush up by 4, Kerry up by 8.
Minnesota: Kerry by 3, Kerry by 12.
Missouri: Bush up by 1,Kerry up by 3.
New Mexico: Bush by 1, Kerry by 5.
Ohio: Bush up 2, Kerry up 4.
Oregon: Bush up 5, Kerry up 5.
Pennsylvania: Bush up by 2, Kerry up 3.
Wisconsin: Bush by 12 (at end of April), Kerry by 8.
Some of these are outliers, those 1 in 20 statistical blips that will be far off the actual numbers for the entire population. But the overall range, with many of these polls taken just weeks apart, suggests that these swing states are swing states for a reason. It's close. There are large numbers of undecided. And any lead inside the margin of error is probably illusory or temporary.
But one side effect of all these polls is that any reporter looking for a sign of Bush's impending triumph or impending doom can just pick the results he choose.
One of the few portents that may be significant are the are the number of states that might have seemed secure but are now considered on the fence. Kerry's got a lot of Gore states to defend: New Jersey, Oregon, Iowa, Minnesota, New Mexico, Wisconsin, and maybe Washington, besides the usual Ohio-Pennsylvania-Michigan-Florida quadrangle dogfight. On Bush's side, he's in trouble of losing New Hampshire's four electoral votes, and will need to work to keep West Virginia in his corner. But it looks like a broader battlefield than 2000.
But the next time some pundit cites a single state poll as a definitive sign of where the race is going, exercise healthy skepticism.
(6/4/2004)
- By Jim Geraghty, National Review Online
zogby.com
My comment:
Interesting to see the picture change after Kerry picks a running mate. |
