I first posted this in another thread. I think it also goes here quite nicely.
Of course there is a strong correlation. The people of the nation and the people of the states are the same people. But the correlation is not absolute, and exists only in so much as the state averages drive the national averages. Some states are more partisan than others.
Massachusetts was certainly a Republican state in 1984. It was also certainly less Republican than the nation, if that's what you want to say.
Philip is correct. The national numbers do not change the state numbers; the state numbers change the national numbers. If California moves to the Republicans by enough to shift the national numbers by 1% that does not mean a single person in any other state will change their mind and vote differently. No one in Rhode Island is going to say "Well, the national numbers moved toward Bush so it is my responsibility to change my vote so my state moves along with them."
As much as some people would like to put Nader's votes onto Kerry and Gore and Buchanan's and Peroutka's votes onto Bush you can't. Those votes were not cast for those candidates or parties; they were cast for a third party for a reason. If the third party had not run things would be different, but they did run so things are not different.
Let's look at a few states:
BATTLEGROUND STATES:Oregon:
Difference in 2000: .4% D
Difference in 2004: 4.2% D
Change: 3.8% to D
New Hampshire:
Difference in 2000: 1.2% R
Difference in 2004: 1.4% D
Change: 2.6% to D
Ohio:
Difference in 2000: 3.5% R
Difference in 2004: 2.1% R
Change: 1.4% to D
Nevada:
Difference in 2000: 3.6% R
Difference in 2004: 2.6% R
Change: 1% to D
Minnesota:
Difference in 2000: 2.4% D
Difference in 2004: 3.4% D
Change: 1% to D
Wisconsin:
Difference in 2000: .2% D
Difference in 2004: .4% D
Change: .2% to D
New Mexico:
Difference in 2000: .06% D
Difference in 2004: .8% R
Change: .86% to R
Iowa:
Difference in 2000: .3% D
Difference in 2004: .7% R
Change: 1% to R
Michigan:
Difference in 2000: 5.1% D
Difference in 2004: 3.4% D
Change: 1.7% to R
Pennsylvania:
Difference in 2000: 4.2% D
Difference in 2004: 2.5% D
Change: 1.7% to R
Missouri:
Difference in 2000: 3.4% R
Difference in 2004: 7.2% R
Change: 3.8% to R
Florida:
Difference in 2000: .01% R (maybe)
Difference in 2004: 5% R
Change: 5% to R
What do we see here? Very little movement. Five states voted more Democrat than last time and 6 more Republican. Most states saw very little change. Kerry may have been helped in New Hampshire by being from a neighboring state. To see if this is real movement we need to wait another cycle or two.
Same goes in Florida. Was the 5% gain real movement or was it bumped by the hurricane relief efforts Bush gave?
The only state I think saw genuine movement that will likely last is Missouri. It seems to be joining the "solid south" for the Republicans. Minnesota, Oregon and Michigan may have also seen some real movement. Everything else is too little or the state has too long a history of being a battleground for the movement to mean much.
SECOND TIER BATTLEGROUND STATESWashington:
Difference in 2000: 5.6% D
Difference in 2004: 7.2% D
Change: 1.6% D
Virginia:
Difference in 2000: 8% R
Difference in 2004: 8.2% R
Change: .2% to R
Arizona:
Difference in 2000: 6.3% R
Difference in 2004: 10.5% R
Change: 4.2% to R
Louisiana:
Difference in 2000: 7.7% R
Difference in 2004: 14.5% R
Change: 6.8% R
Tennessee:
Difference in 2000: 3.9% R
Difference in 2004: 14.3% R
Change: 10.4% to R
These states were all thought at one point to be in play to one degree or another and each was thought to be removed from the table by the end of the campaign.
The only state that moved in the Democrats favor is Washington, everything else moved Republican to one degree or another. Virginia hardly moved at all.
The big jump in Tennessee and Louisiana could be, in part, due to the lack of a southerner at the top of the Democratic ticket. Either way they moved Republican solidly enough to impact the national number.
NON-BATTLEGROUND STATESVermont:
Difference in 2000: 9.9% D
Difference in 2004: 20.1% D
Change: 10.2% to D
North Carolina:
Difference in 2000: 12.8% R
Difference in 2004: 12.5% R
Change: .3% to D
California:
Difference in 2000: 11.8% D
Difference in 2004: 10% D
Change: 1.8% to R
Georgia:
Difference in 2000: 11.7% R
Difference in 2004: 16.6% R
Change: 4.9% to R
Utah:
Difference in 2000: 40.5% R
Difference in 2004: 45.5% R
Change: 5% to R
Rhode Island:
Difference in 2000: 29% D
Difference in 2004: 20.8% D
Change: 8.2% to R
Vermont saw a real, solid move to the Democrats. North Carolina saw a slight move that could possibly be from having the VP on the Democrats be from this state. We'll have to wait 4 years and see.
The big surprise here is Rhode Island running to Bush. I did not see that happening.
So where did the movement occur? Where did Bush gain his popular vote victory? Mostly in mid sized Republican strongholds turning out for him, especially in the south. In many of these states his margin of victory doubled or more having a noticeable impact on the national vote percentages. None of Kerry's big movers were large enough to have that effect.
I think the main cause for the change in the national numbers is the lack of a southerner at the top of the D ticket and the Kerry campaign's "small state" strategy.
If anyone can tell me how voters in California have any effect on the voters in Alabama, I sure would like to know. That is what you are saying when you insist the national average moves the states. If California had voted the exact same way it did in 2000 in terms of percentages the Bush victory would be reduced by roughly .3% to a 2.2% victory.
Have other states changed how much they are trending in one direction or the other? No, they have not changed one bit; only California has changed. But California is big enough that the change had a huge impact on the national numbers so it looks like other states changed if you compare them to the national number.
What you can do is compare a state's trend to the national average and say if it moved more or less than the national change, or if it bucked the change and went the other direction.
Huge changes in the national average are going to be seen in most states numbers. Not because the national numbers changed the states, but because a large change requires a lot of states to contribute to it.