NPR Poll Offers Sliver Of Hope For Democrats

NPR Poll Offers Sliver Of Hope For Democrats
by Mara Liasson

October 15, 2010

With only 18 days left until Election Day, an NPR survey of likely voters in battleground districts found that while the overall field still tilts to the GOP, Democrats are closing the gap in some places.

The good news for Democrats is that they have improved their position in 58 of the battleground districts. In June, Democratic candidates trailed their Republican opponents in these districts by 8 points. Now, the GOP advantage is only 3 points.

A Small Shift
If the election for Congress were being held today, for whom would you vote:



But, says Republican pollster Glenn Bolger, the bad news for Democrats is that there are now more of their seats at risk. The battleground has expanded.

"That's a succinct way to put it," Bolger said. "Having Republicans ahead or tied across 86 districts just gives Republicans a much larger margin for error to win back the House."

'Small Movements Matter'

Seventy swing seats were examined in June. Now, 96 are up for grabs — 86 of them held by Democrats.

Democratic pollster Stan Greenberg sees a small, but silver lining.

"This is still an election that it would be very hard for Democrats to hold the House with these numbers," Greenberg said. "But there is movement here and it's not trivial movement.

<SNIP>

http://www.npr.org/templates/story/story.php?storyId=130576555

That really messes with the validity of any conclusions... but the summary is unclear.

That statistic is problematic to begin with as it doesn't take too many districts going to one side or the other by, say, 10 points to completely disguise what is happening in 10 races that are really within a few points.

The fact that more races are designated as toss ups is actually bad, not good - as they were assumed ours in the earlier race.

If you take two or three dozen democratic districts and added them to the original poll in June... I'm not sure whether the results would have been better than this or worse. You can't even really tell that the playing field has "expanded" from this (though I suspect that most think that it has).

I may have missed it, but you would expect that they would have computed the result for the subset used earlier. They have the data - and it's a pretty obvious thing to do. I wish they would have computed it.

One other problem is that - unlike many other polls - they sampled just land line phones.

'For this latest survey, the polling firms interviewed 1,200 likely voters in the battleground districts over the days Oct. 7-10. The survey was done of "land-line" telephone users only. Calls were not made to cell phone users. Public Opinion Strategies' Bolger says that could mean some younger, possibly Democratic-leaning voters were missed and that the GOP advantage is slightly narrower than the results show. For results based on responses from all 1,200, the margin of error is +/- 2.83 percentage points. Results on subgroups are higher."

They do say that in the 10 battleground GOP districts, the R margin has gone from 16 to 7.