Internals of the McCain-Palin Bump

My NationalJournal.com column, on the "internals" of the McCain-Palin post-convention bounce, is now online.

In the column, I quote from a conversation with Andrew Kohut about his preference to avoid conducting Pew Research Center polls that measure vote preference during the convention period when there is "more fluidity in public opinion." On the New York Times website today, he fleshes out his own thoughts on the McCain bounce:

The question is, how good an indicator is [McCain's bounce] of where the electorate is headed on Nov. 4? A historical look at recent elections shows mixed results. In five of the seven elections since 1980, the candidate with the lead in early September went on to victory in November. In only two elections did the leading candidate go on to lose the election. But the record also suggests that when there was a change in momentum, in three cases that candidate won the election, and in two he was defeated.

Kohut also reviews these historical patterns in more detail and outlines a number of reasons why "this contest is a hard one to call early." It's well worth the click.

Charles Franklin and I put together a series of charts plotting the trends from the conventions to election day for every election since 1980 that illustrates the same trends that Kohut discusses.