Election commentary had a tendency to get into some very fine-grained state-by-state analysis about what does and doesn’t appeal to voters in Pennsylvania or Colorado or the I-4 corridor in Florida and so forth. The evidence, however, was of a pretty boring more-or-less uniform national swing:
My inclination had been to say that press over-emphasis of state-specific factors was probably a holdover from the unusually close 2000 election when the details of the electoral college turned out to really matter. But Andrew Gelman took a more systematic look at the issue and finds that the uniformity of presidential election swings has been experiencing a steady (especially if you include the one-off case of the southern swing to Carter in 1976), decades-long increase:
Plausibly, then, the anticipation of state swings being substantially independent of one another simply reflects the actual experience of veteran campaign operatives and campaign reporters. The swings actually were less uniform in the past. And conventional wisdom among the younger generation of journalists and operatives is still dominated by the lessons taught by their elders rather than by the recent spate of low-variation elections.