Help me Understand the Latest NY Times / CBS Poll

So the latest NY Times/ CBS, poll, available here, claims the following:

1) Obama’s overall approval is 56% with only 33% disapproving (actually down 2% from last month).

2) Overall, 30% “mostly approve” of the President’s health care plan with only 23% “mostly disapprove.” (Among independents the numbers are 22% support 25% opposed)

Here is what I don’t understand.  According the best information I can find, there are approximately 72 million registered Democrats and 55 million registered Republicans, a 1.31 to 1 ratio.  In their poll, 357 respondents identified themselves as Democrats and 289 as republicans, a 1.24 to 1 ratio – not too far off.

However, they decided to “weight” the poll to 37% Democrat and only 22 % Republican! That’s a 1.68 to 1 ratio of Democrats to Republicans.

That would be the same as assuming there are more than 92 million registered Democrats instead of 72 million! So when you don’t get the results you want, just add 20 million Democrats to the mix?

Then they use that mix to post answers to questions such as “who has better ideas for health care reform?” and “who is trying harder for bipartisanship on health care?”

Assuming that my data about party registration is correct, how can a national poll get away with biasing itself so blatantly? It really bugs me that these numbers will be presented to the American people as a true sample of what all American adults think.