Gallup’s newest daily national
poll of likely voters’ presidential preferences show 50 percent
for Romney, 46 percent for Obama.
The numbers represent Romney’s highest support, and Obama’s
lowest support, in recent days (or recent months, and probably
ever, though it’s hard to tell given that Gallup recently changed
from a registered voter model to a likely voter model.)
Given prior results of this survey, including before the model
change during which time Romney was never better than even with
Obama, and that only once, today’s results strike me as dramatic
and significant.
Nevertheless, when Gallup’s alert e-mail arrived in my account,
the text of the note begins: “The race for president is about
tied among likely voters.” Apparently, the e-mail writer at
Gallup is having his or her hopes for Obama’s re-election dashed
and can’t bear to type the words.
Such not-very-subtle bias should be spurned by Gallup if they
want to maintain a reputation of non-partisanship. (Personally, I
have seen too much of this from Gallup to believe that they have no
bias, but others whom I respect think Gallup is basically fair. And
I certainly don’t think they try to influence the actual outcome of
the polls with bias; it’s more about how they report the
results.)


Butch| 10.16.12 @ 4:25PM
Was the plus-four within the margin of error? If so, the statement is technically justified, but I'll bet it would have been written differently if the results went the other way.
Ross Kaminsky| 10.16.12 @ 5:53PM
Butch: Gallup reports margin of error on this poll at 3 percent, so the results were outside. But your second sentence is more along the lines of the point I was trying to make.
spike59| 10.17.12 @ 5:59AM
check the link, and you'll find that, compared to 2008's numbers, ObaMao has lost ground with every demographic listed-EVERY SINGLE ONE.
we're looking at a 'Dewey beats Truman' election...and the MSM will have no answer as to how they could have missed it, much like they missed the mark on the 'even to a blind man could see it coming' 2010 mid-term results
Occam's Tool| 10.16.12 @ 4:58PM
Tonight, I see the Brockton Blockbuster starting slowly, like he usually does. But don't worry; no doubt percussion will accentuate exhaustion and exhaustion will accentuate percussion, and Obama will be blasted off his feet yet again by the Sultan of Salt Lake, the Massachusetts Mauler himself, Mitt Romney!
Paul McGrath| 10.16.12 @ 7:02PM
And in this corner . . . the Kenyan Incompetent, the Hawaiian Has-Been, the Chicago Chump, the Malaysian Mama's Boy . . .
Occam's Tool| 10.16.12 @ 10:14PM
As indeed happened. Obama was reaching for Big Bird to save him, no sh*&%t.
Love your stuff, Paul.
Teflon93 | 10.17.12 @ 7:56AM
It's even WORSE than Kaminsky puts it: pollsters have largely been in the tank for Obama and outright manipulating their data to get him reelected.
The change to likely voters in October is down to pollsters getting graded on their final predictions. They know damned well that registered voter polling massively overstates Democrat performance; they use that for most of the campaign anyway to generate pro-Democrat headlines on behalf of the Left Wing Media clients who pay them. Even when they switch to likely voters, they keep one finger on the scale. You can see this by checking out RealClearPolitics' data on their accuracy with their final predictions: if the error were random, you'd see an equal number of pollsters overstating and understating Democrat performance and by an equal magnitude. Instead you see over 3/4 of pollsters in 2008 overstating Barack Obama's performance, including Zogby overstating it by a full 4.5 pct (D+11 vs D+6.5). But get this: throughout the late election cycle, some pollsters have plugged Zogby's number D+11 into their polling!
This has been a farce. It will not be whisked away under the rug by the usual "the race is tightening" b.s.