Print Story It's raining shoemakers' apprentices
Diary
By TheophileEscargot (Thu Nov 01, 2012 at 01:54:07 PM EST) Reading, Watching, MLP, Politics (all tags)
Watching: "Beasts of the Southern Wild". Reading: "Are You My Mother". Politics. Links.


What I'm Watching
Saw Beasts of the Southern Wild at the cinema. Movie about a young girl called Hushpuppy living with her father in a fragile settlement called Bathtup in the Southern US, outside the protection of the levees.

Does a great job of depicting the world from Hushpuppy's point of view. That includes some fantasy elements, in particular the aurochs, wild beasts that she imagines to be huge shaggy versions of the potbellied pigs they keep, which seem to symbolize the fears and threats that face her.

Good performances from the cast of unknowns, a fascinating setting, and some tearjerking moments. Good film, well worth seeing.

Review. Silly but much-hyped article, rebuttal, interview, filming the aurochs.

What I'm Reading
Finished the biographical comic Are You My Mother by Alison Bechdel, the lesbian comics writer/artist known for the "Bechdel Test". Mostly about the author's relationship with her mother, interspersed with some dreams some theory from psychoanalyst Donald Winnicott, and a bit of Virgina Woolf.

Seemed well-drawn, with some nice character studies and the dream sequences giving a bit of variety from the endless people-in-rooms-talking. I found the psychoanalysis felt a bit claustrophobic, with every coincidence and accicent interpreted as a message from the subconscious, a bit like Woody Allen would be if he didn't have a sense of humour about it.

Overall, interesting if you're OK with microanalysing ordinary relationships; would be deeply dull if you're not.

US election
I hope I'm wrong, but I'm predicting Romney to beat Obama by 1% in the vote share. (And yes, win the electoral college).

Endorsement: Barack Obama. Yes, he's been a pretty crap president, but I think Romney is likely to be worse.

Links
Socioeconomics. Jobs armageddon is overstated, no it isn't. Suicide is now much more of a class issue than it was in the past. How much work are people getting? UK under Cameron.

Politics. Excellent EU speech by Nick Clegg (Not normally a fan but glad he wrote this as I was thinking of writing something similar on the unrealism of the pick-and-choose fantasy.) Direct democracy in Michigan. Abortion is a product of poverty and maternal distress Labour shouldn't return to Blairism. Camerons war on gangs backfired.

Random. Old types of British bar, via. Idiomatic expressions for heavy rain in many different languages.

Articles. Laurie Penny on feminism, utopianism and science fiction. British society: Not so broken.

< Smart As... so far. | Breaking the "no politics" rule >
It's raining shoemakers' apprentices | 42 comments (42 topical, 0 hidden) | Trackback
I think Obama's going to win by georgeha (4.00 / 2) #1 Thu Nov 01, 2012 at 02:03:12 PM EST
Nate Silver and Sam Wang are going with that, polls showing Romney in the lead are odd.


dkos says Gallup won't release a poll until Monday by georgeha (4.00 / 2) #2 Thu Nov 01, 2012 at 02:49:23 PM EST
and Gallup was one of the few polls showing Obama doing poorly. Odd coincidence.


US Election by jimgon (4.00 / 2) #3 Thu Nov 01, 2012 at 05:16:09 PM EST
The election is going to be a long night, but once Ohio and Colorado are called it'll be over.  My guess is that Obama is going to pull this through.  Nate Silver had a blog post today in which he also reviewed other sites that do what he does with the state level polls.  They're all pretty much in agreement that Obama is leading in enough states to likely win.  Of course anything change, but I think the hurricane was a game changer in a way.  Chris Christie couldn't play politics and had to say something nice about Obama.  Obama is doing disaster management while Romney does lowkey campaigning.  Additionally Bloomberg has finally come out for Obama.  Lots of small things that will add up. 




---------------
Technician - "We can't even get decent physical health care. Mental health is like witchcraft here."
And suddenly, climate change by wiredog (4.00 / 1) #4 Thu Nov 01, 2012 at 06:36:53 PM EST
is an issue, and the Republicans have been busily denying it. Given the amount of early voting, I wonder if the election is, essentially, over.

Earth First!
(We can strip mine the rest later.)

[ Parent ]
I think the election probably is over by jimgon (4.00 / 1) #14 Fri Nov 02, 2012 at 05:51:11 AM EST
And I think we don't hear much about climate change because the majority understands that it's real.  We're dealing with the outliers now and they'll start getting less press.




---------------
Technician - "We can't even get decent physical health care. Mental health is like witchcraft here."
[ Parent ]
Christie might also see the internal polls by georgeha (4.00 / 1) #6 Thu Nov 01, 2012 at 07:49:33 PM EST
If he thought Romney had a chance he'd welcome his help.


[ Parent ]
Sandy by ucblockhead (4.00 / 1) #8 Thu Nov 01, 2012 at 09:56:10 PM EST
Sandy response, particularly with one resulting endorsement and one might-as-well-be and endorsement from a Republican are likely to have an affect.

Looking at Nate Silver's site, it seems that there are just very few states where there's really a question.  He's calling it 80% chance of Obama just now, and at this point for Romney to win, you have to assume that there's a systematic error in Obama's favor in states like Ohio.  Ohio in particular is an important one as pretty much everyone agrees Romney has to win it and only two polls out of the last twenty show him ahead.

I don't think we have to wait for Colorado.  If he wins Ohio, Obama's won, unless something very surprising has happened.  (Like Obama losing Pennsylvania.)

Given the amount of early voting, it's even too late for an October surprise.  A Romney victory can only mean the polls have been systematically underestimating his chances.  (Which is possible...I suppose there's a 1 in 5 chance of it.)
---
[ucblockhead is] useless and subhuman

[ Parent ]
Polls by TheophileEscargot (4.00 / 1) #10 Fri Nov 02, 2012 at 04:44:15 AM EST
Firstly, I see a lot of blog comments along the lines "national polls are pointless, it's much more accurate to look at the swing states". I have two problems with that, one theoretical, one practical.

First theoretically, this is very subject to "black swan" events, if one of the non-swing states goes an unexpected way. Let's suppose you can be 99% confident that a non-swing state will go a certain way, and that there are 40 non-swing states. The probability that at least one of the non-swing states will surprise you is then 1 - (0.99 ^ 40), or 33%, quite significant.

Second, practically: I haven't seen a whole lot of evidence that if you actually study national predictions versus swing-state predictions, the swing-state ones do best. There was an interestingly vague post on fivethirtyeight recently that said (my italics):

In recent elections — since state polling data became more robust — it’s the state polls that have done a bit better... There have been other years like 1992 in which the national polls did a bit better.
If you read between the lines, what this tells you is that state polling looks better only if you choose the right timeframe. Too much old data, and it doesn't look better. Too recent data giving a high weight to 1992, it probably doesn't look better either. In other words, it sounds to me like Nate Silver basically went p-value fishing until he found a timeframe that validates his gut feeling and chosen methodology that state polls are better than national.

Now if you look at national polls, what you find is that the race is basically a tie.

The thing is that as well as random statistical noise, all polls are prone to systematic error. For instance in the UK 1992 election the Shy Tory Factor biassed all the polls: people just weren't willing to tell the pollsters they planned to vote for the nasty party. The polls on average were a massive 8.5% out. In the last UK election, the polls overestimated Lib Dem support by about 5% on average.

So, with close polls and the possibility of large systematic errors, I don't think it's possible to be very confident that either Romney or Obama is securely ahead.

However, it seems to me that there are a couple of factors that give Romney an advantage that might not be seen in most polls. The more a poll weights towards likelihood to vote, the better he does: I'm not confident the Obama waverers will bother to turn out. Also I suspect given the opprobrium towards Romney there could well be a "Shy Republican Factor": people who won't admit they want to vote for the callous 53%er but will knuckle down in the end.

I'm not at all convinced that the superstorm and late endorsements will change the race: it's remained remarkably stable up to now in the face of current events. I can't see many voters changing their mind now.

Therefore, I'm giving Romney a slight advantage in the race, though I do think it's very close.
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?

[ Parent ]
Shy voters by lm (4.00 / 1) #12 Fri Nov 02, 2012 at 04:56:12 AM EST
I quite like your "shy voter" point but I think it applies mostly to Obama. That is to say, people are embarrassed that they're going to vote for Obama when polled but will still punch the button for Obama in the privacy of the voting booth.

There is no more degenerate kind of state than that in which the richest are supposed to be the best.
Cicero, The Republic
[ Parent ]
Possibly by TheophileEscargot (2.00 / 0) #13 Fri Nov 02, 2012 at 05:44:00 AM EST
But it was the right wing side it affected in the past.
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?
[ Parent ]
Maybe in the UK by lm (4.00 / 1) #17 Fri Nov 02, 2012 at 06:14:29 AM EST
Nate Silver has looked into that question regarding US elections before and, while left and right do not always map well onto Democratic and Republican presidential candidates, the evidence suggests that Democratic and Republican candidates both outperform their poll numbers.

http://fivethirtyeight.blogs.nytimes.com/2012/09/29/poll-averages-have-no-history-of-consistent-partisan-bias/

The most notable election was 1980 when Reagan outperformed his state level polls by almost 9%. But elections like that are rare, more often we're talking less than 2%.


There is no more degenerate kind of state than that in which the richest are supposed to be the best.
Cicero, The Republic
[ Parent ]
Just what makes you by wumpus (4.00 / 2) #27 Fri Nov 02, 2012 at 04:48:06 PM EST
think that Obama voters are more likely to answer a call with an unknown/blocked phone number and an English only robospeil? Plenty of polls simply require landlines. These are ways to poll people that are dying out, and it is a stretch to claim they favor Republicans. Rightwing blather rationalizes a lot, but this one is really stretching it.

The electoral college appears to favor Obama (even though the big empty states are all red). Romney might need more than a 1% advantage to take Ohio.

Wumpus

[ Parent ]
That would be me. by ammoniacal (4.00 / 2) #20 Fri Nov 02, 2012 at 09:28:41 AM EST
I'm an anecdatum.

"To this day that was the most bullshit caesar salad I have every experienced..." - triggerfinger

[ Parent ]
So if we expand the timeframe by lm (4.00 / 1) #16 Fri Nov 02, 2012 at 06:02:12 AM EST
Back to 1972 -- using Silver's numbers -- we get an average state poll bias of 1.85 and an average national poll bias of 2.11.

So I think his modest conclusion is sensible, ``But the historical evidence weighs in slightly more heavily on behalf of the state polls ... when they seem to contradict the national ones.''


There is no more degenerate kind of state than that in which the richest are supposed to be the best.
Cicero, The Republic
[ Parent ]
national vs. state polls by ucblockhead (4.00 / 1) #21 Fri Nov 02, 2012 at 10:32:01 AM EST
Note that the reason to look at state polls isn't that state polling is statistically better, but that the popular vote doesn't matter.  If state polls are exactly as accurate as national polls, then you'd ignore national polls completely because the national vote totals have no impact on victory.

I don't think it's "gut feeling" so much as a recognition that state polling has greatly increased in size, and polling in general has increased in accuracy.

I think if there were a "shy Republican factor", we'd have seen it in 2008.

---
[ucblockhead is] useless and subhuman

[ Parent ]
Popular vote doesn't matter by TheophileEscargot (2.00 / 0) #23 Fri Nov 02, 2012 at 11:27:04 AM EST
So I'm always told...

http://www.hulver.com/scoop/story/2004/11/2/04046/8003
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?

[ Parent ]
Yes... by ucblockhead (4.00 / 1) #24 Fri Nov 02, 2012 at 11:44:09 AM EST
I made the mistake of thinking the conventional polling wisdom was wrong.

But that's not what I meant.  The popular vote literally doesn't matter.  It's the person with the most electoral votes that wins, and those are decided by individual state races.

If you have two sets of polls, one saying that Romney is ahead nationwide, and another that says that Obama will win based on state polls, both sets can be completely accurate without conflicting.  In such a case, Obama wins.

National polls only matter if they are more accurate than state polls.
---
[ucblockhead is] useless and subhuman

[ Parent ]
Specific comment: by TheophileEscargot (2.00 / 0) #25 Fri Nov 02, 2012 at 11:53:24 AM EST
http://www.hulver.com/scoop/comments/2004/11/2/04046/8003/5#5
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?
[ Parent ]
Yes, it is correlated by ucblockhead (4.00 / 1) #26 Fri Nov 02, 2012 at 01:55:42 PM EST
But again, if state polls are as accurate as the national polls, then they either won't conflict, or if they do conflict, then they are predicting one of the rate popular vote/electoral vote splits.  The national polls should only be preferred if you have reason to think the state polls are less accurate.

In the words of Nate Silver: Mr. Obama is leading in the polls of Ohio and other states that would suffice for him to win 270 electoral votes, and by a margin that has historically translated into victory a fairly high percentage of the time.

Also remember that the splits are most likely when you've got a large number of states where the winner is a given as you can have big popular vote swings in states like California, New York and Texas that have absolutely no electoral vote effect.  That's exactly what happened in 2000, and this polarization still very much exists.  In fact, it may be worse now.

(Though the point may be moot as Obama's leading in the national polls as well.)
---
[ucblockhead is] useless and subhuman

[ Parent ]
Well by TheophileEscargot (2.00 / 0) #28 Sat Nov 03, 2012 at 12:19:23 PM EST
There's a subtle difference between relying on a state poll, and relying on state polls. The decisions of which states to poll most often is a human one. So, I think it's possible for it to miss changes when a state, demographic or within-state community shifts from one party to another.

Also as I said, it means you don't get to weigh up the different methodologies of different pollsters, since you're relying a random grab-bag of methodologies.

Also the curse of the black swan is that by ignoring the tail-end risk (e.g. a safe state changing, a house price collapse) an algorithm can actually look deceptively good when there isn't a black swan event. The Black–Scholes model produced excellent predictions most of time, but when it did fail, it failed big.

Finally, since I've always used national polls before, if I did switch to state polls now, it could be because I subconsciously prefer their results. I couldn't trust my decision to be unbiased.
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?

[ Parent ]
pretty sure silver is aware of this kind of stuff by the mariner (2.00 / 0) #29 Sat Nov 03, 2012 at 12:38:51 PM EST
also, the odds of states like california, new york, massachusetts, etc. going for romney are a lot smaller than 1 in 100.

romney looks good in the national vote because his party wins by large margins in the old confederacy and some states in the west. the problem is that these places, even though they'll sometimes go 70 percent republican, don't have that many people compared to the coastal and great lakes states that have large urban population centers and vote democrat, so while his big wins in smaller states run up his national numbers, obama's lower margin advantage in more populated states run up the electoral votes in the primarily winner takes all electoral college system -- which is all that actually matters in a presidential election. 

this is all just to say that political analysis in the united states generally focuses on state level races and polls (so-called battleground states) for good reason.


[ Parent ]
Please just read this by ucblockhead (2.00 / 0) #30 Sat Nov 03, 2012 at 01:09:23 PM EST
Nate Silver says it better than me.  The only rational reason to expect Romney to win is if you think the state polls are biased in Obama's direction.  That's not in itself irrational, and Silver essentially puts the chance of that being the case at a bit under 16%.

Note that all states are polled.  You don't hear about polls from states like California (Obama +16.5) or Texas (Romney +15.3) because candidates have double digit leads there and victory is almost a certainty.  This is the case with 40 out of 50 states.   The states are polled, the results just aren't reported.  Dog bites man.  There is absolutely no "human decision of what states to poll".    The statement "For Romney to Win, State Polls Must Be Statistically Biased" is based on polling of all 50 states.  It's also a stretch to say that averaging different methodologies invalidates the numbers when 19 out of 20 swing state polls over the last week show Obama in the lead.

"Black Swans" are inherently unpredictable.  Certainly, it makes sense to not be surprised if Romney wins for a reason that is later explainable.  Certainly saying that he only has a 1 in 6 chance of winning allows for this.  But you are saying you think it is more likely to win, which seems tantamount to saying that you think a Black Swan is likely.  That's not a "black swan".
---
[ucblockhead is] useless and subhuman

[ Parent ]
*Sigh* by TheophileEscargot (2.00 / 0) #31 Sat Nov 03, 2012 at 01:23:42 PM EST
Black swan events aren't why I think Romney is slightly more likely to win. They're why I prefer national polls to an aggregation of the states that people have thought worth polling, which is what you asked me.

It's happened only three times in history that the electoral college winner hasn't got a majority of the vote share, and that was usually by less than half a percent. It's not going to happen that one candidate gets 3% more in the popular vote, and lose the electoral college.

So, if there's a significant difference between national polls and an aggregation-of-states-thought-worth-polling, I think them most likely reason is something to do with the thoughts-of-whats-worth-polling.
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?

[ Parent ]
"What's worth polling" by ucblockhead (4.00 / 2) #32 Sat Nov 03, 2012 at 02:34:57 PM EST
How is "what's worth polling" at the state level any different from the national level?  I've never seen any evidence that the methods are different.

Three times in history the electoral college winner hasn't got a majority of the popular vote share.  Zero times in history the electoral college winner hasn't had a majority of the electoral vote share.  State polls measure the latter.  National polls, the former.

You also seem to have completely missed the point about state polling.  State polling happens in all fifty states!  The last poll in California was on 10/28, and has Obama up 23.  The last poll state in Alabama was on 10/14 and has Romney up by 18.5%.  It should be obvious why no one in the media bothered to report either result, but it doesn't mean that the states aren't being polled!  These states aren't labeled "safe Obama" or "safe Romney" based on someone's opinion.  They are labeled so based on polling results.

Also, the national polls aren't remotely showing Romney 3% ahead or even saying he's ahead at all!  As of last Tuesday's polls, Obama was leading in five out of eight of them.  So to say Romney is winning, you also have to say that Nate Silver (and others) are wrong in their national vote predictions as well.  (Silver currently has Obama leading by 2% of the vote, based on national polling numbers.)

What polls are you basing your predictions on?
---
[ucblockhead is] useless and subhuman

[ Parent ]
especially since... by gzt (4.00 / 1) #33 Sat Nov 03, 2012 at 03:13:14 PM EST
...people choose the poll the states with tighter races. all 50 states get polled, the results for CA and AL aren't interesting so they may get polled less frequently, but people poll IA and OH a lot because they're tossups.

[ Parent ]
But by TheophileEscargot (2.00 / 0) #35 Sun Nov 04, 2012 at 04:12:44 AM EST
There's a human element in the decisions of what to poll, that could miss a safe state turning the other way. It's so complicated to work out the chances of that happening, compounded with the possibilities like a "Shy Republican" effect skewing all polls, that I never trust aggregations of state/constituency polls.

 They give an illusion of certainty because you can't just look at them and think "Ok, now what if a Shy Whatever Factor is actually shifting things 3% one way".

But they could well be right, national polls are basically a tossup right now.
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?

[ Parent ]
sure, that's possible. by gzt (4.00 / 1) #36 Sun Nov 04, 2012 at 07:19:47 AM EST
So what kind of probability do you assign to that kind of event?

[ Parent ]
I don't by TheophileEscargot (2.00 / 0) #37 Sun Nov 04, 2012 at 09:13:18 AM EST
With a national poll, if they're doing a reasonable job of randomly getting people across the US, the safe-state-turning uncertainty is included in the standard error calculations.

With aggregations-of-state-polls, that uncertainty isn't dealt with at all, and it can blow up in counterintuitive ways.

E.g. suppose there are 40 safe states, all equally safe (ha ha) with a probability of x that they go for the expected party. The probability y of at least one safe state going the other way is 1 - x^40.

Plug it into a spreadsheet:

x y
0.999 0.0392297893
0.9975 0.0952758684
0.995 0.181679879
0.99 0.3310282414
0.98 0.554299596
0.97 0.7042877126
0.96 0.8046338484
0.95 0.8714878434

It blows up wildly if you do even a slightly worse job this election than in a previous election. Even at pretty low chances it's significant.

It's also hard to correct for from looking at historical data. Suppose x is 0.9975 and y is 0.095. It might not have happened for forty years or so, but there's still nearly a 1 in 10 chance it will happen this year. Nate Silver probably has an estimate for x based on historical data, but because these events are so rare, that estimate must be based such a tiny number of data points, possibly even zero data points. So however brilliant a statistician he is, I don't trust that he has enough data to be as certain as he claims.

Now try looking at "shy tory" or "shy republican" effects on top of that. It seems likely that those effects are going to be more prevalent in a swing state than a safe state. An Ohian is likely to be more shy about admitting he's for Romney than a Texan.

So, you can't really account for those in a state poll. With a national poll, at least you can say "historically Shy X effects have shifted the vote by Y%" and get a feel for how it relates to the margin of error.

So basically, I think the uncertainty is more transparent in national polls. In aggregations of state polls the uncertainties are more opaque, so it's easier to get a false sense of confidence.
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?

[ Parent ]
what i might do this afternoon by gzt (2.00 / 0) #38 Sun Nov 04, 2012 at 09:55:33 AM EST
look at the current polling for each state, bump it a couple points right, use the uncertainties to make a model, run 10000 simulations with each state independent, let you know the answer. i might also do it with some covariance structure on the states, but that would be more work, might take more than one afternoon because I'd have to toy with things.

I mean, yeah, I'm picking up what you're putting down. that right there is part of why Silver is at 84% or whatever instead of 100%. the "shy" whatever effects, sure, there are reasonable concerns that perhaps there are systematic errors in state polls that may or may not be present in national polls, and Silver has admitted as much, that if the state polls are systematically biased, he's boned.

[ Parent ]
84% vs. 100% by ucblockhead (2.00 / 0) #40 Sun Nov 04, 2012 at 11:44:26 AM EST
I believe that it comes down to margins of error.  If a poll aggregate is +2% Obama with a 3% margin of error, Romney will win in some simulations.  If these errors are truly random over all states, Obama will win.  Romney can only win if all the errors are in one direction.

I think he also has a fudge factor based on variance between past polls and election results.

I find the idea of a "shy Republican" effect a bit strange, though.  There was certainly no such effect in 2008.  If anything, Obama beat his polling numbers.  (Though I suppose Bush did outperform his polls by about a percentage point in 2004.  Romney would need to do better than that at this point though.)  My sense is that Republicans in general aren't particularly shy about stating their views even in safe blue states like mine.
---
[ucblockhead is] useless and subhuman

[ Parent ]
Basing it on by TheophileEscargot (2.00 / 0) #34 Sun Nov 04, 2012 at 03:49:31 AM EST
RCP's list of national polls. That still shows it as basically a tie. As I type 4 ties, 2 Romney, 4 Obama. Close enough that I think systematic errors and turnout are important factors: there's no consensus in the national polls.

But it's very close,  the cherries you've picked might well happen to be pointing the right way.
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?

[ Parent ]
Cherries? by ucblockhead (2.00 / 0) #39 Sun Nov 04, 2012 at 11:30:45 AM EST
How is using all state polls "cherry picking"?  I am not "cherry picking".  People who are predicting an Obama win as more likely (which is nearly everyone including Real Clear Politics are generally the ones using all of the data.

If I am "cherry picking", what am I ignoring?

But yes, for Romney to win, there have to systematic errors in state polling.

National polling, of course, doesn't really tell us who is going to win.  A statistical tie at the national election is all the more reason to look at state polls because state polls, using the same methodologies, do not show a statistical tie in the electoral college.
---
[ucblockhead is] useless and subhuman

[ Parent ]
in fairness, by the mariner (2.00 / 0) #41 Sun Nov 04, 2012 at 04:40:10 PM EST
if i were to try to comment on british politics, i'd probably have no idea what's going on. 

[ Parent ]
neither did nate silver by gzt (2.00 / 0) #42 Sun Nov 04, 2012 at 08:14:20 PM EST
IIRC, his attempts at picking british election results were a dismal failure.

[ Parent ]
Ohio and Colorado by jimgon (4.00 / 1) #15 Fri Nov 02, 2012 at 05:53:01 AM EST
I think you are right about Ohio being the technical end of the election night, but the election won't be called until Colorado is done.  Colorado is the only westernish state in play.




---------------
Technician - "We can't even get decent physical health care. Mental health is like witchcraft here."
[ Parent ]
Probably even before than by ucblockhead (4.00 / 1) #22 Fri Nov 02, 2012 at 10:35:56 AM EST
If the four earlier "swing states" (New Hampshire, Virginia, North Carolina, Florida) go as the pollsters expect, it means there's little systematic poll bias, and Ohio will likely go Obama.

If Either North Carolina or Florida go Obama, it means that the polls were biased against Obama, and we're looking  at an easy victory.

If either Virginia or New Hampshire go Romney, then the polls were biased against Romney, and he has a real shot.
---
[ucblockhead is] useless and subhuman

[ Parent ]
The EU by dmg (4.00 / 1) #5 Thu Nov 01, 2012 at 06:42:37 PM EST
Represents bad value for money for the British taxpayer.

It is also anti democratic.

I am against the UK's ongoing membership and would vote "out" were the option presented to me, as would a majority of UK citizens.

But since we will never be given that option I guess there's no point in getting my hopes up.
--
dmg - HuSi's most dimwitted overprivileged user.

Anti-democratic? by TheophileEscargot (4.00 / 2) #11 Fri Nov 02, 2012 at 04:46:06 AM EST
I thought democracy was a sham and we were all supposed to refrain from voting? When did British democracy become a real thing?
--
It is unlikely that the good of a snail should reside in its shell: so is it likely that the good of a man should?
[ Parent ]
It is a sham by dmg (4.00 / 1) #19 Fri Nov 02, 2012 at 07:21:46 AM EST
The EU takes it to even higher levels.
--
dmg - HuSi's most dimwitted overprivileged user.
[ Parent ]
Romney over Obama by ucblockhead (4.00 / 1) #7 Thu Nov 01, 2012 at 09:45:28 PM EST
Seems unlikely.  What do you base it on?
---
[ucblockhead is] useless and subhuman
I don't know if it is unlikely by lm (4.00 / 1) #18 Fri Nov 02, 2012 at 06:39:18 AM EST
Romney is no Reagan but, when handicapping the election, we would all do well to remember that sometimes the polls are off by as much as 10%

My sense is that the odds do lean in Obama's favor but I would not be all that surprised  if Romney won.


There is no more degenerate kind of state than that in which the richest are supposed to be the best.
Cicero, The Republic
[ Parent ]
I'm predicting Romney by dev trash (4.00 / 1) #9 Thu Nov 01, 2012 at 11:30:24 PM EST
yes!  and then 9 months in when Obamacare is still around and taxes are the same or even higher, well.  I'll laugh.  I'm voting third party this year.

--
I DON'T CARE ABOUT YOUR BALLS! ->clock
It's raining shoemakers' apprentices | 42 comments (42 topical, 0 hidden) | Trackback