Poll Vault

[Puppycow]

After the conventions.

It will also be interesting to see if there is any change after veeps are selected.

 

[David Wong]

I guess I get too caught up in the coverage and forget that I'm one of a tiny sliver of Americans who are actually paying attention at this point. There'll be some gaffe by Obama and the pundits will say, "Whoa! This could cost him big time!" and then two days later when the polls haven't moved, the headlines are all, "OBAMA WEATHERS FLIP FLOP GAFFE" when in reality, Obama weathered it because virtually no one is paying attention in June.

It's kind of humorous I guess because the punditry acts like the world begins and ends with the political horse race, how this sound bite or this speech will move the polls, when in reality they're all just talking to each other at this point. The bleachers are empty.

You've got a huge chunk of Americans who have already made up their minds based on some vague sentiment (Republicans are crooks, Obama is a commie, McCain is old, Obama is a muslim, McCain is just like Bush, etc) and of the remaining undecideds, most get their news by browsing headlines on Yahoo before work and from late night talk show monologues.

The 5% or so who are studying policy and trying to make some kind of informed decision won't have much of an effect one way or the other. It's easy to forget that.

 

[Puppycow]

I guess I get too caught up in the coverage and forget that I'm one of a tiny sliver of Americans who are actually paying attention at this point. There'll be some gaffe by Obama and the pundits will say, "Whoa! This could cost him big time!" and then two days later when the polls haven't moved, the headlines are all, "OBAMA WEATHERS FLIP FLOP GAFFE" when in reality, Obama weathered it because virtually no one is paying attention in June.

It's kind of humorous I guess because the punditry acts like the world begins and ends with the political horse race, how this sound bite or this speech will move the polls, when in reality they're all just talking to each other at this point. The bleachers are empty.

You've got a huge chunk of Americans who have already made up their minds based on some vague sentiment (Republicans are crooks, Obama is a commie, McCain is old, Obama is a muslim, McCain is just like Bush, etc) and of the remaining undecideds, most get their news by browsing headlines on Yahoo before work and from late night talk show monologues.

The 5% or so who are studying policy and trying to make some kind of informed decision won't have much of an effect one way or the other. It's easy to forget that.

I hope it's not quite as bad as all that, but I think that you are in the ballpark. I heard some caller to the Diane Rehm show the other day speculating that McCain's trip to Columbia was planned to coincide with the freeing of the hostages. I don't believe that, and I seriously doubt that most people are even aware of it or care.

 

[David Wong]

Well I didn't want to come off as superior to the voting common folk, I just meant that the discussion sounds a lot like what you get on pro football boards around May and June. There's just nothing going on and so you get all these tangential threads about wild trade rumors and which guy got busted for drunk driving in the offseason.

It's more people trying to find things to talk about, rather than actually being engaged in the subject. So it feels the same here, lots of talk but the underlying feeling that none of it will really matter once the real season gets here.

 

[Puppycow]

Well I didn't want to come off as superior to the voting common folk, I just meant that the discussion sounds a lot like what you get on pro football boards around May and June. There's just nothing going on and so you get all these tangential threads about wild trade rumors and which guy got busted for drunk driving in the offseason.

It's more people trying to find things to talk about, rather than actually being engaged in the subject. So it feels the same here, lots of talk but the underlying feeling that none of it will really matter once the real season gets here.

Ah, I get it. Like the difference between casual baseball fans and those who play fantasy baseball and subscribe to Baseball Prospectus and read everything by Bill James, etc.

I've been in the latter camp, but not in the last couple years. Sorta lost interest.

 

[Brainster]

Here's a good example of how different polling methods produce different results:

All told, Obama leads McCain among registered voters by 8 points in this ABC/Post poll, 50-42 percent.

Turnout makes a difference: It's 51-39 percent among all adults overall, but narrows to a close 49-46 percent match among likely voters. As noted, Republicans are among those more likely to vote; young adults, less so.

To recap (since it's presented out of order)

1. Adults: Obama by 12
2. Registered voters: Obama by 8
3. Likely voters: Obama by 3

It's not too hard to see that Obama swamps McCain among adults who are not registered to vote, and that he similarly crushes McCain among adults who are registered but unlikely to vote. Indeed, if we assume that the percentage of adults who are not registered to vote are about 20% of the population, and that the percentage of registered voters who are not likely to vote are about 20% (so that the percentage of adults who are likely to vote is about 64%), then Obama beats McCain by about 2-1 in both groups.

Another interesting result is the sharp decline in intent to vote. Among young voters (18-29), 66% expressed a certainty of voting in March; currently that's down to 46%. There's a similar trend among Democrats in general, with 82% expressing their certainty to vote in March versus 66% now.

 

[Pookster]

Methodology goes even beyond whether it's likely, registered, etc. voters. Many polling organizations also weight/adjust their results based on other factors such as party affiliation, projected turnout, etc.. Two organizations could do the exact same questionnaire with the exact same people with get the exact same answers and report very different results.

Also, "likely" voters is determined differently by many organizations as well. They generally to have a kind of tickler list of questions that triggers what they believe makes a "likely" voter. The only real way to determine who might have the best methodology is to compare them to the actual election results.

This is also why I think "poll of polls" like CNN and RCP does are of no value.

 

[aerosolben]

The only real way to determine who might have the best methodology is to compare them to the actual election results.

This is also why I think "poll of polls" like CNN and RCP does are of no value.

The 538 meta-poll mentioned previous does weighting based on actual previous election results.

 

[Brainster]

Methodology goes even beyond whether it's likely, registered, etc. voters. Many polling organizations also weight/adjust their results based on other factors such as party affiliation, projected turnout, etc.. Two organizations could do the exact same questionnaire with the exact same people with get the exact same answers and report very different results.

Correct. The pollster will tell you that it's what they do after they get the sample results that is the "science" of polling.

Also, "likely" voters is determined differently by many organizations as well. They generally to have a kind of tickler list of questions that triggers what they believe makes a "likely" voter. The only real way to determine who might have the best methodology is to compare them to the actual election results.

Models of likely voter turnout are discussed here. Most use self-reported likelihood of voting with adjustments for recent voting behavior and self-reported interest in the race.

This is also why I think "poll of polls" like CNN and RCP does are of no value.

An average reduces the importance of outliers and so hones in on more likely results, but strictly mathematically speaking I get where you're coming from. I hadn't noticed before, but RCP apparently uses a simple mathematical average, so that Reuters/Zogby's recent poll of 1039 likely voters is considered the same in weight as Rasmussen's poll of 3000 likely voters, even though the latter sample is almost three times as large.

 

[Pookster]

An average reduces the importance of outliers and so hones in on more likely results, but strictly mathematically speaking I get where you're coming from. I hadn't noticed before, but RCP apparently uses a simple mathematical average, so that Reuters/Zogby's recent poll of 1039 likely voters is considered the same in weight as Rasmussen's poll of 3000 likely voters, even though the latter sample is almost three times as large.

It really goes beyond the math aspect of it for me. It's rare that polls ask the same questions the same way. Even if they do, the order of questions is likely not the same. Was the poll web-based, a live person, or an automated voice? If it was a live person, was it a male or female interviewer (yes, women do get less "hangups" than men interviewers)? All of these kinds of things and more can influence answers. The outlier may well be extremely accurate --- for the specifically phrased question asked. Trying to average different polls like this is almost the epitome of comparing apples and oranges. Yes, they're fruit (presidential election polls), but they're often very different things in their own right. It's a futile effort to try to average them to show anything of value.

 

[Undesired Walrus]

The electoral map as it looks today.

 

[gdnp]

Polls in the summer, before the conventions, are like looking at a marathon at the 13 mile mark. Sure, I'd rather be ahead than behind, but there are so many things that could affect the results between now and the finish line that I wouldn't put too much stock in them.

 

[David Wong]

But what will we do about our boredom?

 

[David Wong]

Gallup shows Obama with a 9-point lead (his largest ever) after the world tour. Suggests maybe a little 4 or 5 point bump from all the positive coverage:

http://www.gallup.com/poll/109102/Gallup-Daily-Obama-49-McCain-40.aspx

Doesn't mean much, just interesting that the trip abroad helped him. I'm assuming that's what it was.

 

[gdnp]

I'm not surprised. One of the biggest knocks against Obama is his inexperience--that he isn't ready to lead the free world. Traveling through Europe, Afghanistan, Iraq, and the middle east without any significant blunders, acting presidential and being received in an almost-presidential fashion, was an important "screen test" that he passed with flying colors.

There are lots of ways he could have slipped up. He could have been shunned by the soldiers. He could have gone on a Mike Dukakis tank ride and looked rediculous. He could have made a verbal gaff.

 

[David Wong]

That's actually a good point, and may indicate that there's a significant chunk of voters out there who just want to be reassured that he's not a total moron.

Maybe that's where the debates will be important, it may not be that people will watch to see what his positions are, but will just pay attention to whether or not he can speak intelligently. Especially if all the coverage they've seen so far is gaffe-related.

 

[Undesired Walrus]

As a non-American, what is the best poll to trust? I ask as yesterday I saw a report in which there was little to no bump for Obama, which stands in contrast with the above 4-5 increase.

 

[Brainster]

Gallup's recent poll was of registered voters; Rasmussen's tracking poll of 3,000 likely voters shows Obama with a three-point lead. Any "Berlin bounce" seems to have worn off:

Following his speech in Berlin, Obama enjoyed two very strong nights of polling on Thursday and Friday. His lead grew to six-points for results released on Saturday (see recent daily results). However, polling on Saturday and Sunday showed the candidates much closer with single-day results similar to polling from before the Berlin speech.

 

[chipmunk stew]

As a non-American, what is the best poll to trust? I ask as yesterday I saw a report in which there was little to no bump for Obama, which stands in contrast with the above 4-5 increase.

No single poll taken over a day or two will ever give you an accurate snapshot of the mood of the country (or whomever the polled population is supposed to represent.) I don't care how careful the pollsters are. Better to look at trends.

The "bump" or lack thereof could just as easily be noise in a longer-term trend.

That's why I prefer the tracking polls that bundle a bunch of individual polls over time and apply rolling averages and trendlines. You don't get the immediate feedback that the 24hr news cycle prefers (you have to wait a few days after an event to see what impact it had, if any) but I think you get a more accurate feel for the mood of the country and what is affecting people.

 

[Brainster]

USA Today/Gallup has a stunner:

Republican presidential candidate John McCain moved from being behind by 6 points among "likely" voters a month ago to a 4-point lead over Democrat Barack Obama among that group in the latest USA TODAY/Gallup Poll. McCain still trails slightly among the broader universe of "registered" voters. By both measures, the race is tight.

The Friday-Sunday poll, mostly conducted as Obama was returning from his much-publicized overseas trip and released just this hour, shows McCain now ahead 49%-45% among voters that Gallup believes are most likely to go to the polls in November. In late June, he was behind among likely voters, 50%-44%.

Among registered voters, McCain still trails Obama, but by less. He is behind by 3 percentage points in the new poll (47%-44%) vs. a 6-point disadvantage (48%-42%) in late June.

Clearly an outlier for now; if there's really been a shift it will show up in other polls.