The Latest Poll

In this day of non-stop and up-to-the-minute polling, the latest Presidential-election polls have raised a few questions, in my mind at least.  Let's start with a couple of apparently innocent "facts:"

1) The pollsters keep telling us that there are very few, in fact almost no, "undecided" voters left.
2) The polls keep shifting, almost on a daily basis.

Last week, for example, Obama was said to be moving into a decisive lead in the polls, both nationally and in most of the "battleground" states, the only states that matter in this election, thanks to the Electoral College.

This week, however, after a debate in which Mitt Romney was universally declared the winner, even though, in my humble and apparently irrelevant opinion, he lost badly (see related post) - the polls have shifted dramatically in Romney's favor, to the point where he is, depending on which poll you believe,  currently tied with Obama, or even in the lead.

Here's the question:
Since there are so few undecided voters out there, how exactly can the polls shift so dramatically from day to day?  Since presumably different people are being polled each time, it hardly seems credible that the minimal number of who supposedly still haven't made up their mind are present in sufficient numbers to swing the entire poll results.  It made absolutely no sense until it dawned on me:
The pollsters must only be recording results from self-proclaimed undecided voters.  Here's how I imagine a typical conversation between a pollster and a voter:

POLLSTER: Hello, have you made up your mind yet who you are going to vote for?
VOTER:  Yes, I -
POLLSTER: Thank you, have a nice day.
-Click-

POLLSTER: Hello, have you made up your mind yet who you are going to vote for?
VOTER:  Well, I just can't make up my mind...
POLLSTER: If the election were today,...

etc, etc.  it's the only explanation that can explain the "facts." Clearly pollsters have no interest in the opinion of already-decided voters, since their votes cannot change and therefore cannot provide the endless intrigue of a "neck and neck" election, which sells more papers and magazines, keeps people logged onto websites, and so on.  This could also explain why the media had decided long in advance that Romney would be declared the winner of the first debate (again, see other post).

Of course, this analysis still leaves open the question of which polls, if any, can be believed, and who can be expected to win this election.  Will it be close enough that Republican voter fraud and malfeasance will be able to decide the outcome, as in Ohio in 2004?  Or will it be another Obama landslide, like in 2008?  Stay tuned.