Beware of assumptions

Soren Dayton offers an enthusiastic assessment of how Barack Obama could put the Rust Belt and Northeastern blue-collar white voters "in play" for Republicans in November. Some obvious problems:

• A primary election is not a general election. If it is safe to assume that Democratic voters who didn't vote for Obama in the primaries would be "in play" for Republicans in November, why not assume that Republican voters who didn't vote for McCain in the primaries would be "in play" for Democrats in November? Partisan loyalties tend to be relatively stable, and primary voters tend to be highly partisan.(*) People who vote in a Democratic primary in April are unlikely to vote Republican in November.
• Who says Obama will be the nominee? I mean, besides Obamaphiles, stressed-out Democratic officials, and DC press-corps types trying to "cover their bets" after going out on a limb to declare the nomination a lock for Obama? If every Republican strategist wants to forget about Hillary and spend the next three months preparing for a fall campaign as if Obama were certain to win the nomination, OK. But it's not a certainty.
• Events happen. Exit polls, focus groups and demographic analysis have convinced a lot of people that politics is all about "trends." But trends are after-the-fact intellectual reifications of political behavior. Political trends do not exist independent of political events. A gaffe, a blunder, a scandal, an economic turn, a military action -- these things affect elections as much as any demographic trend.
• Don't bet your hopes. The fact that you greatly desire an outcome doesn't mean that outcome is likely, or even realistically possible. I'm sure Republicans would like to believe that a McCain-Obama match-up would produce a Nixon-McGovern result -- e.g., a GOP landslide with lots of "hard hat" voters turned off by the apparent radicalism of the Democrat. Such hopes, however, cannot justify the huge gamble of investing major campaign resources in a state like Pennsylvania, that hasn't gone Republican since 1988.

Politics is not a science that can be plotted out in advance without taking into account surprises and contingencies. The slide-rule stuff is all very interesting, from an analytical perspective, but it doesn't win elections.

A gut hunch tells me that John McCain can't possibly win more electoral votes in November than Bush won in 2004, and that Pennsylvania -- which viciously dumped Santorum in 2006 -- will remain firmly Democratic.

But who knows? Anything can happen. I'll be content to wait until after Election Day to learn if my gut hunch was more reliable than Soren's micro-demographic analysis.

* I note that my statement here about the general stabilty of partisan loyalty and primary voters as highly partisan might seem to be contradicted by my own glee over the large number of party-switchers (160,000+ Republicans re-registered as Democrats this year) in Pennsylvania. But there is no contradiction at all. If anything, Rush Limbaugh's "Operation Chaos" shows the strength of partisan loyalty, in that Republicans were willing to re-register merely to wreak havoc among Democrats. Thus, demographic mapping showing a white-voter trend toward Hillary in the primary doesn't really suggest any extra "in play" factor in Pennsylvania, since many of those white voters were actually Republican crossover "mischief" voters. They were already Republican to begin with, and thus do not represent some previously untapped reservoir of new potential GOP voters.