Interesting Data from the AP about Undecideds

[mscelina]

http://news.yahoo.com/s/ap/20081031/ap_on_el_pr/ap_yahoo_poll_changeable_voters_9

Apparently, I'm not alone on the fence. With 14% of voters still undecided or admitting that their allegiance could change, the last few days of this campaign are crucial.


Overall, the share of these voters — sometimes referred to as "persuadables" — has barely budged from levels measured in June and September AP-Yahoo News polls, conducted online by Knowledge Networks.
But the survey — which has repeatedly quizzed the same group of 2,000 adults since last November — shows considerable churning below the surface. Of those now changeable, nearly three-quarters said in June their minds were made up, and half said so just last month.


Are we to take from this that the campaign, instead of turning people ON to a candidate has instead turned them OFF? An interesting dynamic, if that's the case.

[rugcat]

Apparently, I'm not alone on the fence. With 14% of voters still undecided or admitting that their allegiance could change, the last few days of this campaign are crucial. Traditionally, undecided voters split fairly evenly. But I believe that this time, the undecideds will go to McCain 2 to 1 or even 3 to 1. (As per my previous post.)

And that is McCains only chance -- will enough undecideds vote for him to overcome Obama's lead? It's not impossible, not by any means.

[Ken]

14% is a lot. How does this compare to percentages of undecideds in past elections?
ps Neat term "persuadables."

[mscelina]

Well, if you take that statistical percentage into the battleground states, it makes for an interesting electoral swing that's for sure.

[mscelina]

14% is a lot. How does this compare to percentages of undecideds in past elections?
ps Neat term "persuadables."

Well, I just mentioned in another thread that Ross Perot took approximately 19% in the '92 presidential election--that might be one way to gauge the undecideds in comparison to that election if nothing else. As I recollect (I'll try to verify this) there were comparable numbers in 2004--most of whom swung to Bush in the last two weeks of the campaign.

[Ken]

...14% would be signifigantly lower # of undecideds than in past years on average as it would seem, then, suggesting McCain & Obama have run effective campaigns, but perhaps not so much so towards the end, as you say.

[mscelina]

Here's the killer:

Half are independents, more than double their proportion among decided voters. But, as with decided voters, more persuadables are Democrats than Republicans. Four in 10 supported Clinton's candidacy this spring.

"She got cheated, I thought," said Chris Markle, 25, who's from Schenectady, N.Y., and now leans toward McCain. "I'm kind of upset about that."

[Ken]

...those dems must really be upset about Hil's dropping out, if they remain undecided with Hil and hubby going all about the country and pleading for people to vote Obama.

[mscelina]

Yep. In effect, it makes Hillary the Ross Perot of this election.

*shudders*

what a terrifying mental image.

[Ken]

...good for Halloween ;-)

[Tirjasdyn]

If 100% of those 14% vote for the same third party...suddenly we'll have a bigger contender for the next election.

Just a thought.

[mscelina]

Hence the comparison to Ross Perot, who took 19% of the electoral vote in 1992.

[blacbird]

So . . . let's look at some real numbers, again from the RCP averages: First of all, RCP doesn't show that level of "undecided" anywhere in recent weeks. Their most recent poll including Barr and Nader (who collectively pull about 4%) show an undecided or uncommitted number of around 7%. Which is effectively the lead Obama has over McCain. Meaning that McCain essentially needs 100% of those undecideds to make it a race. Either that, or he needs somehow to take a bite out of the current Obama support. Tough job, in either case, with four days left.

caw

[maestrowork]

Hence the comparison to Ross Perot, who took 19% of the electoral vote in 1992.

Except Perot was actually actively campaigning, participated in the debates, etc. That he was a viable third candidate in that race. Nobody has emerged in such a fashion this election. Honestly, I'm not even aware of who they are, let alone their platforms. I can't possibly see how even half of those undecided votes go to a third party candidate -- many of them are not even on the ballots. It would take a lot of write-in votes to make that happen.

I tend to agree that these undecided voters would either a) stay home, or b) vote 2:1 for McCain.

[mscelina]

So . . . let's look at some real numbers, again from the RCP averages:

blacbird, this isn't an article from the RCP. It's an independent poll from Yahoo and the Associated Press that has nothing to do with the RCP. Amazingly enough, not everything has to go through the RCP.

The numbers in this poll are just as real and as credible as any from the RCP. It's just from a different source. That's one of the great things about politics--we have lots of different avenues for source material.

[blacbird]

Okay, then, this, from the Yahoo article:

Almost four in 10 persuadables lean toward McCain, and about as many are considering backing Obama, while the rest are either undecided or lean toward other candidates. Viewed another way, about one in every 10 supporters of Obama or McCain says he could still change his mind.

Even so, persuadable voters could be especially fertile hunting ground for McCain in the closing days of a contest in which most polls show him trailing.

Probably the most meaningful thing in it, as it didn't anywhere cite things like the number or kinds of people interviewed, the margin of error, etc. And I'm not disparaging it, or anything else. I use RCP precisely because they collect data from dozens of polling sources, and are scrupulous about reporting real numbers. I got on you a little about your post in the other thread primarily because you didn't really do that, and when I did even a small amount of digging, I was able to find out the real reality of the thing. Which was that it didn't really say what you purported it to say.

caw

[mscelina]

Okay, then, this, from the Yahoo article:



Probably the most meaningful thing in it, as it didn't anywhere cite things like the number or kinds of people interviewed, the margin of error, etc. And I'm not disparaging it, or anything else. I use RCP precisely because they collect data from dozens of polling sources, and are scrupulous about reporting real numbers. I got on you a little about your post in the other thread primarily because you didn't really do that, and when I did even a small amount of digging, I was able to find out the real reality of the thing. Which was that it didn't really say what you purported it to say.

caw

I beg your pardon?

blacbird, I don't give a rat's patootey what source material you use. I really don't. Quite frankly, as I've been zapped for overuse of source citation in this forum before, I'm at a loss here for what this post is implying. So is the new rule that only blacbird's sources are credible? If so, I hope somebody lets me know so that I can adjust my posting habits accordingly.

Bullshit.

Don't make any assumptions on what I purport to say on anything. Don't make any assumptions about the 'real' reality. For every source you list, I can find and cite one that says the direct opposite. The last time I checked, this wasn't softcore P&CE according to blacbird.

[blacbird]

You said, in the other thread, that the poll you yourself cited indicated the Ohio race to be a "statistical tie". Using the same poll data, I simply pointed out that it wasn't. Sorry you have such a problem with that.

caw

[mscelina]

I don't have a problem with that at all. I'll admit that a 1 percentage point difference between 4 points and a usual +/- 3 point margin of error isn't a statistical tie. Would you like me to edit my post so that it reads, "As of this morning, Ohio is one point away from a statistical tie?"

I'd be more than happy to do so if that will appeal to your sensibilities of softcore posting accuracy, blacbird.

[Ken]

...to be honest, all these polls seem like a bunch of phooey. I'll be listening to the news on one channel and they'll say Obama has a double digit lead, and then I'll switch to another channel and they'll say the candidates are tied in a statictical dead heat. It would be comical if the matter wasn't so serious.

[mscelina]

Polls really are a bunch of phooey when you get right down to it. Nobody knows what will happen on the 4th. Hell, we probably won't have a clue what happened on Election Day for three or four weeks after the election.

[Ken]

...we probably won't have a clue what happened on Election Day for three or four weeks after the election.

there seems to be a lot of impatience for the election to be over and settled. If results were delayed that long, the country might really lose it.

[blacbird]

Would you like me to edit my post so that it reads, "As of this morning, Ohio is one point away from a statistical tie?"

It would have been a little better if you'd done that in the first place. It would have been much better if you had provided a link, initially. It also would have been better if you'd pointed out that it wasn't "as of this morning". It was "as of three days ago", if you check the dates on the poll (which I was able to do with minimal trouble). That poll, as I pointed out, was no more recent than several others.

Look, if you're going to cite numbers to bolster your argument, that's fine. Just be sure you know what you're citing, and understand that they'll get checked.

caw

[mscelina]

Whatever, blacbird.

This is softcore, not hardcore. I believe I mentioned in the first post we had about this tiresome and nonending debate that I heard the poll results on the news this morning. I'd really appreciate it if you stopped pursuing me all over the forum so that you can attempt to make some obscure point about how I choose to post. I posted that in context within a conversation, not to get into a debate with you over whether the poll was accurate. Besides, I'm sure that the local news doesn't use a source that meets with your approval.

And I'm quite familiar with the procedure and research necessary to cite numbers, thank you. I can assure you that I am very well aware of what I am citing when I choose to cite anything. Occasionally, I'll even cite RCP just to piss you off.

If you want to complain about my posting habits, please do so in PM. Thanks.

[Sheryl Nantus]

I really don't trust polls - people can and do lie to the pollsters just to get them off of their backs. Not to mention saying something that's politically correct for the area.

If you live in a McCain-friendly area, you're going to say you're voting McCain. Same for Obama.

Only truth that counts is when they actually count the votes.

Until then, nothing's certain.

[blacbird]

If you want to complain about my posting habits, please do so in PM. Thanks.

I got no complaints. I'm not even remotely angry. You can always neg-rep me if it makes you feel better.

caw

[blacbird]

I really don't trust polls

Perfectly acceptable. Come back and talk to us about this on Wednesday.

caw

[whistlelock]

The thing about polls- well, we're a risk averse society. We don't like us some risk. So, we try to predict everything. Even the weather. I mean, you can't do anything about the weather so why worry about it?

But, we do.

So, we try to use observed past behavior to have a good idea of future behavior.

And, if you are curious about their validity over time and how seriously we take them as a culture; go look up Carter's behavior in the days before voting day when he was told he would lose to Reagan in a massive landslide.

[dmytryp]

Perfectly acceptable. Come back and talk to us about this on Wednesday.

caw
So, tell me, what poll do you trust? RCP average takes into account polls like PEW that showed 15 points advantage for Obama while the rest of the polls on the same day (or +- 1 day) showed between 2 and 7 point advantage. Or do you trust Gallup (traditional poll) that showed 2 point advantage Obama and then three days later a seven point advantage.

As an aside, exit polls in Kadima primaries (here) showed something like 10% advantage for Livni. It ended with 430 votes deciding the race.

[whistlelock]

http://www.washingtonpost.com/wp-dyn/content/story/2008/11/01/ST2008110101102.html

A nice write-up on polls.

Which, essentially says, "We have no idea what is really going to happen. None."

[blacbird]

So, tell me, what poll do you trust? RCP average takes into account polls like PEW that showed 15 points advantage for Obama while the rest of the polls on the same day (or +- 1 day) showed between 2 and 7 point advantage. Or do you trust Gallup (traditional poll) that showed 2 point advantage Obama and then three days later a seven point advantage.

I don't exactly "trust" any one of them. I have a little bit of confidence in the statistical spread of many, however, the more the better. And when I see a single one that is, say, three points or more astray of all the rest, in whatever direction, I ignore it, unless it later gets joined by brethren.

The one thing I absolutely don't believe, however, is that all the polls can be skewed in the same direction by communal lying, or some secret conspiracy to do so.

caw

[dmytryp]

I don't exactly "trust" any one of them. I have a little bit of confidence in the statistical spread of many, however, the more the better. And when I see a single one that is, say, three points or more astray of all the rest, in whatever direction, I ignore it, unless it later gets joined by brethren.

The one thing I absolutely don't believe, however, is that all the polls can be skewed in the same direction by communal lying, or some secret conspiracy to do so.

caw
But there is no conspiracy, not even with the stray PEW poll. Election polls depend on weighing which is an indicator of probable showup for the elections. In this election nobody actually knows what is going to happen. This is the reason Obama outperformed many of the polls in the primaries -- high turn up. The pollsters might have adjusted for that. They might not. They might have adjusted for energised Rep base or they may not.

[blacbird]

In this election nobody actually knows what is going to happen. This is the reason Obama outperformed many of the polls in the primaries -- high turn up. The pollsters might have adjusted for that. They might not. They might have adjusted for energised Rep base or they may not.

Nobody knows, obviously, but none of the polls claim that. What they work on are perceptions of probability. And I'm glad you recognize that there is no conspiracy. But we've had people make comments on this very board implying their suspicion that there is such. I guess some people need to have some reason to dismiss the poll numbers. Me, I don't claim them to be perfect indicators of what's going on, but I'm not about to say they are meaningless. We've had more than a few people express the latter opinion, sometimes with remarkably creative feats of rationalization.

caw

[dmytryp]

Nobody knows, obviously, but none of the polls claim that. What they work on are perceptions of probability. And I'm glad you recognize that there is no conspiracy. But we've had people make comments on this very board implying their suspicion that there is such. I guess some people need to have some reason to dismiss the poll numbers. Me, I don't claim them to be perfect indicators of what's going on, but I'm not about to say they are meaningless. We've had more than a few people express the latter opinion, sometimes with remarkably creative feats of rationalization.

caw
Well, if the spread of most polls is around 5% (not sayng it is, I haven't checked lately and they change rather frequently) and there is a signifficant number of undecideds (which many believe would break for McCain) and gathered with the inherent innacuracies, I think the polls have little value.

[blacbird]

I think the polls have little value.

And you're welcome to the opinion, and I respect it. We'll see in three days.

caw

[Bravo]

"Undecided voters are less educated, less affluent, and somewhat more likely to be female than the average voter," Pew said of the survey results, noting the undecided voters are also more likely to attend church regularly.

http://news.yahoo.com/s/nm/20081029/pl_nm/us_usa_politics_undecided

:Shrug:

[whistlelock]

I found one article that said McCain was depending on Wal-Mart moms. Women who shopped wal-mart (obviously), had several children, attended church regularly, and worked part-time.


And that made me say, "wtf? have we fractured our electorate so much that we can drill down to that level of information."

Let's see, have we contacted women in their 20's that attended a few classes at a Community College, work at a discount retailer, unmarried but in a steady relationship with a man but secretly worry they are gay? Has anyone called them yet?!?