Practicalmadness

Why We Cannot Endorse President Bush For Re-Election: From The Tampa Tribune › The editors of a Florida newspaper that has endorsed the Republican candidate in every election since Eisenhower in ‘52 refuses to endorse either Bush or Kerry. They’re skeptical of Kerry’s ability to do the job, but disappointed by Bush’s utter failure to live up to his campaign promises, not to mention his administration’s botching of the war in Iraq.

Using the most current and empirically reliable polling data, Slate has started predicting which way the electoral votes will fall and there’s good news for those of us on the left: if the election were held today, Kerry wins with 276 predicted electoral votes to Bush’s 262. Not a landslide, but a far more decisive (possible) result than we had in 2000.

This is encouraging in part because in making this tally, Slate erred on the side of precedent and gave Colorado and Florida’s votes — expected to be the two most hotly-contested races — to Bush, and assumed that all of Colorado’s electorals are going to just one candidate. (Coloradoans are voting on a proposition to switch from a “winner-take-all” system to a “proportional” one simultaneously with the vote for a new President, so the assumption here is that voters there will reject the new measure and keep the status quo.)

This is smart because while nationwide poll numbers (e.g., the ones that have either candidate ahead by 1-2% with a 2-3% margin of error) make great headlines, they don’t at all reflect the way a President is actually elected in this country. So by breaking it down by state, simulating the way the election will actually go down, Slate may have found the most effective indicator yet.

And it’s encouraging because it takes Florida out of the equation and yet still manages to declare Kerry the winner.

I somewhat rarely agree with the weblog DonkeyRising, written by Democratic analyst Ruy Texieira (he’s hypothetical and bullish about the Democrats’ chances often past the point of reason), but today he has an analysis of the Gallup poll’s flawed methodology that’s about as encouraging as Slate’s electoral vote tally. Basically he argues that Gallup’s profile of “likely voters” mistakenly washes out minorities and young people to an extent that isn’t supported by historical data from, you know, actual elections.

So if the Gallup poll has Bush with an 8% lead over Kerry (with the typical ~2% margin for error) but fails to represent blacks, Hispanics, young voters and the poor (not specifically mentioned by Texieira, but also historically underrepresented in polling) and those groups get the vote out at levels equal to or even slightly lower than in 2000, that 8-point advantage deteriorates quickly.