As Pinkerton Was to McClellan, part 2

Please read the first part of this below for context.

I was having breakfast with a friend and fellow political activist on Saturday. He was expressing his confidence based on the poll averages at Real Clear Politics. I like what they've done, but when it comes to the polls I must say that the average of 8 kinds of excrement is still excrement. No matter how many you average together, all of those polls represent carefully massaged data, intended to please the people who buy the product: the news show biz.

Exhibit A in this argument is in the poll internals. While the pollsters tell us that overall the electorate is evenly divided, in moments of candor they will admit that the among women, the candidates poll about even. This must be more or less consistent through all the polls, because the news show biz has made no effort to market the "gender gap" story.

The "gender gap" is something they marketed vigorously in every election since at least 1980. They used to take glee in beating up Ronald Reagan over how his opponents polled roughly 10-15 points better among women than he did. We heard this also used as a cudgel with which to beat up W when the final polls had Gore 10 points ahead among women.

This year, there are no "gender gap" stories in the news show biz, and the polls that have gone on record show the women's vote evenly divided. So how is it, that the overall electorate is evenly divided if half the electorate has moved 10 points toward W? Shouldn't that mean that W has approximately a 5 point structural advantage over his position 4 years ago?

Well, if you're a pollster you can "adjust" your model of "likely voters" to sample another 5 percentage points of Kerry supporters, and voilà! The electorate is evenly divided. That is, after all, what your customers want to hear...

...and the customer is always right, n'est-ce pas?