Crikey psephologist Charles Richardson writes:

Watching American election results remains a frustrating experience. Big blocks of electoral votes just appear out of nowhere: there is no concept of "swing", no-one ever talks about the pendulum, no-one matches precincts to get comparative results.

It's like Australia of forty years ago: not just pre-Antony Green, but pre-Malcolm Mackerras. So what follows is an attempt to explain what happened in psephological terms that Australians will understand.

In 2000, George Bush Jr received 49.7% of the two-party vote. This year, according to the latest figures, he has 51.6%, a swing of 1.9%.

It is not surprising, therefore, that Bush won a majority in the electoral college. It is smaller than you would expect, though, because so much of his swing was concentrated in his own safe states. A uniform 1.9% swing would have gained him the states of Florida, New Mexico, Iowa, Wisconsin, Oregon and Minnesota, plus one electoral vote in Maine, for a gain of 67 electoral votes.

As it was, however, he only gained the first three of those (the Democrats have not yet conceded Iowa, but Bush is clearly ahead), and lost one of his own states, New Hampshire, for a net gain of only 35 electoral votes. (As explained in my election preview, I count Florida as starting in the Democrat column.)

To win in 2008, the Democrats will need a uniform swing of 1.3% - the median state of course being Ohio, with its 20 electoral votes. In other words, they could win the electoral college without a majority of the vote, because their vote is a little more evenly spread than the Republicans'.

The Republicans won 31 states, 15 of them with margins of more than 10%; the safest was Utah on 22.9%. All but one (Indiana) of the 15 are in the deep south or the west.

John Kerry won 19 states plus the District of Columbia, but only three states by more than 10%; all three are in New England, the safest being his own state, Massachusetts, on 12.9% (the District of Columbia was 40.6%).

It was very much a status quo election: the majority of states swung by less than 2% either way. On current figures, the biggest swing was 5.9%, to the Republicans in Hawaii; the biggest the other way was 4.9% in Vermont.

To the extent that it is moving at all, the American electorate seems to be getting more geographically polarised. Prior to the election, there were 16 states held with margins of less than 3%. Now there are only 11.

The pattern of party support is clear from a glance at a map: Democrats have the north-east and the west coast, the mid-west is marginal, and the Republicans have everything else. It is the Civil War division repeating itself, but with the party labels reversed.

Just as interesting, however, was the pattern of swing this time, which I have not seen mentioned by any other commentator. Although nowhere swung very much, the east swung generally to the Republicans, and the west to the Democrats. Of the 17 states that swung to Kerry, 11 of them were west of the Mississippi, while of the 14 biggest swings to Bush, all but two (Hawaii and Oklahoma) were to the east.