If you were a bookie... Polls and bets on the 2004 elections

Hmm, jp, soz, I dunno ...

I mean, take me. Of all the people I've known reasonably well here in Holland, hung out with, worked with, all through the past ten years, a total of two voted right-wing liberal, noone christian-democrat and noone pim fortuyn.

Well, there's some family who vote right-wing - an uncle on one side, an uncle on the other, perhaps some cousins, and the brother of my granma voted Fortuyn, I'm sure. The family of my Dutch ex-gf's were mixed, too - the parents of one were christian-democrats, and the brothers of another probably voted right-wing. And i had three or four flatmates who voted right-wing liberal, and one who voted christian union.

But all my colleagues now, all my colleagues at my two internships, everyone at the cafe I worked at - they all voted either Green, or Socialist, or Labour, or at most Democrat. Same with my friends - in ten years, just the two exceptions. And there were two anarchists who didn't vote.

Yet there's no denying that those four parties altogether never get more than half of the vote, nationally.

Its amazing how you can have the most diverse body of colleagues, friends and acquaintances - different styles, ages, professions, interests, looks, ethnic backgrounds - but seemingly still all part of the same overall national political bracket. I had to go mingle on the "Livable Rotterdam" forum to interact with Fortuyn supporters.

Kind of a disconcerting thought.

Well, soz, the Pack's done it's job. Now it's the citizens'.

What will happen tomorrow depends on who will be the "new voters" and who will stay at home.
Will it be the young University students?
Will it be the new-born Christians who were not confident on 2000 on the former addict Bush?
Will it be the ethnic minorities?

After that, pollsters will give all kinds of explanations. "See... that's how it was, but we had to adjust the measure of voter trunout..."

Oh, that's definitely true. (That our immediate group of contacts is less politically diverse than we might think.)

I know... 4 people who say they will vote for Bush. One other can't stand Kerry but can't stand Bush either and will be voting for someone else. That leaves, whoa, REALLY A LOT of people I know who will be voting for Kerry -- certainly not representative.

I don't know about Holland, but the US has, traditionally, a bimodal distribution of votes, according to prescincts.
This means that Republican and Democrat voters tend to cluster together (because of economical, racial and cultural reasons). So it's normal for an American's contacts to be not representative of the whole of the electorate.

Hmm ... but even a "Democrat precinct" will usually have 20, 30, 40 percent Republicans, and vice versa ...

Plus, you know many people beyond your own precinct, of course.

Yet still clustered within the overall political container category, apparently ...

Well, thats human nature I guess. Not many right-wingers work for NGOs or universities, live downtown, study social sciences, do (non-church) volunteer work, work at filmfestivals, et cetera. Birds of a feather.

Missed your post there, fbaezer. Yep, the new voters... which way will it go?

FreeDuck mentioned a cell phone poll that was 55-40 (?) for Kerry.

Still in push/ pull, panic/ calm mode re: the day of reckoning finally drawing nigh. Be over already! No we need more time! Bush will win! Kerry will win!

And then I think "at least it will finally be over tomorrow night", only to be followed by "unless it's NOT... "

Oy.

An oft-discussed, much missed demographic:

From www.electoral-vote.com :



As much as this will be pooh-pah'ed by the conservatives,

It sure doesn't look BAD for Kerry!

Cycloptichorn

Thanks for the link and further info, Cycloptichorn.

Huh!

Gotta get out the vote. The local Kerry office just told us to bring umbrellas -- it might rain tomorrow, and there might be lines, and we want to keep people in line! :-)

The poll that counts will be taken tomorrow.

And on January 20th, John Kerry will be sworn in as president.

Love your unflagging optimism, Frank.

Sure hope you're right.

What makes a bi-modal distribution is precisely that few prescincts are 60-40 and a lot of prescincts are 80-20.

This are results by county, 2000, 40 counties taken at ransom (the method tends to choose bigger counties):

Baldwin, Ala: Bush 72%
Hempstead, Ark: Gore 54%
Los Angeles, Ca: Gore 73%
Monterey, Ca: Gore 58%
Tulare. Ca: Bush 60%
El Paso, Co: Bush 64%
Newcastle, Del: Gore 60%
Marion, Fla: Bush 55%
Banks, Ga: Bush 71%
Chicago, Il: Gore 80%
Monroe, In: Bush 52%
Wichita, Ks. Bush 79%
Owsley, Ky: Bush 80%
Lafayette, La. Bush 62%
Baltimore City, Md: Gore 83%
Falmouth, Mass: Gore 56%
Detroit, Mich: Gore 94%
Grand Traverse, Mich: Bush 59%
Yellow Medicine, Minn: Bush 47%
Christian, Mo: Bush 64%
Carter, Mt: Bush 89%
Cherry, Ne: Bush 81%
Cornish, NH: Bush 49%
Passaic, NJ: Gore 58%
Santa Fe, NM: Gore 65%
Bronx, NY: Gore 86%
Suffolk, NY: Gore 53%
Randolph, NC: Bush 72%
Cuyahoga, Oh: Gore 62%
Washington, Or: Gore 49%
Philadelphia, Pa: Gore 80%
Central Falls, R.I: Gore 79%
Dorchester, S.C: Bush 62%
Walworth, S.D: Bush 69%
Putnam, Tn: Bush 50%
Bandera, Tx: Bush 77%
Starr, Tx: Gore 77%
East Montpelier, Vt: Gore 52%
Appomatox, Va: Bush 62%
Pierce Wa: Gore 52%

As you can see, the norm (for counties, even more for prescincts) is that the winner has a clear advantage over the runner-up.
In this random sample, only in one out of every four counties, the winner gets less than 60% of the vote.

I bet this doen's happen in the Netherlands.

'K, I'm not buying the cellphone stuff. I've seen calculations on it that all showed it to be very marginal, but I dont have any links anymore ... and i cant remember any numbers, alas.

But lets go out on a limb here. How many people nowadays have a cellphone, but no landline at all? That'd be mostly youngsters, right? Say, 10-15% of those aged 18-30? Thats roughly 2-3% of the overall electorate.

Throw in a few postmodern, single urban 30-somethings that did away with their landline too & some folks living out in a trailer or in the boonies that might never had one, and you get to perhaps 4% of the voting-age population, 5% at most?

Next question, how does their turnout compare? Youngsters turn out notoriously less than older people. Urban folks less than those from smaller towns. Singles turn out less than married couples (and couples, I assume, are more likely to have a landline phone). So can I guess that the cellphone-only folks will turn out less? Brings 'em down to, say, 3% of the voters?

I'm just pulling these numbers from the air, really. Going on what I would guess, generally speaking. But if we're talking 3% of the electorate that will vote Kerry by 55%/40% instead of by 50%/50%, then we're not talking bout much of any impact, whatsoever ...

Eg., you have 1000 voters (100%). 30 of them (3%) vote for Kerry 55% to 40% - ie, 17 votes for Kerry, 12 for Bush, instead of breaking down according to the national poll average, 15 for Bush, 14 for Kerry. Then all in all we're talking of a 3 vote difference compared to what the polls would have shown. Three on the total of 1000 = 0,3%.

Course, race is so tight, a 0,3% difference might just flip it around. But compared to the issue of likely voter samples pretty much excluding most first-time or irregular voters, period, or pollsters finding that even of those with landlines, 2 out of 3 refuse to answer their questions, it seems a bit of a non-issue. Unless someone can debunk all my assumptions above, of course. But just use your own and then calculate what the impact would be - you might be surprised.

Zogby, Rasmussen, and Gallup all have commented on cellphones. In a nutshell, all discount any impact, as their methods entail completing interviews with a sufficiently large population as to encompass a representative sample of each significant demographic, age, gender, geographic distribution, party identification, etc. While a poll's results may be drawn from the responses of a few hundred respondents, those few hundred selected as the polling sample are drawn at random from among a much larger randomized pool of completed responses.

Concern over the undersampling of cellphone users is about as well founded as concern over undersampling folks who drive green cars or wear boxer shorts.

Yeah ... just calculated the average of the above, makes for 65% for yer typical election winner.

And that was just basically my point, really ... even with 65% for the winner on average, that means that yer average American will still have one fellow-inhabitant of his county thats from the other party for every two thats from his own.

All this talk of red states versus blue states makes the impression of monolithic territorial blocks - the coasts versus the fly-over states, stuff like that - where in each, folks of the other party are simply unheard of. Thats generally a rhetorical exaggeration, one that better fits the reality of strikingly-coloured powerpoint charts than the messy diversity on the ground.

Of course, what we were observing just now is that even though generally speaking there must still be plenty of folks of the other party around in your own town, you're still relatively unlikely to mingle with all that many of 'em because your workplace, free-time activities, field of your studies, whatever, are often culturally stratified as well.


Well, depends how you count, of course.

I mean, we dont just have two parties, we have ten. So a single party getting over 50% in a town is rare. Nowadays, anyway.

(It used to be quite common, because of the way Dutch society, until the sixties, was ordered by the system of "pillarisation", with Catholics, Protestants and Socialdemocrats all going to their own soccer club, camping site, scouts organisation, allotment gardens plot, perusing their own radio station, TV channel, newspaper, going to their own (catholic, protestant or general) school or university, et cetera. In the south, the Catholic People's Party easily ratcheted up 60%, 70% of the vote for example; in the very northeast, the Communists got 50%. But that on an aside. I once wrote a lot about it all for a thread here on A2K, see here)

So yeah, no party with over 60%. But if you group the Dutch parties together in left and right-wing ones, you get the same thing of course. Leftwing parties getting two-thirds of the vote in the major cities, right-wing parties getting it in the rural areas and luxury towns.

Thats why I mentioned, you know - my circle of acquaintances always seemed pretty diverse, what with socialists, green lefters, labour party socialdemocrats and the odd wishy-washy liberal all regularly disagreeing with each other (just like you got many different kinds of Democrats in the States). But when the **** hits the fan, they're all just "left-wingers", of course - so there is a similar kind of pattern. There's little party loyalty, lots of floating voters - but most people I know only ever "float" from one left-wing party to another.

Ah, I'm just rambling now ... havent really got any specific point to make. Just random observations ;-)

Gettin' confusing.

So, yesterday, there was a CBS/NYT poll out that I posted here, based on interviews of 824 registered voters done Thursday through Saturday, that had Bush up at 49% and Kerry at 46% among likely voters.

Today, CBS has a new one out, based on the same 824 interviews plus an additional 370 done on Sunday.

The new results based on the Thu-Sun totals, likely voter sample, now are:

Bush 47%
Kerry 46%
Nader 1%

Now, I would conclude that this means Bush had a very bad day on Sunday, right?

But the trouble is that this article here, after providing the above horserace numbers, subsequently only compares data for the Thu-Sat sample with those of the totals for Sat and Sun together (700 interviews in all).

And on that count, it notes that "in the interviews conducted over the weekend days Saturday and Sunday, presidential preference results are the same as for the entire four days reported here."

So ... does that mean Bush in turn had a good day on Saturday?

Nevertheless, on the specifics of who "will protect the country from terrorism" and has "strong qualities of leadership", Bush did do noticeably worse in the weekend overall than in the Thu-Sat sample.

Eeeeehhh ... count me confused. And call the race "too close to call". ;-)

Here's an entertaining little summary take on the cellphone thing - and some other stuff:

Do the Amish Have Cell Phones?

And here's an update to that practical state-by-state table of polling data averages.

If from that table, you flip Ohio (B +0,1) and Iowa (B +1,3) to Kerry, but not Wisconsin (B +1,1) - you get a tie in the EC ...

So Kerry needs to win both Ohio and Wisconsin, or Florida and any other state now still coloured red in the table.

All he needs, thus, is a 1,2% last-day swing in these states ... which is funny and in a way just, because thats also exactly what he would need to overcome the average Bush lead in the past week's national polls (currently at 1,1%).

The Casino's are sliding back towards Bush... with 3 hours left before the cutoff.



(That's 8 to 5)

Just heard the funniest thing on the radio. CBC host interviewing the Canadian leaders of Democrats Abroad and Republicans Abroad. The Republicans Abroad talking head said that it was clear (?) that George Bush would win the popular vote. She was concerned about the Electoral College however. I literally shook my head and looked at the radio. Turns out she wasn't kidding.

On the cellphone-only group: Setanta and a few of his friends are in that group. Lord only knows what demographic they're supposed to fit in with, or what that would mean for the election. All I know for sure is they only have cellphones, and they are voting.

I too only have a cell phone and I'm voting. My married with 1 child friends in Wisconsin recently switched to a 3 cell plan and canceled there home plan as well.