Why bother?

Today, many are reporting on a new NYT/CBS poll that claims massive support for the President and the direction of the country. Taegan Goddard from Congressional Quarterly summarized it this way:

Polls Finds More Optimism
Americans have grown more optimistic about the economy and the direction of the country in the 11 weeks since President Obama was inaugurated, according to the latest New York Times/CBS News poll.

"Fully two-thirds said they approved of his overall job performance. By contrast, just 31 percent of respondents said they had a favorable view of the Republican Party, the lowest in the 25 years the question has been asked" in the poll.

But what good is a poll if the sample is skewed? To me, the two most interesting questions of this poll had nothing to do with approval rating or direction of the country. Instead, it was the questions asking 'who did you vote for?'(43 Obama-25 McCain) and 'what Party do you affiliate yourself with?' (39D-23R-30I).

While I can understand Party affiliation fluctuating over time, this poll seems to go to the extreme in their sample. I accept that there are fewer Republicans than Democrats, but 16 percent fewer than the D's and less than 1/4 of the population?

What is particularly telling is the 'who did you vote for' number. As we all know, the election was much closer than the 43-25 differential that this poll provided. There are few numbers in politics that are stable, but the ultimate poll of the nation comes every four years. If you want to properly gauge Americans, wouldn't it be more effective to fit your sample around how they actually aligned themselves?

With such skewing, why should we even bother paying attention to these numbers?