Tuesday, August 05, 2008

Making Sense of the Polls

A conversation I just had with Allan Loudell on WDEL got me thinking about the polling trends. Allan, who talks frequently with pollster John Zogby, has highlighted Zogby's latest phone poll that shows an 11 point swing from early July. The Zogby poll also shows some head-scratching results like a 9 percent drop for Obama among Democrats and a 36 percent swing to McCain among young voters. (By the way WDEL posts podcasts of recent interviews for those who can't always listen live.)
Zogby's polling looks more volatile than Gallup's tracking poll, which shows McCain consistently in the narrow range of 40 to 44 percent and Obama consistently in the range of 44 to 49 percent.
Pollster.com's chart that compiles results from national polls shows more of a modest bump for McCain than a drop in support for Obama. This would be consistent with the theory that McCain's negative advertising was as much meant to boost his standing among right-leaning voters as it was to cut down Obama's standing among undecideds.
How much do the polls mean at this point? Political Arithmetick has an interesting comparison of the relative standing of Obama, Kerry and Gore during the last three presidential election years:
You can see that the numbers moved quite a bit in the late summer and early fall, as the candidates showcased themselves at their conventions. The good news for Obama is that he has consistently trended ahead of McCain for most of the year, and polled better than either Kerry or Gore did in 2000 and 2004. Gore only led for a short stretch in the late summer. Kerry showed a weak lead in the summer before falling sharply and never recovering.


Blogger Nancy Willing said...

I was hoping that you would flesh out some of the comments you made in critique of the D'Anna poll. You said that you had discussed the poll with an unidentified person and knew of problems with its methodology. I consider that something of a slur. Would you mind expounding?

7:15 AM, August 07, 2008  
Blogger Tom Noyes said...

It seems to me that mentioning "problems with methodology" hardly rises to the level of a slur. If I feel free to raise questions of methodology about a prominent pollster like Zogby, why shouldn't I do the same with Vince D'Anna?

I have been told that the poll was conducted over a several weeks, and that the questions changed over time. I have not heard anything specific about the sample, which is where I would start looking for an explanation for it anomalous results.

I just find it odd that, after a year of the most intense statewide primary campaign we've seen in a generation, the two candidates could only muster 45 percent between the two of them. The poll shows Jack Markell, who was reelected two years ago with 70 percent of the vote, with only 14 percent -- a truly Protackian number.

Carl Sagan famously said that extraordinary claims require extraordinary evidence. D'Anna's poll results are so extraordinary that raising questions about the methodology should be expected and not dismissed as a slur.

8:58 AM, August 07, 2008  
Anonymous Anonymous said...

It is only a slur if you leave the incidentals 'up in the air' so thanks for pointing out the specifics of what you heard.

From what you have written here, I am fairly certain of your source and I can assume that it was perhaps given to you second hand. It is also total bullshit, by the way.

Too bad you listened to bad sources. There has been some pretty raw things coming out of a certain campaign's volunteer that is adding to this flap. That volunteer now knows that there were four completely disparate polls run. The particular poll you and he blabbed about was discarded precisely because of the length of time it was taking to gather the data (done over the fourth of July).
There never was a poll that had variant questions. The person I spoke to about the misinformation he was spreading now understands what happened. I recommend that you go back for some verification.

As far as the analysis, Ron Williams simply put out the raw head to head data. If he wanted to do other than make sensational news, he'd have actually used all of the data on the poll to show where things actually stand. Basically, Markell and Carney are tied. There are a ton of Delawareans who don't know who Jack is and barely know John and this poll was taken just before ad-gate and most of Jack's teevee commercial thrust.

I am sorry that a few people were flippant enough to trash this poll. I certainly intend to pursue some satisfaction from the campaign from where this misinformation came.

1:47 PM, August 07, 2008  
Blogger Tom Noyes said...

I've already devoted more time to this poll than I feel it deserves. I never actually posted anything on TommyWonk on this, but offered a comment elsewhere, as I recall.

I have heard other comments on the matter, which I am not inclined to pass on to anyone. If you have an issue with someone you guess may have offered a disparaging comment about the poll, that's your business. Leave me out of it.

2:08 PM, August 07, 2008  
Anonymous Anonymous said...

Tom, if you chose to drop loose rumors around the blogosphere, you surely must expect receiving direct appeals to fairness and disclosure.
I have gone ahead and contacted a particular campaign about the presumed rumor-mongerers responsible for the bullcrap being passed along for DEM loyalists disgestion and I did get some satisfaction there.
Primaries are here to stay and all DEMs should get on board with that fact.

11:11 AM, August 15, 2008  

Post a Comment

<< Home