That's what I learned to use. Pollsters err when they ask party affiliation as opposed to whether a voter self identifies as liberal or conservative. Actually they ask both in exit polls but too many pre-election polls ask merely party affiliation.
For example, in 2000 and 2004 Democrats won every state with at least 25% self-identified liberals according to the exit poll, and Republicans won every state with at least 37% self-identified conservatives. No state had both at the same time. So that simple method went 100-0 combined.
I mention those cycles because they were very tight in the national vote. That allowed the self-identified numbers to find their natural spot. It can vary somewhat when there is a tilt toward one side or another. After more than a decade of winning every presidential statewide bet I made -- using those self-identified numbers -- I lost two statewide wagers in 2008 because Obama's national advantage narrowly swayed two states into his column that have a high number of self-identified conservatives -- North Carolina and Indiana. I think I lost one by a tenth of a point and the other by three tenths. In 2012 both returned to their normal setting.
Some states can change as the population shifts. Virginia now has far fewer self-identified conservatives than when I started charting this. It was in the 39% range then dropped all the way to 31% in the 2012 exit poll. Likewise New Mexico and Colorado. However, states like that should never be taken for granted because the number of self-identified conservatives remains high. They hardly have the classic DNA of states Democrats figure to carry regardless of national margin. New Mexico still had 37% self-identified conservatives in the 2012 exit poll. We'd be moronic to ignore that state, despite the Hispanic numbers. Nevada remains in the 35-37% conservative range. It hits the higher number in midterms and the lower rate in presidential years. And that's all the difference, the makeup of the electorate.
I always have to laugh when a knucklehead like Rachel Maddow touts pickup opportunity in some state like South Carolina. She is good for that type of ignorance at least 2 or 3 times every cycle, based on some irrelevant poll during a Democratic uptick during the race. South Carolina always reports a range of 17% liberals and 40% conservatives, along with a very low percentage of Hispanics. There simply isn't logic toward a reversal, unless the national margin is overwhelming.
It will be interesting to see what the national self-identified percentages are this year. In 2012 it rose to 25% liberals according to the national exit poll. That was fantastic. Perhaps not sustainable. Conservatives were 35%.
= new reply since forum marked as read
