2004: +38.1* (R)
2008 Early Forecast: +39% (R)
2008: Strong Republican**
Trend Since 1976: none
Trend Since 1992: Republican***
Trend Since 2000: Democrat
Opponent Lead in Popular Vote to make state a tossup: 32%
Since 1976
Democrat wins: none
Republican wins: 1976, 1980, 1984, 1988, 1992, 1996, 2000, 2004
Positively correlated states****: Arkansas, Idaho, Kentucky, Minnesota, Missouri, Montana, South Dakota, Texas, Utah, West Virginia
Negatively correlated states: Connecticut, Delaware, DC, Florida, Illinois, Maryland, Massachusetts
*All numbers are 2-party margin versus national vote. For instance, Bush won Wyoming by 40.6%, which was 38.1% more than his national vote margin of 2.49%. Thus, 38.1 is listed here. In effect, all numbers assume that there is no margin between the Democrat and Republican Candidate in the national popular vote.
**
Strong Republican means that if the national popular vote is tied the Republican will win the state even if the forecast margin is off by more than 2 weighted standard deviations. "Lean" means greater than 1 but less than 2 standard deviations. "Tossup" means less than 1 standard deviation.
*** "Democrat" as a trend in a Republican state indicates an expected decline in support for the Republican candidate in that state from the 2004 result, calculated from the date indicated. Of course, in Wyoming, this doesn't matter.
**** If a state is positively correlated with another, then an increase for the Republican in one state will result in a similar increase for the Republican in the positively correlated state. Decreases are similarly correlated. If negatively correlated, than in increase for the Republican in one state tends to mean there's a corresponding increase for the Democrat in the negatively correlated state. So, if the GOP candidate is gaining ground in Wyoming, it is likely he's also gaining in Minnesota but losing ground in Florida.
EVEN MORE detail on what this all
means is here.