PURPLE Poll for CO FL IA MN NC NH NM NV OH PA VA WI
       |           

Welcome, Guest. Please login or register.
Did you miss your activation email?
March 28, 2024, 10:24:18 AM
News: Election Simulator 2.0 Released. Senate/Gubernatorial maps, proportional electoral votes, and more - Read more

  Talk Elections
  Election Archive
  Election Archive
  2012 Elections
  2012 U.S. Presidential General Election Polls
  PURPLE Poll for CO FL IA MN NC NH NM NV OH PA VA WI
« previous next »
Pages: [1]
Author Topic: PURPLE Poll for CO FL IA MN NC NH NM NV OH PA VA WI  (Read 4170 times)
Ben Romney
Hillary2012
Jr. Member
***
Posts: 395
Germany


Show only this user's posts in this thread
« on: March 21, 2012, 08:29:59 PM »

PURPLE Poll: Obama: 46/50

CO NV NM:
Obama: 43/51
Obama/ Romney: 45/46
Obama/Santorum: 45/45

IA MN WI:
Obama: 48/47
Obama/Romney: 50/44
Obama/Santorum: 51/41

NH OH PA:
Obama: 44/51
Obama/Romney: 48/43
Obama Santorum: 48/43

FL NC VA:
Obama: 46/51
Obama/Romney: 47/46
Obama/Santorum: 49/42

http://www.purplestrategies.com/wp-content/uploads/MarchPurplePoll12.pdf
Logged
Fuzzybigfoot
YaBB God
*****
Posts: 4,211
United States


WWW Show only this user's posts in this thread
« Reply #1 on: March 21, 2012, 08:39:00 PM »

wut
Logged
The Ex-Factor
xfactor99
Jr. Member
***
Posts: 1,241
Viet Nam


Political Matrix
E: -5.42, S: -6.43

Show only this user's posts in this thread
« Reply #2 on: March 21, 2012, 09:10:45 PM »

This trend of releasing surveys of multiple states (I blame Rasmussen) is pretty dumb.
Logged
Earthling
Jr. Member
***
Posts: 1,129
Netherlands


Show only this user's posts in this thread
« Reply #3 on: March 22, 2012, 04:29:15 AM »

Obama is leading in Florida but trailing in Colorado? Joke poll.
Logged
Eraserhead
Atlas Legend
*****
Posts: 44,405
United States


Show only this user's posts in this thread
« Reply #4 on: March 22, 2012, 06:21:51 AM »

This stupidity has to stop.
Logged
Democratic Hawk
LucysBeau
Atlas Icon
*****
Posts: 14,703
United Kingdom


Political Matrix
E: -2.58, S: 2.43

Show only this user's posts in this thread
« Reply #5 on: March 22, 2012, 08:32:06 AM »

For comparative purposes, Obama vs Romney in 'Purple Poll'

Jan (overall): Obama 46 - Romney 44

Jan (The Wild West: CO, NV, NM): Obama 47 - Romney 42
Jan (The Heartland): IA, MN, WI): Obama 49 - Romney 42
Jan (The Rustbelt: NH, OH, PA): Obama 45 - Romney 43
Jan (The Southern Swing: FL, NC, VA): Obama 44 - Romney 46

Feb (overall): Obama 47 (+1) - Romney 43 (-1)

Feb (The Wild West: CO, NV, NM): Obama 47 (nc) - Romney 44 (+2)
Feb (The Heartland): IA, MN, WI): Obama 48 (-1) - Romney 39 (-3)
Feb (The Rustbelt: NH, OH, PA): Obama 44 (-1) - Romney 46 (+3)
Feb (The Southern Swing: FL, NC, VA): Obama 47 (+3) - Romney 45 (-1)

Mar (Overall): Obama 48 (+1) - Romney 44 (+1)

Feb (The Wild West: CO, NV, NM): Obama 45 (-2) - Romney 46 (+2)
Feb (The Heartland): IA, MN, WI): Obama 50 (+2) - Romney 44 (+5)
Feb (The Rustbelt: NH, OH, PA): Obama 48 (+4) - Romney 43 (-3)
Feb (The Southern Swing: FL, NC, VA): Obama 47 (nc) - Romney 46 (+1)
Logged
King
intermoderate
Atlas Star
*****
Posts: 29,357
United States


Show only this user's posts in this thread
« Reply #6 on: March 22, 2012, 12:26:02 PM »

At least the pairings here make a bit of sense in terms of regional culture. It's not Scott's "Romney leads in Ohio/Oregon/Arkansas" crap.
Logged
Likely Voter
Moderators
Junior Chimp
*****
Posts: 8,344


Show only this user's posts in this thread
« Reply #7 on: March 22, 2012, 12:35:28 PM »

what is the point of these?
Logged
Tender Branson
Mark Warner 08
Atlas Institution
*****
Posts: 58,173
Austria


Political Matrix
E: -6.06, S: -4.84

Show only this user's posts in this thread
« Reply #8 on: March 22, 2012, 12:44:27 PM »

I hope they switch to polling all of these states individually after Romney is declared the nominee.
Logged
Gustaf
Moderators
Atlas Star
*****
Posts: 29,770


Political Matrix
E: 0.39, S: -0.70

Show only this user's posts in this thread
« Reply #9 on: March 24, 2012, 09:55:59 AM »

That seems to hint at a map looking something like this, right:

Logged
Tender Branson
Mark Warner 08
Atlas Institution
*****
Posts: 58,173
Austria


Political Matrix
E: -6.06, S: -4.84

Show only this user's posts in this thread
« Reply #10 on: March 24, 2012, 10:24:34 AM »

That seems to hint at a map looking something like this, right:



You forgot Colorado for Obama, but otherwise yes.
Logged
Gustaf
Moderators
Atlas Star
*****
Posts: 29,770


Political Matrix
E: 0.39, S: -0.70

Show only this user's posts in this thread
« Reply #11 on: March 24, 2012, 10:26:14 AM »

If Romney leads Obama by 1% in CO+NV+NM he must lead in at least one of the 3 states. Looking at the 2008 results it seems reasonable to think that he would be ahead in Colorado. You could potentially make a case that he could also be ahead in Nevada.
Logged
Tender Branson
Mark Warner 08
Atlas Institution
*****
Posts: 58,173
Austria


Political Matrix
E: -6.06, S: -4.84

Show only this user's posts in this thread
« Reply #12 on: March 24, 2012, 10:31:09 AM »

If Romney leads Obama by 1% in CO+NV+NM he must lead in at least one of the 3 states. Looking at the 2008 results it seems reasonable to think that he would be ahead in Colorado. You could potentially make a case that he could also be ahead in Nevada.

Or in New Mexico, because the sample size for NM would be about 100.
Logged
ajb
Jr. Member
***
Posts: 869
United States


Show only this user's posts in this thread
« Reply #13 on: March 24, 2012, 10:37:52 AM »

If Romney leads Obama by 1% in CO+NV+NM he must lead in at least one of the 3 states. Looking at the 2008 results it seems reasonable to think that he would be ahead in Colorado. You could potentially make a case that he could also be ahead in Nevada.

Or in New Mexico, because the sample size for NM would be about 100.
In 2008 the results were
NM: Obama 57 McCain 42
NV: Obama 55 McCain 43
CO: Obama 54 McCain 45

So if he's leading in only one of the three (which I think is unlikely, but is what the poll suggests), it's most likely NM.

Which also makes sense when you look at the % of each state that is Hispanic: NM 46%; NV 27%; CO 20%.
Logged
🐒Gods of Prosperity🔱🐲💸
shua
Atlas Star
*****
Posts: 25,663
Nepal


Political Matrix
E: 1.29, S: -0.70

WWW Show only this user's posts in this thread
« Reply #14 on: March 24, 2012, 10:40:52 AM »

NH a rustbelt state?
Logged
Tender Branson
Mark Warner 08
Atlas Institution
*****
Posts: 58,173
Austria


Political Matrix
E: -6.06, S: -4.84

Show only this user's posts in this thread
« Reply #15 on: March 24, 2012, 10:48:17 AM »

The western subsample is weird anyway.

Recent polls have shown that NV is lean-Obama, and NM is strong-Obama and CO is at least a toss-up.

Which makes this Purple Poll subsample odd, because if we assume that Obama leads in NV and NM by about 10%, Obama would need to trail by about 11% in CO, assuming that the CO-part of the sample has about 250 people and NV has about 150 and NM about 100. But without an additional variable from the poll it's impossible to say who leads in which part.
Logged
Tender Branson
Mark Warner 08
Atlas Institution
*****
Posts: 58,173
Austria


Political Matrix
E: -6.06, S: -4.84

Show only this user's posts in this thread
« Reply #16 on: March 24, 2012, 10:55:47 AM »

The subsets for CO, OH, PA and FL would actually make decent polls.

The description says that "Purple Poll" uses an oversample in each region with a MoE of 4.1%

Using a population of 10 Mio. in the western region, this would lead to a sample of roughly 580.

Because CO has about 50% of the 10 Mio. people, the CO subset alone would have about 300 respondents. 300 people would not be too bad of a poll.

The same is true for OH and PA, with PA having about 300 respondents and OH about 250.

FL would also have 300 respondents in a regional 580 oversample.
Logged
ajb
Jr. Member
***
Posts: 869
United States


Show only this user's posts in this thread
« Reply #17 on: March 24, 2012, 10:59:25 AM »

The subsets for CO, OH, PA and FL would actually make decent polls.

The description says that "Purple Poll" uses an oversample in each region with a MoE of 4.1%

Using a population of 10 Mio. in the western region, this would lead to a sample of roughly 580.

Because CO has about 50% of the 10 Mio. people, the CO subset alone would have about 300 respondents. 300 people would not be too bad of a poll.

The same is true for OH and PA, with PA having about 300 respondents and OH about 250.

FL would also have 300 respondents in a regional 580 oversample.

Wouldn't it make more sense to weight by electoral vote? In which case, CO would be 45% of the sample, NM 30%, NV 25%, for example.
Logged
Vosem
Atlas Icon
*****
Posts: 15,622
United States


Political Matrix
E: 8.13, S: -6.09

Show only this user's posts in this thread
« Reply #18 on: March 24, 2012, 11:41:43 AM »

Here's my attempt to break this poll down into digestible chunks (I'm probably going to go back and do something similar for old Rasmussen polls too):

2008:
Colorado voted Obama 54/44
Nevada voted Obama 55/43
New Mexico voted Obama 57/42

Without weighting for population purposes, these three states voted 55/44 Obama. Therefore, a 46/45 Romney lead implies that

a) 9% of the population is undecided or voting third party; since 1% did that in the real 2008 results, the simplest thing to do is to just remove 4% from both figures; we get a 51/40 Obama lead; and then a uniform 12-point swing either way gets Romney 46/45

Doing this to Colorado, first we must remove the voters that have become undecided (remove 3.5% from each candidate, but .5% means rounding upwards, so just remove 4 from both) gets an Obama lead of 50/40; then the uniform 12-point swing means a 46/44 Romney lead.

Doing this to Nevada, first we must remove the voters that have become undecided (remove 3.5% from each candidate, but .5% means rounding upwards, so just remove 4 from both) gets an Obama lead of 51/39; then the uniform 12-point swing means a 45/45 tie.

Doing this to New Mexico, first we must remove the voters that have become undecided (remove 4% from each candidate); gets an Obama lead of 53/38; then the uniform 12-point swing gets us a 47/44 Obama lead.

2008:
Iowa voted Obama 54/44
Minnesota voted Obama 54/44
Wisconsin voted Obama 56/42

Without weighting for population purposes, these three states voted 55/43 Obama. Therefore, a 50/44 Obama lead implies

a) 6% of the population is undecided or voting third party; since 2% did that in the real 2008 results, the simplest thing to do is to remove 2% from both figures, which gets us a 53/41 result; and then a uniform 6-point swing either way gets a 50/44 Obama lead.

Doing this to Iowa, first we must remove the voters that have become undecided (2% from each candidate); this gets us an Obama lead of 52/42. Then, the uniform six-point swing gets us an Obama lead of 49/45.

Since Minnesota had the same result in 2008 as Iowa, the math would be identical; and would result in a 49/45 Obama lead.

Doing this to Wisconsin, first we must remove the voters that have become undecided (2% from each candidate); this gets us an Obama lead of 54/40. Then, the uniform six-point swing gets us an Obama lead of 51/43.

2008:
New Hampshire voted Obama 54/45
Ohio voted Obama 51/47
Pennsylvania voted Obama 54/44

Without weighting for population purposes, these three states voted 53/45 Obama. Therefore, a 48/43 Obama lead implies that:

a) 9% of the population is undecided or voting third party; since 2% did that in the real 2008 results, the simplest thing to do is to remove 3.5% from both figures (the .5 rounds it up to 4%), which gets us a 49/41 Obama victory. Then, to get to 48/43 Obama, we must remove 1% from Obama's total and add 2% to Romney's total.

Doing this to New Hampshire, first we must remove the voters that have become undecided (4% from each candidate); this gets us an Obama victory of 50/41. Then, we subtract 1% from Obama's total and add 2% to Romney's total; this gets us an Obama victory of 49/43.

Doing this to Ohio, first we must remove the voters that have become undecided (3.5% from each candidate, which rounds up to 4%); this gets us an Obama lead of 47/43. Then, we subtract 1% from Obama's total and add 2% to Romney's total; this gets us an Obama victory of 46/45.

Doing this to Pennsylvania, first we must remove the voters that have become undecided (3.5% from each candidate, which rounds up to 4%); this gets us an Obama lead of 50/40. Then, we subtract 1% from Obama's total and add 2% to Romney's total; this gets us an Obama victory of 49/42.

2008:
Florida voted 51/48 Obama
North Carolina voted 50/49 Obama
Virginia voted 53/46 Obama

Without weighting for population purposes, these three states voted 51/48 Obama. Therefore, a 47/46 Obama lead implies that:

a) 7% of the population is undecided or voting third party; since 1% did that in the real 2008 results, the simplest thing to do is to remove 3% from both figures, which gets us a 48/45 Obama victory. Then, a uniform 2-point swing gets us to a 47/46 Obama victory.

Doing this to Florida, first we must remove the voters that have become undecided (3% from each candidate); this gets us an Obama lead of 48/45. Then, a uniform 2-point swing gets us to a 47/46 Obama victory.

Doing this to North Carolina, first we must remove the voters that have become undecided (3% from each candidate); this gets us an Obama lead of 47/46. Then, a uniform 2-point swing gets us a 47/46 Romney victory.

Doing this to Virginia, first we must remove the voters that have become undecided (2.5% from each candidate; the .5 rounds up to 3%); this gets us an Obama lead of 50/43. Then, a uniform 2-point swing gets us a 49/44 Obama victory.

Results By State Summary:
Colorado: 46/44 Romney
Iowa: 49/45 Obama
Florida: 47/46 Obama
Minnesota: 49/45 Obama
Nevada: 45/45 tie
New Hampshire: 49/43 Obama
New Mexico: 47/44 Obama
North Carolina: 47/46 Romney
Ohio: 46/45 Obama
Pennsylvania: 49/42 Obama
Virginia: 49/44 Obama
Wisconsin: 51/43 Obama
Logged
Bacon King
Atlas Politician
Atlas Icon
*****
Posts: 18,822
United States


Political Matrix
E: -7.63, S: -9.49

Show only this user's posts in this thread
« Reply #19 on: March 24, 2012, 12:46:35 PM »

The numbers for New Colovada and North Floroginia are a bit concerning.

Honestly though, I can't treat a pollster as a serious company if it does useless polls that combine states like this. Or, for that matter, if it's named after a color in all capital letters, as if a first grader was shouting out the word as he grabbed his crayons. Still better than naming your polling company after yourself, though, which ironically seems to be the companies with the worst records out there.
Logged
Gustaf
Moderators
Atlas Star
*****
Posts: 29,770


Political Matrix
E: 0.39, S: -0.70

Show only this user's posts in this thread
« Reply #20 on: March 24, 2012, 12:55:08 PM »

Tender, it seems unlikely that New Mexico would be the least Democratic in the trio. Colorado was the most GOP in 2008 and also the biggest one.

Vosem, you probably have to weight by population for this to make sense. Especially since in most of the trios there are large differences in size between the states.
Logged
Vosem
Atlas Icon
*****
Posts: 15,622
United States


Political Matrix
E: 8.13, S: -6.09

Show only this user's posts in this thread
« Reply #21 on: March 24, 2012, 12:56:36 PM »

Tender, it seems unlikely that New Mexico would be the least Democratic in the trio. Colorado was the most GOP in 2008 and also the biggest one.

Vosem, you probably have to weight by population for this to make sense. Especially since in most of the trios there are large differences in size between the states.

But there's zero indication that the Purple Pollsters are weighting it by population...
Logged
Pages: [1]  
« previous next »
Jump to:  


Login with username, password and session length

Terms of Service - DMCA Agent and Policy - Privacy Policy and Cookies

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

Page created in 0.061 seconds with 16 queries.