P, what wasn't mentioned by the liberal blog MEM cited is that the Newsweek polling data was skewed.

Of the 1000 people polled, only 22% were registered Republicans, while 35% were registered democrats.

Most of the rest (37%) identified themselves as "independents." However, of the people who called themselves "independents," a whopping 52% considered themselves as democrat-leaning.

In other words, of the 1000 people polled more than half were democrats/liberals.

Of course that is going to affect the results.