>>>>http://www.talkingpointsmemo.com/news/2009/04/ap_poll_americans_high_on_obama_direction_of_us.php
>>>
>>>How to produce high approval ratings for Obama?
>>>http://www.americanthinker.com/blog/2009/04/how_to_produce_high_approval_r.html
>>>http://www.ap-gfkpoll.com/pdf/AP-GfK_Poll_Topline_April_2009.pdf
>>
>>I read over the poll information and the American Thinker rebuttal. The numbers the American Thinker references for 2009 seem to be correct, but the breakdowns appear [to me] to be reasonably similar over the periods noted starting in Nov 2008, so I can't figure out where the information for their next paragraph comes from. Am I missing something (probable)?
>>
>>In the 2008 election the spread between Democrats and Republicans was 6.5 percentage points not 18 and independents made up 22% of the vote not 26%.
>
>They're referring to the difference between the percentage breakdowns of actual voters during the last election verses the breakdowns used in the poll.

Ah. I get it. I thought they were using the published numbers in the poll.

>
>Mind you, polling breakdown issues are not a new development. There are a number of issues which make it difficult for polling companies to achieve proper diversity. This is why the breakdowns are published and must be accounted for when examining the results.

Although, if the poll results are correct, then it's worth noting that the poll breakdown numbers are not much changed over the period. That suggests that republicans (at least those polled) are feeling better about the direction of the country too.