Newsbusters' Tom Blumer sees some kind of liberal media conspiracy of silence in the lack of media coverage of a Gallup poll finding that more people self-identify as "conservative" than "liberal" at the state level as well as nationally. Blumer seems to think this finding has great significance, though Gallup provides no historical data for comparison, so we don't know which way things are trending.
And, as I've mentioned a time or two in the past, such labels are so imprecise and meaningless to many if not most Americans that these self-ID questions are of limited value. Indeed, the Gallup poll itself provides evidence that these questions don't mean much: Gallup finds that even in Massachusetts and Vermont more people self-identify as "conservative" than "liberal."
But Blumer thinks this one-off poll that is quite consistent with years and years worth of national-level polling is hugely important. Maybe that's because he doesn't really "get" how polling works. Here's Blumer:
The margins may not be "statistically significant," but the reported result still shows conservatives on top in HI (+5), VT (+1) and MA (+1). I also have to wonder how you can have a 5-point or more margin of error in a poll of 160,000 people. [Emphasis added]
Wonder no longer, Blumer:
Results are based on telephone interviews with 160,236 national adults, aged 18 and older, conducted Jan. 2-June 30, 2009, as part of Gallup Daily tracking. For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±1 percentage point.
The margin of sampling error for most states is ±3 percentage points, but is as high as ±7 percentage points for the District of Columbia, and ±6 percentage points for Wyoming, North Dakota, South Dakota, Delaware, and Hawaii.
That little bit of explanation was carefully hidden in the Gallup article Blumer linked to and quoted. On the first page. Under the heading "Survey Methods." A heading that was presented in bold font.