Big Lizards:Blog:Entry “Polling Keeps a-Leaping”

« Peaceful, Easy Baghdad Feeling? | Main | Embryonic Stem Cells: Static Analysis Strikes Out Again »

August 22, 2006

Polling Keeps a-Leaping

Hatched by Dafydd

Commenter Terrye tipped us to an incredible new poll out from Gallup, written up by their partner USA Today (Terrye hat-tips "AJStrata" at the Strata-Sphere).

As you know -- or as you should know, if you've been conscientiously reading Big Lizards -- the so-called "generic congressional poll" has been running pretty grim for Republicans lately. But in a Gallup poll taken from August 18th through the 20th, 2006, both the generic congressional poll and also Bush's job-approval have taken a sharp turn upward:

In a USA TODAY/Gallup Poll taken Friday through Sunday, support for an unnamed Democratic congressional candidate over a Republican one narrowed to 2 percentage points, 47%-45%, among registered voters. Over the past year, Democrats have led by wider margins that ranged up to 16 points.

Now 42% of Americans say they approve of the job Bush is doing as president, up 5 points since early this month. His approval rating on handling terrorism is 55%, the highest in more than a year.

Note that according to Polling Report, among "regular voters" (which I reckon means respondents who vote regularly), the generic congressional question was a dead-even tie, 48 to 48. That is superb! Of course, USA Today found no occasion to mention that datum.

So how does Gallup and USA Today explain the sudden jump? They see it as the reaction by the American people to the exposure of the London airplane bombing plot:

The arrest of terror suspects in London has helped buoy President Bush to his highest approval rating in six months and dampen Democratic congressional prospects to their lowest in a year....

The boost may prove to be temporary, but it was evidence of the continuing political power of terrorism.

“The arrests reminded people that terrorists were out there, and this is his strong suit,” says political scientist Gary Jacobson of the University of California, San Diego. Now, as in 2002 and 2004, Bush and GOP congressional candidates argue that they can be better trusted to combat terrorism.

The alleged plot to bomb flights to the USA “also changes the subject of public discussion from the war in Iraq, which people are not very happy about,” says Christopher Gelpi, a political scientist at Duke University.

Interestingly, Mort Kondrake was crowing about (and Fred Barnes was lamenting) the fact that the shattered terror plot was not having any effect on the electorate; Mort said something about how terrorism was clearly no longer the dispositive factor in national elections (he didn't use the word "dispositive"). Well, the crower can eat crow on Saturday.

But now, some obligatory words of warning; being a math guy, I can hardly stop myself:

Gallup has a history of not weighting its sample for party affiliation; while ordinarily, this results in too many Democrats (which is controlled by where and when polls are conducted), occasionally they manage to oversample Republicans or Independents. If they did so this time, then of course the poll would show much better numbers for Republicans than would actually be the case in a perfectly sampled poll.

Even if the poll were conducted flawlessly, there is always the chance that the particular batch of people polled were less representative than usual.

Polls are typically reported with what is called a "margin of error" attached, and this one is no exception: "the telephone survey of 1,001 adults has a margin of error of +/-3 percentage points." That means is that there is a 95% chance that this poll accurately reflects the current opinion of American adults within a -3% to +3% range.

Looking only at registered voters would increase that range somewhat, perhaps to ±3.5%. But even so, the important number is the 95%: by definition, 5% of well-conducted polls will nevertheless be outside that ±3.5% range; such polls are called outliers, and they are wild cards that cannot be predicted nor prevented.

The only way to tell whether this is an outlier, and if so by how much, is to watch other polling on the generic congressional question. Alas, no other pollster has polled over a similar range; the only one polling in this period of time, CNN, last polled in June (they saw a 6% gap in June and a 9% gap now)... so there is no comparable case.

PRE-POSTING UPDATE: a late-breaking CBS-New York Times poll, reported in Polling Report but nowhere else as of this moment, shows the Democratic advantage much wider (15%) and climbing. But they also include a choice of "Depends," in addition to Republican, Democratic, Other (which would scoop up the Green Party, the Libertarian Party, the Beer-Drinker's Party, and so on), and Unsure; Depends sucks up 12% of the response, and we have no clue how to allocate that.

As far as I'm concerned, from a polling perspective, the extra weasel-option invalidates the entire poll. No other pollster adds "Depends" as a choice, so you cannot compare the CBS/NYT poll to any other. (I wouldn't even include it on the Real Clear Politics average, though I suppose they probably will.)

But I'm somewhat cheered by the concomitant increase found in President Bush's job-approval numbers on the new Gallup poll -- from 40% last time to 42% this time -- because that is similar to the other two polls conducted at the same time, which showed similar increases:

CNN went from 40% at the beginning of the month up to 42% now;

Rasmussen went from 40% to 43%;

This tells me that they probably didn't have a "runaway population sample;" they're in line with other sampled increases.

In any event, the current Real Clear Politics average for the generic is 8.0% in favor of the Democrats; without the new Gallup poll, it would be 10.5%. But the average for the entire month of August, not including the most recent Gallup or CNN numbers, is 51.0% Democrat, 37.8 Republican, for a gap of 13.2% in favor of the Democrats.

This represents a huge leap forward by the Republicans, an improvement of 39% over the early part of the month. If this polling turns out not to be an outlier, then the trendline is moving towards completely erasing most of the electoral gains the Democrats anticipate for this fall. But there is, naturally, a third caveat:

As they say in the financial biz, past performance is no guarantee of future results. It's still possible that something could happen to dramatically boost the Democrats' chances in November, which could throw all these careful weighing of averages into a crockpot.

Of course, it's more likely that things will continue happening to help the Republicans, as with the London terrorist plot. If we can believe Gallup's own explanation for the strong movement towards the GOP -- a response by the American people towards a reminder of the terrible danger we face from terrorist attack -- then that itself is likely to happen several more times between now and the election.

Too, per an earlier post, the situation in Iraq is likely to improve between now and election day, and we might even begin bringing troops home. Since Iraq is the biggest drag on both Bush's job approval and the generic congresssional numbers, an improvement in voters' perception of how the Iraq war is going can have a stunning impact on how they vote on November 7th.

It's still anybody's game; but prospects are definitely looking up for the Republicans now.

Hatched by Dafydd on this day, August 22, 2006, at the time of 6:30 PM

Trackback Pings

TrackBack URL for this hissing: http://biglizards.net/mt3.36/earendiltrack.cgi/1133

Comments

The following hissed in response by: Baggi

It all comes down to Immigration, which perhaps the left hasn't realized yet.

For months the Republicans and the Bush administration had the left and the right beating up on them. Just take a look at how the rhetoric has changed by the polipundit kidz website. It was all immigration all the time for awhile there was a big shakeup because 4 out of the 5 posters there weren't anti-Bush due to immigration issues.

So the ever-mature Polipundit kicked them all from his website (He's got every right to do that) and thus Wizbang Politics was born.

My point being is that it has taken months for things to simmer down. If they re-ignite prior to the election, the Republicans will go down in flames. But if they can keep Immigration off the news for the next few months, things will go swimingly for the Republicans.

So, keep your eyes open. If the media starts to focus on immigration again prior to elections and the Republicans don't appease the far right wing radicals on the issue, we'll lose the senate and the house and the governorships.

The above hissed in response by: Baggi at August 22, 2006 7:23 PM

The following hissed in response by: Mr. Michael

Too, per an earlier post, the situation in Iraq is likely to improve between now and election day, and we might even begin bringing troops home.

Of course, today they announced that we are calling up more Marines for Iraq and Afghanistan... it's just 'way too early to really spot any trends there. On the one hand, if the situation in Iraq needs more troops, the danger over there will help focus the public on the threat, which generally would help the Republicans... but then the need for more troops would hurt them.

I'm more focused on how the Politicians are presenting themselves and their core values (if any) than I am trying to determine the impact of world events. I acknoledge their impact, but I don't see where we can use them as predictors. They work better as excuses.

The above hissed in response by: Mr. Michael at August 22, 2006 9:52 PM

The following hissed in response by: Dan Kauffman

Summary of Party Affiliation
Rasmussen Reports Tracking Surveys

Approximately 15,000 Interviews per Month

The OTHER category becomes more and more important and Republicans may do better there in the Center than Democrats who seem determined to purge their ranks of the politically impure.

The above hissed in response by: Dan Kauffman at August 23, 2006 12:18 AM

The following hissed in response by: Don

I had a look at the RCP site for generic congressional polls. The Gallup poll does appear to be an outlier.

We'll see. As Jay Cost keeps reminding us - the generic poll doesn't matter that much because people don't vote for or against a generic congressman. They vote for or against a candidate, typically an incumbent.

There may be reason for optimism for all of the reasons you mention. I think Bush has troughed and is on the way up; how far remains to be seen. I doubt Ned Lamont is helping the Peace Democrats nationwide; I can't see a richer than God peacenik playing in the South - nor will the nutroots idiots play there. People like Leiberman; they don't like the people who bounced him much; so the more play that race gets nationwide the worse for Democatic hopes.

The above hissed in response by: Don at August 23, 2006 5:49 AM

The following hissed in response by: Dan Kauffman

LOL does this strike a chord

The Democratic Party seems to be bound and determined to run off the edge of a cliff.
Howard Dean, Moveon.org and other assorted irrational Bush haters, rather than trying to avoid this,seem to be petulantly complaining they are not stampeding fast enough.

The Lemming Left

The above hissed in response by: Dan Kauffman at August 23, 2006 6:14 AM

The following hissed in response by: MTF

Can there be a good argument the upcoming midterm elections might just as easily result in Republican gains in Congress as losses? I wonder.

I haven't seen the Democrats achieve traction on any single issue. Remember the Democrats “Plan for Real National Security”? It’s gone. When terror events happen, major party figures all flock to support the President, driving the Daily Kos crowd absolutely crazy, and reminding voters that even the Democrats believe Bush is better on the GWOT than they are.

The above hissed in response by: MTF at August 23, 2006 10:22 AM

The following hissed in response by: Terrye

They may be calling up more Marines but overall the recruitment quptas have been met. In fact I saw a headline the other day to the effect that in spite of violence in Iraq etc the military was making its quotas. I don't think calling up more people will make that much difference, because it does not effect that many people and if it helps quell the violence it can be a plus.

Immigration can cut both ways as well.

But this poll might be a outlier or not, truth is the whole idea of generic candidates might be so subjective that there is no way to get an idea.

The above hissed in response by: Terrye at August 23, 2006 1:19 PM