Trends for 2008

[Shira]

Trends, of course, develop over time, so predicting the outcome in 2008 will depend not only on what data set one uses, but how it is interpreted.

In the current polarized and partisan political situation, only swing states matter.  Either they are in play and the election will be close (like 2000 and 2004), or one side has such a dominant position they don't matter (like 1996 or 1988).

2000 and 2008 provide an excellent data set since one candidate, Bush, acts as an internal control to identify trends.  Looking at the increase in Republican (Bush) vote in the two elections, and continuing that trend, predicts a near Republican sweep of the swing states.

Below is a list of swing states and their change in % vote for Bush from 2000 to 2004, with a prediction for 2008 based on the same percentage change.  This predicts an easy Republican win in 2008:

                Bush%     Bush%   Projected  Winner
                 2000   2004      2008
MN             45.5   47.6   49.7   R
MI             46.1   47.8   49.5   D
PA             46.3   48.4   50.4   R
OR             46.5   47.2   47.8   D
WI             47.6   49.3   51.0   R
NM             47.9   49.8   51.7   R
NH             48.1   48.9   49.7   R
IA              48.2   49.9   51.6   R
FL              48.9   52.0    55.1      R
NV             49.5   50.5   51.5   R
OH             50.0   50.8   51.6   R
CO             50.8   51.7   52.6   R


Alternatively, we can assume most of Bush's gains in 2004 were due to the incumbent effect.  The 5 incumbents re-elected since 1950 all increased their margin, from 2% (Eisenhower) to 17% (Nixon).  If so then the real trend is how the swing states changed relative to the national result.  One can predict the 2008 result by subtracting the 2.8% national gain Bush had from his 2004 result in the swing states.  This predicts an sweep of the swing states by the Democrat:

                        Projected
                 2000   2004    2008
MN             45.5   47.6   44.8   D
MI             46.1   47.8   45.0   D
PA             46.3   48.4   45.6   D
OR             46.5   47.2   44.4   D
WI             47.6   49.3   46.5   D
NM             47.9   49.8   47.0   D
NH             48.1   48.9   46.1   D
IA              48.2   49.9   47.1   D
FL              48.9    52.0    49.2   D
NV             49.5   50.5   47.7   D
OH             50.0   50.8   48.0   D
CO             50.8   51.7   48.9   D

Same data, 2 interpretations, which will come closest to the 2008 result?

Trend is an important indicator if it covers the last 4 or 5 elections. Try to look at the trends since 1988. The number a candidate got in a state should be compared to the national number.

[Shira]

The term “Battleground State” does have a meaning only if the elections are close. If for, example, the winner in 2008 would get 54% nationally, then it is guaranteed that this person will win  all of the battleground states.
If you want to measure how conservative a state is, look at the number Bush got in that state and compare it to his national number 50.6%. I would say so: A state where Bush got by 2.5% above 50.6% is a safe state for the Republican candidate in 2008 (in close elections), while every state where he got less than 48% is guaranteed for the Democratic candidate (in close election certainly). All the states that in 2004 were in the 48%-53% range will be the swing states in 2008. The major and deciding swing states will again be OH, FL and PA.
As to VA, Bush got there 53.7% which is by 3% above his national number, but this number (the 3%) was 6.4% in 1988 and since then it is gradually and consistently shrinking. (migration from DC?). If the Democratic candidate will get 52% nationwide, then he/she will carry VA as well.

[Shira]

Trends from '80, '84, '92, and '96 are worthless because 84 was a landslide year and the rest had strong third parties whose votes could've broken either way.  Trends in themselves are pretty worthless anyway.

Third party or landslide victory doesn’t matter if you look into the really significant measure: The state number minus the national number.
Here are 2 examples as to GOP numbers:

In TX
1998 – 2.58%
1992 – 3.11%
1996 – 8.04%
2000 – 11.43%
2004 – 10.36%

In FL
1998 – 7.5%
1992 – 3.45%
1996 – 1.6%
2000 – 0.98%
2004 – 1.37%