Was the financial crisis an important factor in McCain's trouble with voters?:
Crisis? What Crisis?, by Phil Klinkner: Lots of chatter, especially from the GOP that the September financial crisis did McCain’s chances. I’m skeptical. Look at this graph from Pollster.com. It shows the poll average trendline from August 22 (the day Obama announced his pick of Joe Biden for VP) to the election. I’ve set the sensitivity to high so that it will highlight even small day-to-day changes in the polls.
Obama got a bit of a bounce out of his convention, but that receded pretty quickly in the face of the Palin announcement and then McCain’s convention bounce. But the graph shows that McCain topped out at about September 7 and that Obama began to bounce back two days later on September 9. By September 13, the day Lehman declared banruptcy and the crisis began to develop, Obama was already back in the lead.
If the crisis clearly moved voters in Obama’s direction, you would expect to see a sharp break in the trendlines sometime after September 13. But you don’t. The lines pretty much stay on course from early September until early October, when McCain bottoms out and starts to come back a bit. The timing of Obama’s rise in the polls seems to suggest that it resulted from the fading of the GOP convention bounce, not the financial crisis. In fact, the graph above tracks almost exactly with the predicted impact of the convention bounces predicted by 538’s Nate Silver ...
There isn't a sharp break in the polls around the crisis date, but it's still possible that the crisis - and how each candidate reacted to it - mattered in terms of helping to shape voters' general perception of each candidate. That is, voters make decisions at different times, some decide early, others decide much later. One possibility is that the crisis causes many of the people who have already decided to change their minds, and that it also causes many of the people still thinking it over to suddenly support one of the candidates. In this case, you would expect to see an abrupt change in the polling data. But another possibility is that as the many people who haven't been paying much attention begin to tune in and decide how to vote - and many people wait until the election is fairly close to do this - the perception of how the two candidates handled the financial crisis could help to sway the decision of who to support one way or the other. If so, then you would see a slower, more persistent separation in the polls rather than an abrupt change right at the time of the crisis.