You guys bring up good points, but increased polarization has likely increased the predictive value of early polls. I doubt we'd see any crazy swings like this in the modern era, barring a major game changer such as a scandal or economic collapse.
For example, in Romney vs. Obama numbers, there was nothing TOO crazy and wildly different from the end result in there, except during Obama's peak of popularity. Especially if there was enough polls to create an average, which there currently is for Hillary vs. the GOPers.
Good point. Here are some average leads Obama had on Romney according to the RealClearPolitics poll averages at different stages/days during the two years prior to the 2012 election (21 last months to be exact, as RCP started tracking the race on February 2, 2011):
February 9, 2011: Obama +4%
April 17, 2011: Obama +4%
April 19, 2011: Obama +3.6%
May 6, 2011: Obama +4%
August 10, 2011: Obama +3.6%
September 24, 2011: Obama +3%
February 6, 2012: Obama +4%
March 14, 2012: Obama +3.9%
April 5, 2012: Obama +4.1%
April 26, 2012: Obama +3.7%
June 6, 2012: Obama +3%
June 29, 2012: Obama +3.8%
August 8, 2012: Obama +3.9%
August 17, 2012: Obama +3.9%
September 11, 2012: Obama +3.6%
September 25, 2012: Obama +4%
October 1, 2012: Obama +4%
In other words, during these 20 months above, one can hardly argue that the Obama-Romney polls changed much. Romney basically needed an earthquake at least of some proportions in order to change the narrative. And his 47% comment coupled with the hurricane Sandy arguably weren't the type of earthquakes that Romney was searching for.
In fact, on the days of Obama's strongest leads, June 2-6, 2011, Obama was still only +7.3% ahead of Romney. In other words, one can not say that polls changed very much over time, at all. During those two years, Romney was just ahead of Obama in the polls during four extremely short periods (basically a week each time), namely during:
*September 6-13, 2011 (8 days)
*October 6-11, 2011 (6 days)
*October 9-17, 2012 (9 days)
*October 22-30, 2012 (9 days)
In other words, over the lapse of the last
642 days of the campaign, Romney was just leading Obama in the aggregate of polls on
32 days, that is
just 5% of the time.
