Self Delusion among Liberals

I read a column by Ruy Teixeira in the New Republic in which he explains why president Obama's "victory" in 2012 will look a lot like the one in 2008. The column was an exercise in self-delusion on a scale I rarely see. Ruy takes apart a recent Pew poll which found Obama 8% ahead of Romney and explains that Obama's victory margin is coming from all the same voters as last time. But here's the rub: Ruy acknowledges that the average of recent polls shows Obama ahead by 4% (which includes this Pew poll with the 8% margin). That means that all the other polls are showing a very close race, not an 8% Obama victory. For anyone familiar with statistics, those numbers indicate that the Pew poll about which Ruy opines is an outlier, its results not actually indicative of reality. To make matters worse, the Pew poll shows that Obama is not performing as well as last time among minorities; instead, Obama has increased his lead over his 2009 victory margin by picking up a bigger percentage of whites. Really? Poll after poll has shown that Obama has lost significant support in the white community since 2008. Ruy, however, is telling us about how Obama has actually picked up support there.

I cannot tell which Pew poll is being discussed in the column. My guess, however, is that it is a poll in which the partisan mix of respondents oversamples Democrats and undersamples Republicans.