2012 White Obama Lovers Map by County Project: COMPLETE!
       |           

Welcome, Guest. Please login or register.
Did you miss your activation email?
March 28, 2024, 04:26:02 AM
News: Election Simulator 2.0 Released. Senate/Gubernatorial maps, proportional electoral votes, and more - Read more

  Talk Elections
  General Politics
  Political Geography & Demographics (Moderators: muon2, 15 Down, 35 To Go)
  2012 White Obama Lovers Map by County Project: COMPLETE!
« previous next »
Pages: 1 2 3 4 5 [6] 7 8
Author Topic: 2012 White Obama Lovers Map by County Project: COMPLETE!  (Read 51266 times)
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #125 on: April 28, 2015, 10:41:10 PM »

One big update/dump with all remaining NE states added!

VT 2012 statewide share of the white vote for Obama: 66.3%
RI 2012 statewide share of the white vote for Obama: 60.2%
MA 2012 statewide share of the white vote for Obama: 57.8%
ME 2012 statewide share of the white vote for Obama: 55.5%
CT 2012 statewide share of the white vote for Obama: 52.6%
NY 2012 statewide share of the white vote for Obama: 51.2%
NH 2012 statewide share of the white vote for Obama: 50.3%
Logged
Bandit3 the Worker
Populist3
Junior Chimp
*****
Posts: 7,951


Political Matrix
E: -10.00, S: -9.92

Show only this user's posts in this thread
« Reply #126 on: April 28, 2015, 10:48:43 PM »

A bunch of the counties in Kentucky that start with 'M' are jumbled. For instance, I clicked on Mason County, and it gave me the results for Marion County.
Logged
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #127 on: April 28, 2015, 10:54:49 PM »

A bunch of the counties in Kentucky that start with 'M' are jumbled. For instance, I clicked on Mason County, and it gave me the results for Marion County.

Thanks for the heads-up. Like I said yesterday regarding the "S/Saint" screw-ups, these things can happen and I did a good job at weeding them out initially. What has caused this problem again is a second set of conflicting orderings when I added them to this new map. When comparing how the counties are ordered by racial stats versus election results, I know I fixed those prior to this. Now, however, I'm taking that organized data from before (when I was hand-drawing things) and correlating it with yet another ordering of counties (the geographic data points).

Specially, what does it for some of the "M" counties is "Mc". In some alphabetical listings of counties, the "Mc"s come first. In others, they come after any counties that begin with "Ma". Give me a few minutes and let me see what I can do.
Logged
Bandit3 the Worker
Populist3
Junior Chimp
*****
Posts: 7,951


Political Matrix
E: -10.00, S: -9.92

Show only this user's posts in this thread
« Reply #128 on: April 28, 2015, 10:57:21 PM »

Also, someone mentioned Kosciusko IN. I think I read once that this county has some very conservative religious sect, and that's why it's so Republican.
Logged
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #129 on: April 28, 2015, 11:55:27 PM »

A bunch of the counties in Kentucky that start with 'M' are jumbled. For instance, I clicked on Mason County, and it gave me the results for Marion County.

Thanks for the heads-up. Like I said yesterday regarding the "S/Saint" screw-ups, these things can happen and I did a good job at weeding them out initially. What has caused this problem again is a second set of conflicting orderings when I added them to this new map. When comparing how the counties are ordered by racial stats versus election results, I know I fixed those prior to this. Now, however, I'm taking that organized data from before (when I was hand-drawing things) and correlating it with yet another ordering of counties (the geographic data points).

Specially, what does it for some of the "M" counties is "Mc". In some alphabetical listings of counties, the "Mc"s come first. In others, they come after any counties that begin with "Ma". Give me a few minutes and let me see what I can do.

This was much more difficult to fix in-system than I had anticipated. Now for Kentucky I have four separate sets of data for each county from trying to merge correct data, thinking it would override the old data. Roll Eyes Instead, it just duplicates it and I had to dig in there, find the messed-up entries and make the edits.

At any rate, it should all be fixed now. Let me know if the names for the following are in the right place (they seem to be, based on my glance):

McCracken
McLean
McCreary
Marshall
Martin
Mason
Madison
Magoffin
Marion
Logged
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #130 on: April 29, 2015, 01:12:17 AM »

Texas has been added.

TX 2012 statewide share of the white vote for Obama: 19.6%*

This one will have to be taken with a huge grain of salt. No matter how I adjusted my formulas, there were wide discrepancies between rural/urban counties, as well as heavily white/heavily minority counties. The formula I settled on resulted in a lot of negative numbers in rural counties that had to be manually adjusted, with varying levels of increased Latino drop-off from the statewide formula and varying levels of Latino support for Obama.

It's very hard to project the Latino share of the vote and Latino support for Obama in Texas, due to the large numbers of both established and non-established Latino populations; in most other states, it's the latter and not the former. Some of these Latino groups will be 70% Obama, while others might be 40-50% Obama. I projected statewide Latino turnout at 24%; I found one source that said 27% but this literally would have meant whites went from 26% Obama in 2008 to 15% in 2012; I'm not buying that. The swing as-is looks almost unbelievable.

The white support in some of the South TX counties is likely off in this case (when we're dealing with blacks and whites, it would be more acceptable due to reasons I mentioned on the previous page). Many of these have been manually adjusted, too, but with several of these counties having 80%+ Latino populations and <15% white populations, there is a lot of room for error.

I welcome poking and prodding on this particular state because I am sure its individual projections can be improved.
Logged
Bandit3 the Worker
Populist3
Junior Chimp
*****
Posts: 7,951


Political Matrix
E: -10.00, S: -9.92

Show only this user's posts in this thread
« Reply #131 on: April 29, 2015, 08:48:17 AM »

Those counties in Kentucky are in the right locations now.
Logged
nclib
Atlas Icon
*****
Posts: 10,300
United States


Show only this user's posts in this thread
« Reply #132 on: April 29, 2015, 09:47:07 AM »

Are you counting Hispanic whites as white?
Logged
YaBoyNY
NYMillennial
Jr. Member
***
Posts: 1,469
United States


Show only this user's posts in this thread
« Reply #133 on: April 29, 2015, 09:58:19 AM »

Yeah. I can't say that I have an easy solution for this, but white Obama % probably a lot lower in South Texas, and slightly higher in East Texas. Especially Harris and the areas around it. They're undoubtably very anti-Obama, but it's probably closer to the wherabouts of 25 - 30% rather than low teens.
Logged
Bandit3 the Worker
Populist3
Junior Chimp
*****
Posts: 7,951


Political Matrix
E: -10.00, S: -9.92

Show only this user's posts in this thread
« Reply #134 on: April 29, 2015, 10:00:49 AM »

I notice for Potter County (which includes most of Amarillo), the white vote for Obama is ZERO.
Logged
nclib
Atlas Icon
*****
Posts: 10,300
United States


Show only this user's posts in this thread
« Reply #135 on: April 29, 2015, 10:23:17 AM »

Are you counting Hispanic whites as white?

If not, the TX border counties are extremely high.
Logged
Mr. Illini
liberty142
Junior Chimp
*****
Posts: 6,838
United States


Political Matrix
E: -4.26, S: -3.30

Show only this user's posts in this thread
« Reply #136 on: April 29, 2015, 11:24:50 AM »
« Edited: April 29, 2015, 11:27:02 AM by Mr. Illini »

Why are Prince George's whites considerably more Democratic than Montgomery's whites? One would think they would be about the same.
Logged
Reginald
Jr. Member
***
Posts: 802
Show only this user's posts in this thread
« Reply #137 on: April 29, 2015, 11:41:09 AM »
« Edited: April 29, 2015, 11:54:59 AM by Reginald »

Could you disclose part of your methodology and some of the assumptions you're making on Latinos in TX? I have to agree that the south and west of the state are pretty far off. A lot of these counties are showing Dem percentages for whites that are higher than the numbers for the county as a whole, which is impossible obviously.

Also, are you using the Spanish Surname Voter Registration statistic? It's a nice proxy for capturing Latinos who are actually eligible to vote.
Logged
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #138 on: April 29, 2015, 08:55:28 PM »

If Texas has some sort of voter registration/turnout stats by county and by race, this would make things a whole lot easier. As I've said, for many of these, I have to take old exit polls, census data and when available, voter registration and turnout data that can be easily compiled by county. If there is no voter registration by race and by county data available or it's scattered to the four winds for each of them, then I'm not going to spend 10 hours or more compiling a single state when my formulas almost always (except in cases like Texas, of course) hit the result spot-on when comparing it to top-line data available in exit polling.

Are you counting Hispanic whites as white?

No, but in a state like Texas, it's possible that a larger share of Latinos identify as white anyway, even with the Census' prowess. I imagine New Mexico is going to be a pain in the ass.

Could you disclose part of your methodology and some of the assumptions you're making on Latinos in TX? I have to agree that the south and west of the state are pretty far off. A lot of these counties are showing Dem percentages for whites that are higher than the numbers for the county as a whole, which is impossible obviously.

Also, are you using the Spanish Surname Voter Registration statistic? It's a nice proxy for capturing Latinos who are actually eligible to vote.

With regards to Latino calculations, the statewide/base calculation was that their share of the electorate was equivalent to 63% of their population ("38% of the population, 24% of the electorate"). It was also assumed in the base formula that 66% of Texas Latinos voted for the President (some estimates out there showed 70%, but as I mentioned before, this would drive dozens of counties into extreme negative white vote percentages).

Using the two numbers above (Latinos as a share of the electorate being equivalent to 63% of their share of the population; 66% of them voting for Obama), there were still dozens of counties with negative white vote percentages for Obama. Some were relatively small (-1%, -3%, etc) but there were still many that were way off (-9%, -16%, etc). At this point, it really became a guessing game with gradual and consistent adjustments to the formula for those particular counties.

For most of the non-Rio Grande Valley counties, I started by lowering Latino share of the electorate even more, then would lower Latino preference for the President (usually from 66% to 60%), then lower "Other" share of the electorate, then adjust their preferences, then lower black share, and so forth - until I got something that looked reasonable). In some cases, I'd wind up adjusting Latino numbers down more than once.

It's important to note that in some of these counties, I had to reduce by more than half the baseline adjusted Latino share of the electorate, drop their support to ridiculously low levels, cut non-white, non-Latino turnout and support drastically, and still would have counties with negative numbers.

Here's an example of how much tweaking I had to do in some places to still wind up with negative white numbers:

Initial Pop -> State Formula Adjusted Turnout -> County-Specific Formula Adjusted Turnout

Latino Pop/Turnout : 45% -> 28% -> 18%
Latino Support : ## -> 66% -> 40%
Black Pop/Turnout : 8% -> 8% -> 5%
Other Turnout : 4% -> 3% -> 1%
Other Support: ## -> 60% -> 30%

I'm also considering how each region may have quite substantial fluctuations in the percentage of Latinos who can vote and who do vote. For instance, it wouldn't surprise me if there is higher turnout in South Texas and higher support for Democrats. It would seem to me that CVAP populations would be higher here, as most who come here as undocumented would likely head for urban areas rather than hang out around the border these days. Taking into account something like this and adjusting it would dramatically lower white numbers in South Texas. The problem with a lot of these counties is that the base starting point (Census) shows that whites are <15% in a lot of these counties; at that point, even a small error in the formula can dramatically shift overall white support in one direction or another. In most of these areas by default, however, white share of the electorate is two times or more as large as their share of the population, for obvious reasons.
Logged
muon2
Moderator
Atlas Icon
*****
Posts: 16,787


Show only this user's posts in this thread
« Reply #139 on: April 29, 2015, 11:01:50 PM »
« Edited: April 29, 2015, 11:08:17 PM by muon2 »

From what I've seen on DRA for 2008 there are significant differences in voting patterns for urban Latinos compared to their rural counterparts. As with whites, I might expect the Latino D vote % to be higher in urban areas. There's also the problem of turnout, especially in rural areas.

Consider rural Bailey county, NW of Lubbock.

VTD 1: HVAP 47.7%, SSVR 30.2%, McCain 76.4%.
VTD 2: HVAP 28.6%, SSVR 19.4%, McCain 83.1%.
VTD 3: HVAP 78.6%, SSVR 67.8%, McCain 39.6%.
VTD 4: HVAP 62.2%, SSVR 48.2%, McCain 61.1%.

A fit from HVAP to Obama would predict the white vote is negative 12% Obama. At least the SSVR comes out with a white vote of negative 2%. It still isn't real, but it's closer.
Logged
nclib
Atlas Icon
*****
Posts: 10,300
United States


Show only this user's posts in this thread
« Reply #140 on: April 29, 2015, 11:45:28 PM »

Also in Texas, the Hispanics are more Democratic along the border and in areas with high % of Hispanics.
Logged
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #141 on: April 30, 2015, 12:27:06 AM »
« Edited: April 30, 2015, 12:36:09 AM by RG Griff »

Thanks for the input, guys. Correcting Texas will have to be a gradual thing for me and even then, I'm not sure how accurate it will be ultimately. I imagine it's going to be similarly difficult throughout a huge chunk of the Southwest. Sad Which is why...I've decided to skip that part entirely and head to the Upper Midwest! Cheesy



Michigan, Wisconsin, Minnesota and Iowa have been added.

IA 2012 statewide share of the white vote for Obama: 50.1%
MN 2012 statewide share of the white vote for Obama: 48.4%
WI 2012 statewide share of the white vote for Obama: 48.2%
MI 2012 statewide share of the white vote for Obama: 44.3%
Logged
Ebsy
Junior Chimp
*****
Posts: 8,001
United States


Show only this user's posts in this thread
« Reply #142 on: April 30, 2015, 12:28:16 AM »

The Rust Belt sends its regards.
Logged
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #143 on: April 30, 2015, 09:42:59 PM »

North Dakota, Nebraska and Oklahoma were added this morning, but I didn't feel like posting a new batch so soon to the thread. However, I'm not at home and don't have the precise statewide percentages on hand; will post those later.
Logged
ElectionsGuy
Atlas Star
*****
Posts: 21,107
United States


Political Matrix
E: 7.10, S: -7.65

P P
Show only this user's posts in this thread
« Reply #144 on: April 30, 2015, 10:04:14 PM »

Hey Adam, Rhode Island has higher white vote percentages for Obama in the 3 counties other than Providence than the counties themselves, is that a mistake or how is that possible?
Logged
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #145 on: April 30, 2015, 11:16:11 PM »

Hey Adam, Rhode Island has higher white vote percentages for Obama in the 3 counties other than Providence than the counties themselves, is that a mistake or how is that possible?

Hey, thanks for pointing that out. It appears that I somehow didn't copy Bristol into one of the two columns (you probably noticed that there was a empty space over Bristol County where you could actually see the map) and it transposed the white percentages in the wrong counties. That's why the results didn't match. I've added Bristol's geography in and fixed all of the results.
Logged
Ebsy
Junior Chimp
*****
Posts: 8,001
United States


Show only this user's posts in this thread
« Reply #146 on: April 30, 2015, 11:21:27 PM »

Were there any counties where Obama won the white vote but didn't get over 50%?
Logged
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #147 on: April 30, 2015, 11:28:44 PM »

Were there any counties where Obama won the white vote but didn't get over 50%?

Quite a few, actually. If you now go to the "White Vote Winner by County" map, all of the pinkish-red counties are places where Obama got at least 48% of the vote but not a majority. As explained there, because my formula is more likely to underestimate white Obama support than overestimate it, I'm assuming that he won all counties where he got 48% or more of the vote.
Logged
Ebsy
Junior Chimp
*****
Posts: 8,001
United States


Show only this user's posts in this thread
« Reply #148 on: April 30, 2015, 11:31:02 PM »

Silly me not scrolling down.
Logged
Adam Griffin
Atlas Star
*****
Posts: 20,090
Greece


Political Matrix
E: -7.35, S: -6.26

Show only this user's posts in this thread
« Reply #149 on: April 30, 2015, 11:41:48 PM »


Oh, I didn't realize you hadn't seen it. Prior to your question, I had all of the counties at 48% or greater uniformly shaded as red. When you asked, I decided to go ahead and make those 48-49% counties salmon for easier discernment. I'll probably change them back to red later, though, for what it's worth. Tongue
Logged
Pages: 1 2 3 4 5 [6] 7 8  
« 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 0.045 seconds with 11 queries.