17
   

Get yer polls, bets, numbers & pretty graphs! Elections 2008

 
 
nimh
 
  1  
Reply Fri 14 Dec, 2007 05:14 pm
So much for Giuliani's rope-a-dope strategy of hoping that the early states will be divided among two or three different rivals while he hangs back in second or third place waiting to win big time in the February 5 states...

Florida a firewall no more!

Quote:
A new Rasmussen Reports automated survey of 685 likely Republican primary voters in Florida (conducted 12/13) finds former Gov. Mike Huckabee running slightly ahead of former Gov. Mitt Romney (27% to 23%) in a statewide primary; former Mayor Rudy Giuliani trails at 19%
0 Replies
 
nimh
 
  1  
Reply Fri 14 Dec, 2007 10:48 pm
Graphs for the Republican primary race in Iowa according to the same recipe as above:

http://img100.imageshack.us/img100/5482/iarepsonlyminepp5.png

http://img509.imageshack.us/img509/5567/iarepsbothpv5.png
0 Replies
 
nimh
 
  1  
Reply Fri 14 Dec, 2007 10:53 pm
nimh wrote:
So much for Giuliani's rope-a-dope strategy of hoping that the early states will be divided among two or three different rivals while he hangs back in second or third place waiting to win big time in the February 5 states...

Florida a firewall no more!

Quote:
A new Rasmussen Reports automated survey of 685 likely Republican primary voters in Florida (conducted 12/13) finds former Gov. Mike Huckabee running slightly ahead of former Gov. Mitt Romney (27% to 23%) in a statewide primary; former Mayor Rudy Giuliani trails at 19%


Oops, just saw that Cyclo had already posted this one. Interesting nevertheless!
0 Replies
 
nimh
 
  1  
Reply Fri 14 Dec, 2007 11:15 pm
And just for the heck of it - ok, also because it's practical - here's the two Republican graphs next to each other, so you can compare the state of the races in New Hampshire and Iowa with each other directly.

The graph for New Hampshire is also updated compared to the one I posted on Thursday - there's been three new polls out in the past two days.

http://img206.imageshack.us/img206/6314/ianhrepsminerj6.png

Huckabee's rise in Iowa has been truly phenomenal. Like it's taking place in some parallel dimension where the laws of gravity dont hold or something.
0 Replies
 
msolga
 
  1  
Reply Sat 15 Dec, 2007 01:19 am
bm
0 Replies
 
nimh
 
  1  
Reply Sat 15 Dec, 2007 07:29 pm
Survey USA, Rasmussen and Quinnipiac all regularly publish state-level polls that match up one or more of the Democratic candidates against one or more of the Republican candidates. Other pollsters do so much more rarely, or only for the odd state here and there.

These are interesting. National match-up polls, for sure, are cool because they give a good idea of the overall mood in the country and the overall viability of the different candidates: which ones do good not just within their party, but in a match-up against the frontrunners of the other side? But in the end of course the race is going to be decided on a state to state level, and more particularly in a dozen or so swing states.

So I've been tracking how Hillary has been doing in these state-level match-up polls for a few months now. First I recorded only how she did against Giuliani, Thompson and Romney, at the time the three frontrunners. In this post, you can find back two related graphs, for how Hillary did in these match-ups in August/early September and in September/early October. And, well, the results looked very, very good for Hillary.

Now I've started tracking how she does against McCain and Huckabee as well. In the table below, I've paired up the poll results for November/early December from Survey USA (green), Rasmussen (orange), and Quinnipiac (in blue), for each of the hypothetical match-ups: Hillary vs Giuliani, McCain, Romney, Thompson and Huckabee.

The numbers are listed by state. The numbers you see listed are the margins of 'victory' (in blue) or 'defeat' (in orange) for Hillary. For comparison purposes, the margins of victory/defeat for Kerry against Bush are given on the left.


http://img149.imageshack.us/img149/2037/hillarystatematchupsdecms0.png


I think these results are interesting on two levels.

On the one hand, they make visible just how large a shift in political preference is appearing on the horizon - very tentatively, of course - and for now, anyhow - if, at least, the Republicans don't see the light and nominate John McCain.

On the other hand, it shows up striking differences in how the Republican candidates stack up against Hillary.

From the previous tables it was already clear that Giuliani does much better than Fred Thompson, and Romney does outright disastrously. This new table shows Huckabee doing worse still than Romney (for now!); and it shows McCain being clearly the most competitive Republican candidate, which makes it so poignant that the Republicans show little sign of warming up to him yet, outside New Hampshire.

Now I've been very inclined to attribute these differences in how the candidates stack up to name recognition. That doesnt hold up for Giuliani and McCain, because both of those are widely known, so the difference between them would appear to simply show that McCain really has the greater appeal to independents and moderates. But it would, logically, hold for Romney.

Romney is simply much less known than Giuliani or McCain, and thus people are less likely to express a preference for him. Instead, both the number of undecideds and of those choosing Hillary over this unknown quantity are higher, and thus Hillary has a much more positive margin over him than over Rudy or McCain. Once Romney becomes better known too, his numbers will automatically pick up.

That was the theory. But I dont put half as much stock in that explanation anymore, for two reasons:

  • Romney has campaigned for over half a year now, and has attracted a fair share of media attention over time. But his numbers show little sign of improving at all. In the August/early September polls, his margins against Hillary were on average 16 percentage points worse than Bush's 2004 margins against Kerry had been. In September/early October, he was doing 15 points worse. In October/early November, it was still 15 points. Now, it's 12 points. There's hardly any improvement at all, despite the cumulative effect of the media coverage that he's gotten would have had on voters' familiarity with him.

  • Huckabee started out just as unknown as Romney. As a logical consequence of his relative unknownness, his margins against Hillary in this table are on average 15 percentage points worse than Bush's 2004 margins against Kerry, and worse still than Romney's. But unlike Romney, Huckabee is picking up in these match-up polls very quickly. Take Ohio: last month, as this table shows, the Survey USA poll saw Hillary beating Huckabee by 17 points. Now there have been 3 new polls this month already that are not included in this table yet: a new Quinnipiac poll on New Jersey, and new polls by both Survey USA and Quinnipiac on Ohio. And in those Ohio polls, Hillary now leads Huckabee by just 2 points according to Survey USA, and 7 points according to Quinnipiac. That's a big improvement upon November. So Huckabee's numbers in these match-ups apparently do seem to quickly pick up as his campaign catches the limelight. Which makes Romney's inability over the months to budge his standing against Hillary in much any way look a lot more meaningful.
All of which confirms my hope that Romney will be the Republican nominee...
0 Replies
 
nimh
 
  1  
Reply Sun 16 Dec, 2007 09:46 am
Crossposting this great post of Old Europe's in the Republicans thread, which brings together the best of all those pretty maps we sought out in 2004 in one post:

old europe wrote:
okie wrote:
Bush's base was almost the entire country, see the map.

http://www.usatoday.com/news/politicselections/vote2004/countymap.htm



Wasn't.

It's deceptive to only look at a map colored according to counties that went to a specific candidate:

http://www-personal.umich.edu/~mejn/election/countymapredblue.png


It's a diagram, and it doesn't take into account population density or the percentages of people voting for a candidate. In leaving out that data, it's giving you a skewed image.


Looking at percentages of people voting for the respective candidates by county, the map would look like this:

http://www-personal.umich.edu/~mejn/election/countymaplinear.png


However, that still doesn't account for population density. Now, there are ways of integrating that data into a visualisation, too. Here's one:

http://i17.tinypic.com/6pr19as.jpg

(displaying population density as heights proportional to votes per area)


And here's another one, called a cartogram:

http://www-personal.umich.edu/~mejn/election/countycartlinear.png

(rescaling the sizes of the counties according to their population)


(more stuff here and here)
0 Replies
 
nimh
 
  1  
Reply Sun 16 Dec, 2007 05:50 pm
OK, so here's my latest gimmick (one day, I will outdo Peter Snow).

Two posts above, I posted an overview of the results of state-level match-up polls in November/early December.

Like I said, I've been tracking those since August. I already gave the link for the tables of August/early September and September/early October; meanwhile, in this post from a couple weeks back on the Hillary etc thread, you can find a table with the results for October/early November (the top one).

OK, now on the basis of these state-level polls, what would the electoral map look like if elections were held tomorrow, and the candidates were called Hillary Clinton and Rudy Giuliani? How would the Electoral College be made up?

Leaving states where polls pointing to either of the two as winner balance each other out blank, giving extra importance to the more recent polls, and extrapolating the result for states that have not been polled at all yet on the basis of how states where Bush fared comparably in '04 are showing up on the polls, this would be the result:


http://img89.imageshack.us/img89/4094/hillarygiulianimap12120oo9.png


I've used the interactive make-your-own-map thingie at www.270towin.com to make this map. It's a cool thingie.

Meanwhile, this would be what the map and numbers look like if the race was between Hillary and Romney, and was held this week:


http://img262.imageshack.us/img262/5735/hillaryromneymap121207xf3.png


Those wondering about just how much worse Romney stacks up I refer to my post two points up, in particular the text that followed the table.

I made these maps on 12 December, but nothing has changed since.
0 Replies
 
Cycloptichorn
 
  1  
Reply Tue 18 Dec, 2007 07:34 pm
Man, Clinton spanks Romney!

Iowa Electric Market for Dems:

http://images.dailykos.com/images/user/363/DConv08.png

Wow, looks like Hillary is going to be the big winner. How well do these things work, anyways? Let's look at 04:

http://images.dailykos.com/images/user/363/dconv04.png

Wuh, yeah. So, I guess it's safe to say that pre-Iowa polls are pretty useless and don't represent the bounce that Iowa winners get.

At this point I don't see Hillary getting the nomination at all.

Cycloptichorn
0 Replies
 
nimh
 
  1  
Reply Wed 19 Dec, 2007 01:59 pm
Cycloptichorn wrote:
Wuh, yeah. So, I guess it's safe to say that pre-Iowa polls are pretty useless and don't represent the bounce that Iowa winners get.

Well, those are not polls - it's like a political stockmarket that enthusiasts like us can speculate on..

But yeah, doesnt change the point much, for sure... I rarely look at who's ahead in the national horserace polls on the primaries anymore. I look at the polls for the early states, and what the national polls say about the balance between the Democrats and Republicans, which Democrat does best vs the Republicans, what issues are paramount, that kind of thing. Who's ahead in the primaries in national polls seems mostly a function of who's getting the most national media coverage + prior familiarity.

Not that the polls on the individual early primary states are necessarily mirrored in the end result.. they just show what the mood is right at the moment.

The only thing I've recently picked up on from the national primary polls is how quickly Huckabee has picked up and surged into second place, even though he was just as unknown as Romney originally was - whereas none of Romney's lengthy front-runnerdom ever lifted him out of the low tens. So I guess that says something about their comparative likability. Perhaps.

That was a great juxtaposition of graphs for that electronic market that you posted there, anyway... sure does make one relativate!
0 Replies
 
nimh
 
  1  
Reply Wed 19 Dec, 2007 05:58 pm
Check out the current chart at pollster.com for the Republican primary race in Iowa.

Who would have ever thought that, just two weeks before the caucuses, this would be the ranking of the Republican candidates Exclamation

1. Mike Huckabee
2. Mitt Romney
3. Fred Thompson
4. Rudy Giuliani
5. Ron Paul
6. John McCain
0 Replies
 
okie
 
  1  
Reply Wed 19 Dec, 2007 06:12 pm
nimh wrote:
Crossposting this great post of Old Europe's in the Republicans thread, which brings together the best of all those pretty maps we sought out in 2004 in one post:

old europe wrote:
okie wrote:
Bush's base was almost the entire country, see the map.

http://www.usatoday.com/news/politicselections/vote2004/countymap.htm



Wasn't.

It's deceptive to only look at a map colored according to counties that went to a specific candidate:

http://www-personal.umich.edu/~mejn/election/countymapredblue.png


It's a diagram, and it doesn't take into account population density or the percentages of people voting for a candidate. In leaving out that data, it's giving you a skewed image.


Looking at percentages of people voting for the respective candidates by county, the map would look like this:

http://www-personal.umich.edu/~mejn/election/countymaplinear.png


However, that still doesn't account for population density. Now, there are ways of integrating that data into a visualisation, too. Here's one:

http://i17.tinypic.com/6pr19as.jpg

(displaying population density as heights proportional to votes per area)


And here's another one, called a cartogram:

http://www-personal.umich.edu/~mejn/election/countycartlinear.png

(rescaling the sizes of the counties according to their population)


(more stuff here and here)

I looked at the second map again and I think an objection is in order. It is a nice trick, to make Democrats look better geographically, because if you look at counties where Republicans narrowly won, mixing blue with red makes the color look more like blue than red, when in reality that is not a true rendition. For example, 53% red with 47% blue produces a shade closer to blue than red in the mind's eye. Just an opinion, but I prefer the first map. And the third map is not pertinent because what a map shows is geographic support, not popular vote support.
0 Replies
 
okie
 
  1  
Reply Wed 19 Dec, 2007 06:15 pm
I will post the following site because it tends to show a reversal of Huckabee's burst of support has possibly reversed and Romney is possibly upswinging again.

http://www.realclearpolitics.com/epolls/2008/president/ia/iowa_republican_caucus-207.html
0 Replies
 
nimh
 
  1  
Reply Wed 19 Dec, 2007 06:27 pm
On a related note, this week saw the first poll of Republican voters in Iowa that actually had Giuliani down in sixth position.

It was an Insider Advantage poll, done on 16-17 December, and it had Romney at 28%, Huckabee at 25%, Thompson at 11%, McCain at 7%, Ron Paul at 6% and Giuliani at 5% - just 1 point ahead of Tom Tancredo.

For the moment it's the only poll yet that has him that low. But there might soon be more, since he's certainly trending downward rapidly in the state (and not only there).

Similarly, the Insider Advantage poll had a result for the Democrats in Iowa that was roughly in line with the trends, but in an exaggerated enough looking way to make for an eye-catching result: it was the first poll since mid-August that had Hillary Clinton down in third place, below both Obama and Edwards.

Again, that's not where the other polls have her yet, though, and in this case it's unlikely to be replicated a lot soon: Hillary is trending down and Edwards up, but both in a fairly gradual way, and the difference between them is about 4 or 5 points on average.
0 Replies
 
nimh
 
  1  
Reply Wed 19 Dec, 2007 06:47 pm
okie wrote:
I will post the following site because it tends to show a reversal of Huckabee's burst of support has possibly reversed and Romney is possibly upswinging again.

http://www.realclearpolitics.com/epolls/2008/president/ia/iowa_republican_caucus-207.html

Thanks! Interesting, that.

The sudden downward trend for Huckabee in that rolling average is basically down to two new polls: the Insider Advantage poll that has him only at 25%, trailing Romney by 3; and the new Rasmussen poll, which has him at 28%, above Romney by just 1.

The Rasmussen one is the more interesting one. Especially in Iowa, where it's very difficult to estimate who will actually turn out to caucus, different pollsters using different methods to identify likely caucusgoers can get very different results. But Rasmussen regularly polls Iowa, so here we can compare apples with apples, by comparing its new numbers with what it found the last couple of times:

Code:Rasmussen - Iowa - Republicans

Date Romney Huckabee Thompson Giuliani Paul McCain

12/17 27 28 8 8 6 14

12/10 23 39 8 8 5 6

11/26-27 25 28 11 12 5 4

11/12 29 16 14 15 4 6



The McCain upsurge could be a fluke, no other poll is showing it, though one is tempted to credit the Des Moines Register endorsement for it. The downward trend for Giuliani and Thompson is as expected, they might be hitting their 'floor' now. But the interesting bit is what Okie said: has Huckabee peaked and is he going downward again, while Romney is recovering somewhat as a result? Are some Iowan Republicans going back to their longstanding frontrunner after rethinking Huckabee as flavour of the day?

It's possible.. but it will take a few more polls to really tell; two new polls is very little to go on yet.
0 Replies
 
nimh
 
  1  
Reply Wed 19 Dec, 2007 06:50 pm
okie wrote:
I looked at the second map again and I think an objection is in order. It is a nice trick, to make Democrats look better geographically, because if you look at counties where Republicans narrowly won, mixing blue with red makes the color look more like blue than red, when in reality that is not a true rendition. For example, 53% red with 47% blue produces a shade closer to blue than red in the mind's eye.

To my eyes, those evenly divided counties just look purple on the second map.. is purple "a shade closer to blue than red"? :wink:
0 Replies
 
okie
 
  1  
Reply Wed 19 Dec, 2007 07:38 pm
nimh, it appears the Rasmussen poll shows that perhaps McCain is mostly capitalizing from the Huckabee drop from his possible peak, and Romney to a less extent, but McCain has a long way to go to win. And also, I look at the numbers and wonder if the sort of undecided or weakly decided voters went from Thompson and Giuliani to Huckabee, but now that more about him is being known, their support was sort of soft, and now they are drifting into McCain's or Romney's camp.

On the color subject, I agree, perception of color is debatable, but anyway I still prefer the either red or blue to clearly show which counties were either won or lost by either party, in a geographic fashion. The other renditions return the perception to showing the roughly 50/50 popular vote. I think we should be reminded however that the founders thought representation by state instead of population was an important part of the representative government by the simple virtue of the fact that there are 2 senators from each state, regardless of size or population, and also by virtue of the fact that the electoral college was instituted, as a part of electing the president.
0 Replies
 
old europe
 
  1  
Reply Wed 19 Dec, 2007 07:43 pm
okie wrote:
I think we should be reminded however that the founders thought representation by state instead of population was an important part of the representative government by the simple virtue of the fact that there are 2 senators from each state, regardless of size or population, and also by virtue of the fact that the electoral college was instituted, as a part of electing the president.


In that case, you should also have objections against a map that shows majorities by county.
0 Replies
 
okie
 
  1  
Reply Wed 19 Dec, 2007 07:46 pm
Okay, I concede your point.
0 Replies
 
old europe
 
  1  
Reply Wed 19 Dec, 2007 07:47 pm
Here's the map that you should prefer, according to what you just posted:

http://www-personal.umich.edu/~mejn/election/statemapredblue.png


Then again, you might want to account for the effects of the electoral college, going by what you said - which would leave you with this cartogram of the States:

http://www-personal.umich.edu/~mejn/election/statecartcollege.png
0 Replies
 
 

Related Topics

 
Copyright © 2024 MadLab, LLC :: Terms of Service :: Privacy Policy :: Page generated in 0.08 seconds on 11/21/2024 at 07:36:59