Talk:Statewide opinion polling for the 2012 Republican Party presidential primaries/Archive 1

Archive 1

Old Polling Data

Starting this Saturday, I plan to begin updating the map to dim the colors of states where the most recent polling data is older than two months every Saturday. I also added a link to the existing page containing straw poll data to this page. Utesfan100 (talk) 05:43, 9 November 2011 (UTC)

• While you are doing that, you could delete a lot of the old information. Some of these polls are 18 months old and more. The article is way too big and the information is largely irrelevant now. OldSquiffyBat (talk) 10:59, 17 December 2011 (UTC)

Metro area polls

I don't think that polls of individual metro areas or cities belong on this page, which is explicitly for statewide polls. However, I see no reason why they shouldn't be included on another similar page, perhaps Local opinion polling for the Republican Party (United States) presidential primaries, 2012. Any thoughts? Thanks. – Hysteria18 ^{(Talk • Contributions)} 21:52, 18 February 2010 (UTC)

Actually, a better solution might be to put the metro area polls back on this page (in a different section from the statewide polls) and rename it to Statewide and local opinion polling for the Republican Party (United States) presidential primaries, 2012. Given the (current) lack of much of either, perhaps this would be best? – Hysteria18 ^{(Talk • Contributions)} 22:33, 18 February 2010 (UTC)

At this point, I think we have to put up all information we've got, and simply offer disclaimers. We have so little information right now, that even a half-way reliable poll is INCREDIBLY valuable. The map needs to go back, perhaps with an asterisk on those stats which have only local polling done, but at this stage, with so little available, we have to be thankful for what we have - and try to make predictions with what we are given. I included a disclaimer on each local poll, which in my mind is perfectly acceptable. I we want to rename the page, I would be alright with that, but not with selectively using polls. Mburn16 (talk) 17:43, 21 February 2010 (UTC)

In general elections people in metro areas vote different than people outside of them, and I think that applies to primary polls as well, so these metro area polls would not be representative for the state. So if they should return to the page, it should be under its own header, and with a huge disclaimer. Ambi Valent (talk) 13:44, 28 February 2010 (UTC)

I would agree with that - so how about renaming the page "Statewide 'and local' opinion polling for the Republican Party (United States) Presidential Primaries, 2012" (note, I'm also open to a shorter name for the entire page) - including the local polls under their own header, and filling them in on the map as diagonally shaded in the correct color? If we ever get around to working out a delegate count for these contests, we won't include them... 76.230.57.176 (talk) 18:04, 28 February 2010 (UTC)

Proposed Map Coloring Rules

With the recent [unusual] Minnestoa poll, which probably requires the state to be split between Romney and Pawlenty, we've pretty much run out of colors to use, and the map is really starting to look bad with all the divisions. I think its time we establish some rules for coloring the map in the future. Here is my suggestion:

1. The candidate leading in the most states will have his/her states colored RED
2. The candidate leading in the second most states will have his/her states colored BLUE
3. The Candidate leading in the third most states will have his/her states colored GOLD
4. The candidate leading in the fourth most states will have his/her states colored GREEN
5. Candidates who only lead in their home states will be grouped together as "local candidates", and colored DARK GRAY
6. Candidates, beyond the top four, who lead in only one state, but not their home state, will be grouped together as "other" and colored BROWN

If we could get someone to write up the code for a table to put in the infobox, listing all the candidates and how many states they lead in, with the appropriate color, that would be great too. Mburn16 (talk) 01:48, 3 September 2010 (UTC)

I replaced the numbers with a hash, so the proposal appears as a numbered list as was intended. (Are such edits legal or illegal?) Ambi Valent (talk) 20:37, 6 September 2010 (UTC)

As for the proposal, I think among the leading candidates, each candidate should receive a permanent color, because the order may change at times. The actual order would then be shown in the table below the map. Ambi Valent (talk) 20:37, 6 September 2010 (UTC)

You're in the clear, I don't work on the talk pages enough to get it right all the time. As for the permenent color, that is the way we started out, but now that we're up to at least four major candidates, the colors kind of become a mess - particularly since the green doesn't look all that great with the others, yet the amount of green on the map had been growing substantially. Mburn16 (talk) 20:53, 6 September 2010 (UTC)

More confusion: Of the 5 states that are split, there are 3 different outcomes:

1. In Pennsylvania and North Carolina, the polls are really split. No problems here.
2. In Louisiana, the poll that Jindal won is not the last poll - but Jindal didn't appear in a poll after the one he won. Does he still count as split winner, since he didn't lose?
3. In Minnesota and South Carolina, the winner of one poll appears in only one of the polls - but beats the winner of the other part of the split poll. So is there a winner (Romney in MN, DeMint in SC) or is there a split?

Plus, I have a little problem on the aesthetic side: the stripes appear too thin (at least here) when the map appears in the sidebar, making it unclear which split it is. Another possibility besides redrawing could be that split polls receive their own color (or pattern) independent of which candidates the split is between. Ambi Valent (talk) 21:02, 6 September 2010 (UTC)

The Louisiana coloring is truly a mistake - I was redoing the entire system this AM and apparently slipped up. I will make that correction in a bit. I'll see what I can do about the shading of the split states as well - I might be able to make it a bit better. As for the polling in South Carolina and Minnesota, it is beyond me to try and figure out how things would poll if all candidates were compared at once. In Minnesota, for example, would Romeny still lead if Palin was included? Or does much of Romney's support come from her? Mburn16 (talk) 00:00, 7 September 2010 (UTC)

I've decided, party out of laziness, partly out of the desire to avoid confusion, and party out of the recognition that there was never a consensus on the coloring, to allow Gingrich and Huckabee to retain their current colors, even though they have technically exchanged position. I will instead update the ranking image to show the change. As long as we don't give each candidate an individual color, instead grouping single-staters and local candidates together, I think we can give each of the big four their own, permanent color. Mburn16 (talk) 00:02, 18 September 2010 (UTC)

Split leads

So far, polls which happened at the same time were counted as equal, and if they produced different winners, the lead was counted as split. But so far, that only applied to polls by the same polling firm with different candidates.

Now we have Newt Gingrich winning a Sep 2-6 poll, and Sarah Palin winning a Sep 3-8 poll. To find out whether one was earlier, or both count as equal, we could:

1. Compare the average dates. If they're less than a day apart, the polls are equal, otherwise the latter one wins. Or
2. Compare the start date. Or
3. Compare the end date. Or
4. Compare all the dates. If more than half of the dates of one poll are before the other poll started, the other poll is later. Ditto the other way around.

The first method returns average dates of Sep 4 and Sep 5/6, so there's a winner. The second and third method also produce a winner. The fourth method returns that only one day of the Sep 2-6 poll was before the other poll, and that only two days of the Sep 3-8 poll are after the other poll. Both polls share 4 days. Since neither the "before" days nor the "after" days outnumber the shared days, the polls count as equal.

So, which is it? Ambi Valent (talk) 20:16, 17 September 2010 (UTC)

Its Palin...the poll that contains the most recent results rules, and in this case, polling was conducted later in the poll which shows Palin to be the leader. Regretable, in my view, since that particular poll is muddied up by third-tiers who won't be in the race come ballot time, but nonetheless, its her. Mburn16 (talk) 23:41, 17 September 2010 (UTC)

So for the next time, does the later average date or the later end date count? Both go to Palin right now, but in future polls one method could return a winner and the other a tie (if one poll is much shorter than the other, even a different winner). Ambi Valent (talk) 19:02, 18 September 2010 (UTC)

Infobox Rankings

Is there anyone able to use include code from my user page in the infobox, to better display the rankings? I've tried, but could not get it to work, and, while the current format (an image) works reasonably well, it is also more difficult to update. Mburn16 (talk) 23:56, 17 September 2010 (UTC)

Where do straw polls go?

--Hodgson-Burnett's Secret Garden (talk) 16:07, 26 January 2011 (UTC)

I think the results (in short form) would belong into the general article on Republican primaries, as long as that article mentioned the straw poll in the first place.Ambi Valent (talk) 21:22, 2 February 2011 (UTC)

Remove Thune, DeMint, and Ryan from the map.

All three of these candidates have announced that they will definitely not be running for president, so they should be removed. 99.174.92.174 (talk) 00:40, 5 April 2011 (UTC)

Map Not Updating?

There have been more polls which have come out, changing some states (Florida is an example), yet the map doesn't change. I tried to edit but my file wouldn't upload - can anyone fix the map? 75.75.161.174 (talk) —Preceding undated comment added 21:45, 25 April 2011 (UTC).

State frontrunners

I don't think that with Huckabee's removal from the race that the second-place finishers in the states in which he lead should automatically be considered the new leaders on the map. We have no way of knowing where Huckabee's support ends up, and to assume that whoever was behind him will by default move into the lead in the state is faulty reasoning, especially in states where the other candidates were closely grouped behind him. I think the map should continue to reflect Huckabee's leads until new polling comes out, at which time the map can be recolored piece by piece. —Preceding unsigned comment added by 173.230.169.70 (talk) 19:11, 16 May 2011 (UTC)

'Oppose' - leaving Huckabee as the frontrunner in each state makes the map pointless, providing the lead to the second-place candidate at least gives a reasonable projection of who could very well be ahead...we know for sure that Huckabee isn't. I'm replacing the strikethroughs as well. Mburn16 (talk) 17:22, 17 May 2011 (UTC) -On second thought, I'll wait to replace the strikethroughs - if anyone here can offer a good reason we shouldn't indicate that a candidate has dropped out of the race. Mburn16 (talk) 17:30, 17 May 2011 (UTC)

The map is already pointless if you arbitrarily give leads to people that can't be backed up with polling. I think the whole map should be scrapped, honestly. —Preceding unsigned comment added by 173.230.169.70 (talk) 22:14, 17 May 2011 (UTC)

For the simple fact that all of Huckabee's supporters will probably not all fall for Romeny or Palin or Gingrich, we need to just not even count them until more recent polling data becomes available. —Preceding unsigned comment added by 66.213.14.210 (talk) 23:20, 17 May 2011 (UTC)

I think we would be okay to leave Huckabee on the map for now - if polling organizations love to be relevent, there will be polling that shows what people in the states feel without Huckabee running. Of course, there is also the chance that he decides to jump back in - I say revert back to with Huckabee and update as changes are made with polls without him. 75.75.161.174 (talk) —Preceding undated comment added 00:48, 19 May 2011 (UTC).

Agreed we need to change this back to as if Huckabee was still a contender. —Preceding unsigned comment added by 64.233.244.102 (talk) 18:19, 19 May 2011 (UTC)

'Oppose' - the Map should show our best guess of who will win which state if the primaries were to happen today. This means that candidates who have withdrawn their candidacy should not show up. I do, however, agree that something has to be done since a former-candidate's support will not be equally distributed to other candidates. We should mark those states with an asterisk, or perhaps a lighter/darker version of the coloring, in order to indicate that those results are our best guess given the current status of the polling. This way, we include all the information we have, but we make all the caveats known. Nick2253 (talk) 22:35, 13 July 2011 (UTC)

Definition of "Shared Lead" in Map

Right now, the map indicates states that have a shared lead only in the case that the candidates share the exact same percentage of support. However, these opinion polls have a margin of error, and I think that if the candidates' support is within the margin of error, then they should be considered a "shared lead". Otherwise, we are essentially saying the margin of error on our map is zero. Nick2253 (talk) 22:38, 13 July 2011 (UTC)

Updated Poll for Texas

http://www.azimuthpolls.com/?p=11 — Preceding unsigned comment added by 174.149.78.194 (talk) 06:06, 17 July 2011 (UTC)

I believe that poll is already mentioned in the article, and in any event it took place over a month ago. Difluoroethene (talk) 15:17, 17 July 2011 (UTC)

Right now the map shows Perry as sharing the lead in Texas, but that's only if you consider both halves of a split poll, the point of which was to see who Texas would vote for with Perry either in or out of the race. Since he's definitely in the race now, the only half of the poll that matters now is the half with him in it, which he won by a large margin. Perry should get all of Texas to himself. It just makes sense. 76.106.47.236 (talk) 21:09, 26 August 2011 (UTC)

Merge Statewide opinion polling for the Republican Party presidential primaries, 2012

The article Statewide opinion polling for the Republican Party presidential primaries, 2012 was started recently. It clearly needs to be folded into this article. Safiel (talk) 16:05, 16 September 2011 (UTC)

Never mind, I tagged the other article for CSD A10, as there is nothing there that is not already here. Safiel (talk) 03:14, 17 September 2011 (UTC)

Idea for the map

So, I know I am not active in this topic, but I have been viewing it regularly. I had an idea I wanted to throw out there--especially to those that update this page often. What the striped states were used for actual ties and the "lighter" color was used for states where the leaders lead was smaller than the margin of error. I feel like this better represents that the current leader is actually ahead in the State, but it is small enough that they could be overtaken. Whereas, the striped states seem to more inaccurately represent the states are tied. For example, a margin of error of 2.5% is much smaller and could be considered a tie if they are within that, whereas a margin of error of 5% is much larger (especially with so many undecided in each poll) that it seems off. Again, this is a suggestion I hope everyone can discuss; at least to me it makes logical sense and better represents what the page is setup to show (leaders in States). Bullshark44 (talk) 18:45, 30 September 2011 (UTC)

Virginia Poll Released on 10/11/11

In a Quinnipiac Poll released on 10/11/11, Romney and Cain are now tied for first in Virginia with 21% of registered Republican voters and "Republican leaners". The link is here: http://www.quinnipiac.edu/x5822.xml?ReleaseID=1658.

On this wiki, the most recent poll listed is a Quinnipiac Poll that was released on 10/3/11, before it was known whether or not Chris Christie or Sarah Palin would jump into the race. This old poll had Rick Perry holding a statistically significant lead of 4% over Romney. Rick Perry's new numbers show that he currently holds support among only 11% of Republicans and Republican leaners. Therefore, the map should also be re-colored to reflect this development. — Preceding unsigned comment added by 71.108.134.163 (talk) 14:28, 11 October 2011 (UTC)

South Carolina poll released on 10/10/11

On 10/10/11, American Research Group (ARG) released a poll showing that Herman Cain is holding a statistically insignificant 1% lead over Mitt Romney. Link: http://americanresearchgroup.com/.

Cain is polling at 26%. Romney at 25%. Rick Perry has suffered a significant decrease in support. Nevermind. Someone just updated the polls and the map. Someone should do the same for Virginia in light of the new Quinnipiac poll released on 10/11/11. — Preceding unsigned comment added by 71.108.134.163 (talk) 14:46, 11 October 2011 (UTC)

Colours used

I am colour blind and can't differentiate between the colours you are using on the map for Mitt Romney and Herman Cain. Please consider changing them. Thanks. 82.46.139.174 (talk) 15:55, 13 October 2011 (UTC)

I would like to keep them red and green, because those colors are very different to those of us who aren't colorblind, so I've tried some variations on Romney's shade of red. Can you distinguish between any of these? ThrawnRocks (talk) 18:55, 13 October 2011 (UTC)

Yes, I can clearly distinguish the colours in option 2. Thanks very much. 82.46.139.174 (talk) 21:17, 13 October 2011 (UTC)

What we have now

  Mitt Romney

  Herman Cain

Option 1

  Mitt Romney

  Herman Cain

Option 2

  Mitt Romney

  Herman Cain

Option 3

  Mitt Romney

  Herman Cain

Option 4

  Mitt Romney

  Herman Cain

Option 5

  Mitt Romney

  Herman Cain

Option 6

  Mitt Romney

  Herman Cain

Some thoughts

I'm wondering if, to accomodate colorblind users, it might be possible to use an animated GIF? It'd be easier for those users to distinguish between animations than colors (for example, Romney's states could go from dark to light every 2 seconds, while Cain's states would oscillate faster, perhaps every 1.5 seconds)? On the other hand, some users might find that annoying. Another possibility would be a graphic that shows the candidate's name whenever you hover over the state with your mouse, though I don't know if that's currently feasible. Difluoroethene (talk) 20:18, 13 October 2011 (UTC)

I think a GIF would be annoying to some users, but having the name (or names) of the leading candidates appear when you hover over a particular state is a good idea. I've seen it done before on photos of groups of people, although I've no idea how it works! 82.46.139.174 (talk) 21:19, 13 October 2011 (UTC)

I updated the map with option 2. I hope that is more readable for you now. With regards to the name-hover thing, have you seen it on wikipedia at all? If so you could post a link to such an image here and I could try to figure out how to do that for the map. ThrawnRocks (talk) 22:06, 13 October 2011 (UTC)

Ah yes, that's much better now. Thank you very much. I will have a look and see if I can find an image with it. 82.46.139.174 (talk) 22:25, 13 October 2011 (UTC)

Illinois

New Illinois poll shows for Cain.--Metallurgist (talk) 00:29, 22 October 2011 (UTC)

Well, I had fun learning how to download and editing SVC files and have the new update. Unfortunately my skills at updating the file are lacking. Utesfan100 (talk) 03:42, 22 October 2011 (UTC)

South Carolina

South Carolina is miscolored. The latest Public Policy Polling (PPP) Poll has Romney in the lead at 30% which is outside the 2.9% margin of error. The poll is already up but the map still shows a statistical tie between Romney and Santorum. http://http://www.publicpolicypolling.com/pdf/2011/PPP_Release_SC_107925.pdf — Preceding unsigned comment added by 71.108.131.53 (talk) 15:57, 6 January 2012 (UTC)

Minnesota

Minnesota should be colored dark red to reflect Mitt Romney's lead in the state which has been outside the margin of error since the KTSP study was conducted. The KTSP poll wsa conducted on June 15-17. Mitt Romney's percentage in that survey was 29%. His nearest competitor was at 23%. The margin of error was 4.2%. He has a clear lead in this state. — Preceding unsigned comment added by 71.108.134.163 (talk) 21:09, 24 October 2011 (UTC)

It's probably because Tim P. has left the race, and they need new data to see to whom his supporters backed. Light-jet pilot (talk) 03:02, 25 October 2011 (UTC)

Just show the state as faded dark red, then. That's what faded colors are for (to indicated predictions based on old data or candidate withdrawal). Difluoroethene (talk) 23:21, 26 October 2011 (UTC)

They should change the color and put the state into Romney's column until a new poll is taken. The same rule is being applied in Oregon, Kentucky, Missouri, and Montana. Let's have some consistency. — Preceding unsigned comment added by 71.108.134.163 (talk) 22:02, 27 October 2011 (UTC)

No i dont believe so, old data is one thing but when a withdrawn candidate has 20% of the vote it creates a margin of error so large that the poll becomes statistically useless. I dont think there should be faded colours at all.XavierGreen (talk) 23:58, 29 October 2011 (UTC)

I do think the faded colors are counterproductive, and could artificially inflate a candidate’s numbers. Light-jet pilot (talk) 03:45, 30 October 2011 (UTC)

Mississippi - Cain/Gingrich??

Mississippi shows a Cain/Gingrich "shared lead" (yellow-green) in the map even though there are seemingly only two sets of polls for the state. The most recent one shows a Gingrich lead and the older one says either Haley Barbour or Mike Huckabee. It seems to me that the state should be entirely yellow in this map. Can anyone explain otherwise? Was green a Huckabee colour? I hope it is not because the 3 point Gingrich lead is within the margin of error since I am sure that can be said of other states that are solid. — Preceding unsigned comment added by 193.75.58.39 (talk) 07:58, 9 November 2011 (UTC)

You are correct in that, the reason the map shows a tie is because Cain is within the margin of error. This has been our practice at least 4 months now. If you can find "other states", you should point them out, or change them. Light-jet pilot (talk) 13:36, 9 November 2011 (UTC)

Pennsylvania - Cain/Romney/Gingrich/Santorum

How do we color a state with four candidates within the margin of the leader? The error of the Quinnipiac poll is 4.1%. Cain and Romney are at 17%, while Gingrich and Santorum are at 13%. Utesfan100 (talk) 15:15, 10 November 2011 (UTC)

Old data striping on the map

I believe the fading on the map should be removed, it does nothing but confuse readers since several of the faded colours are barely discernable from one another.XavierGreen (talk) 16:33, 11 November 2011 (UTC)

I absolutely agree with XavierGreen on this issue. Romney's support is either dark red or a light pink, depending on an equation that has no consistent application. Whereas Herman Cain's color is bright green regardless of his degree of support. Let's go back to one solid color for each candidate, please. — Preceding unsigned comment added by 71.108.134.163 (talk) 18:21, 11 November 2011 (UTC)

I also agree, and I am the one who made the change. I think the idea that each candidate should have one distinct color. This leaves effects for old data. I have toyed with a striping effect perpendicular to the mixed polling stripes. Let me know what you think. Utesfan100 (talk) 04:37, 12 November 2011 (UTC)

Wow. The striping is really bizarre! Personally, I don't think it works - it makes it look like a patchwork quilt! It's quite confusing as well. My first thought was that it was for states that had 3 candidates tied for first place. Tiller54 (talk) 19:33, 12 November 2011 (UTC)

I was thinking they may be a bit too closely spaced. It certainly is clearer than the shading effect. Utesfan100 (talk) 21:49, 12 November 2011 (UTC)

It is, certainly. I don't think it's needed, though. It's far too confusing. Tiller54 (talk) 00:15, 13 November 2011 (UTC)

Agreed that the stripes look better now then they did at first. Also agreed that I don't really think it's needed. I feel like the stripes (and the fading before them) are basically a meaningless hodgepodge category, and they should either be there without the stripes, or grayed out. Minnesota should definitely be gray in my opinion, and probably RI as well simply for the sake of consistency. Finally, I feel like the two month cutoff for what is considered old is fairly arbitrary. New leading candidates happen almost instantly, as happend with Perry and Cain before, and my be happening with Newt now. What this means is that by the timescale on which opinions swing, two months is still pretty long, and a lot can happen in that time period. ThrawnRocks (talk) 21:21, 15 November 2011 (UTC)

The weird patchwork effect, particularly in states with a light and dark color like Montana, is significantly improved by juxtaposing the light and dark lines and spacing them further apart. This also reduces the visual impact below that used to indicate a tied state.

Utesfan100 (talk) — Preceding unsigned comment added by 98.230.89.129 (talk) 20:21, 13 November 2011 (UTC)

Yes, that is much better. Tiller54 (talk) 23:06, 14 November 2011 (UTC)

Colorado and Conn. should be striped now, right? Light-jet pilot (talk) 04:15, 1 December 2011 (UTC)

Iowa

Iowa is miscolored. The latest Insider Advantage poll (released on 11-8-11) shows that Romney, at 18.7%, is now within the margin of error (+/- 4.8%) of Cain, the leader at 23.3%. The state should be striped for Romney and Cain. [[1]] — Preceding unsigned comment added by 71.108.134.163 (talk) 18:31, 11 November 2011 (UTC)

Fixed, along with South Carolina, which is now solid Cain. Utesfan100 (talk) 04:38, 12 November 2011 (UTC)

Iowa seems to be miscolored as the latest poll ARG shows Gingrich as the leader. Another thing of note is that the Iowa Revolution PAC/TeleResearch Corp poll that the current coloring is internal polling from the Ron Paul PAC. I am not sure what the policy is with posting internal polling, but it is rarely unbiased, often inaccurate, and tends to only be released for political gain. — Preceding unsigned comment added by 24.225.64.67 (talk) 18:41, 23 November 2011 (UTC)

Iowa is again miscolored and the map erroneously states that there is a three-way tie in Iowa as of the evening of 12/27/11. The latest Public Policy Poll (PPP) (released on 12-27-11) shows that Paul and Romney are in a statistical tie with Paul at 24%, Romney at 20%, and the MoE at 4.1%. Gingrich is in a distant third place at 13% and should therefore no longer be included in the tie. http://www.publicpolicypolling.com/pdf/2011/PPP_Release_IA_1227925.pdf — Preceding unsigned comment added by 71.108.131.53 (talk) 05:57, 28 December 2011 (UTC)

Irrelevent Data

New Mexico, Oregon, and Rhode Island are included in the map but have no relevent data at all. Since the polls used contian withdrawn candidates, the effective margin of error becomes so large for these polls that the data is meaningless. They should be removed from the map. Any poll whos margin of error +totals of withdrawn candidates exceeds the difference between the leader and other candidates should not be included in the map, as those polls are not a reliable indicator of anything at all.XavierGreen (talk) 22:19, 15 November 2011 (UTC)

These states, and Minnesota, have been striped to show the diminished value of the data, due to their age alone. The established procedure for coloring appears to be as follows:

• States whose latest poll show a current candidate with a lead over all candidates larger than the margin of error for that poll are colored solid for that candidate.
• States whose latest poll shows two current candidates with a lead over all other candidates are colored by striping the colors of the candidates.
• States whose latest poll have a candidate who has withdrawn from the race within the margin of error of the lead are reset to no data.

Further, the following ideas have been recently employed:

• States whose latest poll shows three or more current candidates within the margin of error of the lead are colored black, with the candidates listed in the legend.
• States whose polls are older than two months (updated on Saturdays) are modified with black and grey stripes.

Utesfan100 (talk) 23:02, 15 November 2011 (UTC)

Then states like Minnesota should either be removed or made black, since the effective margin of error is 27.2% there would be a six way tie between romney, bachman, paul, cain, gingrich, and perry. Old data is one thing, but completey meaningless data is another. What good does a poll with a 27.2% margin of error do?XavierGreen (talk) 23:19, 15 November 2011 (UTC)

Similarly Montana should be black or removed from the map, since the most recent poll has an effective margin of error of 14% giving bachmann, romney, and gingrich a tie.XavierGreen (talk) 23:22, 15 November 2011 (UTC)

Rhode Island is the worst with a effective margin of error of 41.2% that makes every candidate in the poll that is still in the race tied. It too should be removed from the map or at least coloured black.XavierGreen (talk) 23:28, 15 November 2011 (UTC)

I am not entirely sure that your assertion that the effective margin can be calculated by adding the candidates who have stepped down to the polling margin of error linearly. Suppose we have taken a poll with a margin of error of 5%. If we find that 25% of our data values are systematically not valid and throw them out, our margin of error would only raise to 5.8%. Even if we through out 40% of the data, the margin of error goes to 6.5%. In fact, for all states mentioned here the leader has the lead when this methodology is applied.

Now your argument is reduced to whether not counting those who previously supported a candidate who stepped down or elected not to run produces a sample that is not representative of the population at large. That argument is much harder to make without polling data on the new population. Once that data is available it would supersede the data above anyways.Utesfan100 (talk) 00:08, 16 November 2011 (UTC)

If you dont know what 40% of the population will chose, than in effect you have a margin of error of 40%. This effective margin of error i am talking about is not due to the sample size of the poll as the true margin of error is. You cant throw out 40% of the voting population and get any reliable statistics at all.XavierGreen (talk) 16:06, 16 November 2011 (UTC)

Colors

I suggest modifying the colors so that Huntsman has more red to distinguish from Bachmann as show below.

All proposed colors in the map are adjacent at some edge in this image

Utesfan100 (talk) 04:35, 16 November 2011 (UTC)

Huntsman? Why? He is polling around 3% and highly unlikely to win a poll in any state. Tiller54 (talk) 02:19, 17 November 2011 (UTC)

Well, a color is reserved for his name in the internals of the file, it clashes with the color reserved for Bachmann and Bachmann still has color painted on the map making Huntsman easier to move.

Though after New Hampshire I expect his name to be removed from the file.Utesfan100 (talk) 02:55, 17 November 2011 (UTC)

He's still not going to lead in a single poll, though. In New Hampshire he's in single digits and even in his own state of Utah he might be polling in the teens but Romney is polling in the 60s to 80s. Just seems a bit pointless to me. Tiller54 (talk) 12:29, 17 November 2011 (UTC)

No more pointless than having him in the file now. Since he is there, he should have a color that does not conflict with Bachmann. Utesfan100 (talk) — Preceding unsigned comment added by 150.176.192.118 (talk) 17:36, 17 November 2011 (UTC)

Yeah, fair enough. The way the frontrunner keeps changing, he may yet have his moment in the sun! Tiller54 (talk) 00:42, 18 November 2011 (UTC)

New polling

New polling shows Gingrich clearly leading in PA and leading Romney outside of the margin of error by 0.1% in AZ Tiller54 (talk) 13:59, 22 November 2011 (UTC)

RevPAC poll

I am not familiar with what standards apply to the inclusion of polls in articles such as this. The RevPAC poll was commissioned by Revolution PAC, a political action committee affiliated with Ron Paul's campaign. My first instinct is to remove the poll on this basis, but I would not want to do this if there is a precedent for including internal polling data in such articles. Would a stipulation about the poll's affiliation with the candidate suffice for WP:NPOV or should it just be removed?--The Devil's Advocate (talk) 23:43, 23 November 2011 (UTC)

I agree. Every candidate has their own polling that shows themselves ahead. I don't think it should be included. Light-jet pilot (talk) 02:11, 24 November 2011 (UTC)

Having worked on several political campaigns i can tell you that there is generally (but not always) a bias with internal polling to show the campaigns candiate doing better than he actually is. However the only real way to tell if the poll is biased would be to take a look at the procedures under which the poll was run. For example if the pollsters identifed that they were affiliated with one particular candidate or another before asking the questions responses will generally be biased towards that candidate. Generally in polling compilation websites such as real clear politics polling organizations that may be biased are notated to which party they are biased towards, for example PPP is a democratic organization. Realclear politics does not have this poll included in their aggregations, but have not updated their iowa aggregation since this poll was released. I would wait and see what they do since they are (in my opinion) the best polling aggregator out there at the moment. Its unusual for in-house polling results to be released to the media at all though since this was from a PAC and not a campaign its a bit more understandable considering the poll showed their candidate winning. If realclear politics includes the poll in their aggregation i would include it here as well with a notation. If they dont include it i would not include it here either and add a statement that PACS affiliated with specific candidates are not included in the listings.XavierGreen (talk) 03:08, 24 November 2011 (UTC)

Good thinking! Light-jet pilot (talk) 13:36, 24 November 2011 (UTC)

Cain withdrawal

Presumably the pure-green states should simply be reverted to gray, but what about the green-striped states? For instance, should Virginia become entirely Romney-maroon or should it revert to gray? LANTZY^TALK 19:05, 3 December 2011 (UTC)

Got rid of the striped states as well. It's been the policy here before and make sense given that all those Cain supporters didn't just disappear, and Cain was just as much the leaders of those states as the other candidate, so that data is inaccurate. The real question is what to do with all the remaining states, where Cain didn't win, but was still included in the polls and sometimes still had a large minority of support in the state. I think at the very least we should stripe all the current states to indicate they are from polls from before Cain withdrew, though I think the most honest representation of the data would be to grey everything out right now. I realize these are both kind of drastic actions, so I'd like to see what other people think. ThrawnRocks (talk) 20:21, 3 December 2011 (UTC)

Another possibility is to replace Cain's green stripes with gray stripes, so that Texas, for instance, would become gray and blue striped. This would illustrate which states were recently trending towards Gingrich, Perry, or Romney, while not giving the false impression that these candidates enjoyed a commanding lead. LANTZY^TALK 21:55, 3 December 2011 (UTC)

Romney was leading in several states before the Cain withdrawal. Now that Cain has withdrawn, those states should be colored solid red unless another candidate is within the statistical margin of error with Romney. Instead, in states like Nevada, where Romney was leading, the state is just grey, as if no polling has been done there. This is a change in the rules for coloring and there is no valid explanation for it. — Preceding unsigned comment added by 71.108.129.64 (talk) 03:02, 4 December 2011 (UTC)

The states which have been reverted to gray are those in which Cain was statistically tied with some other candidate (Romney, Gingrich, or Perry). So Romney was not in an absolute lead in any of them, and there's no reason to surmise that Cain's exit has put Romney over the top. My suggestion is simply to replace all of Cain's green turf with gray, including the formerly green striping, whilst retaining the striping of currently active candidates. This way, people will know that Romney was tied for first in Illinois, that Gingrich was tied for first in Mississippi, etc, without being misled into thinking that any candidate ever had an absolute lead. LANTZY^TALK 03:12, 4 December 2011 (UTC)

There is also the problem that Cain actually hasnt withdrawn, merely suspended his campaign. This means that in some states he will still be listed on the ballot even though he is no longer actively campaigning. Its possible that some polling agencies will continue to include Cain among their list of prompted candidates when asking voters which candidate they support. It would especially make sense to do so in New Hampshire and the other early primary states where it is already listed on the ballot and will not be removed.XavierGreen (talk) 03:21, 5 December 2011 (UTC)

Unless he offically withdraws, i would leave his colouring in place.XavierGreen (talk) 03:33, 5 December 2011 (UTC)

Ballot access is a poor criterion by which to measure a candidate's political relevance. Thirty-thirty candidates will appear on the New Hampshire Republican primary ballot, and most of them are not actively campaigning. This map is intended to be a relevant overview, not an exhaustive survey. Its purpose is to provide the most current possible view of the state of the race. A central fact of the race is that Herman Cain is no longer in it. Leaving the map covered in huge swathes of green would be very misleading. LANTZY^TALK 04:32, 5 December 2011 (UTC)

Herman Cain's polling numbers will likely still be more relevent than John Huntsman or Rick Santorums in many states. You state that Herman Cain is no longer contending the presidential nomination, but in fact he still is he merely is not doing so actively. He still is a factor of concern since he is listed on the ballot, and in some states may gain a signifigant portion of the vote despite his suspended campaign.XavierGreen (talk) 14:59, 5 December 2011 (UTC)

I agree with XavierGreen. --Dezidor (talk) 16:20, 6 December 2011 (UTC)

I feel we should stick to the protocol described in the irrelevant data section above (as was done by ThrawnRocks).

• States whose latest poll show a current candidate with a lead over all candidates larger than the margin of error for that poll are colored solid for that candidate.
• States whose latest poll shows two current candidates with a lead over all other candidates are colored by striping the colors of the candidates.
• States whose latest poll shows three or more current candidates within the margin of error of the lead are colored black, with the candidates listed in the legend.
• States whose latest poll have a candidate who has withdrawn from the race within the margin of error of the lead are reset to no data.
• States whose polls are older than two months (updated on Saturdays) are modified with black and grey stripes.

On a related note, I need to update the recently expired states that is two Saturdays behind.Utesfan100 (talk) 15:38, 6 December 2011 (UTC)

For full disclosure, I was a county captain for the Cain campaign. My view that it was a correct decision to remove Cain from this map in no way should suggest that I do not still support Cain and his Plan B.Utesfan100 (talk) 15:49, 6 December 2011 (UTC)

As i stated above the problem is that data involving Herman Cain is not irrelevent, because he has not withdrawn merely suspended his campaign. The difference between him and other non-active candidates is that he is widely known and did have an active campaign. Thus he is likely to maintain signifigant levels of support in some states where he is still listed on the ballot.XavierGreen (talk) 21:09, 6 December 2011 (UTC)

Polls are showing, and likely will continue to show, that Cain's support base is going largely to Gingrich. Does this mean we should color old green states yellow? Certainly not. We should clear them and wait for valid data. Utesfan100 (talk) 17:53, 8 December 2011 (UTC)

Cain's support is no longer being measured by opinion polls, so any attempt to track his support would be impossible. We would merely be cluttering the map with outdated information. LANTZY^TALK 04:37, 7 December 2011 (UTC)

Thats nice, then why were three new polls just listed on the page today that contain data for Herman Cain. And if you folks want valid data i suggest the removal of polls that have actual withdrawn candidtates totaling more than the winner of the poll like Rhode Island has.XavierGreen (talk) 20:46, 8 December 2011 (UTC)

The Pennsylvania and Ohio polls include Cain only because they started collecting data a week before he dropped out. Notice that not a single poll which began collecting data after December 3 includes Cain. LANTZY^TALK 02:30, 9 December 2011 (UTC)

And I agree with you about the old data. Bachmann and Perry should be taken off the map. Perry hasn't topped a statewide poll since early September, and Bachmann hasn't topped a poll since June! Surely we're getting to the point where there is sufficient data coming in that we can discard any data older than a month without completely depopulating the map. LANTZY^TALK 02:35, 9 December 2011 (UTC)

The latest suffolk university new hampshire poll included Herman Cain in its prompted list of names during its poll. So some polling agencies are still including him in states where he is still on the ballot (as they should if they truely want to get an accurate reading of voter opinions).XavierGreen (talk) 15:49, 15 December 2011 (UTC)

Michigan

Mitt Romney is leading Newt Gingrich well outside of the margin of error in Michigan. Can someone who knows what they're doing please update the map to reflect this? Thanks Tiller54 (talk) 16:08, 12 December 2011 (UTC)

Actually the two most recent polls have conflicting results and overlapping timeframes. Why does a poll that was in the field from December 6 to December 9 completely trump a poll that was in the field on December 8? Half the data in the poll that shows Romney ahead is older than the data in the poll that shows Gingrich ahead. Surely Michigan should be striped Romney/Gingrich to indicate the conflicting results. We can't privilege one overlapping poll over another. 76.106.47.236 (talk) 03:05, 14 December 2011 (UTC)

Because that's the order they are sorted in, with the one that finished most recently at the top. Tiller54 (talk) 18:13, 14 December 2011 (UTC)

What if the two polls had been in the field on the exact same days, instead of just overlapping days? Are you telling me we would just randomly pick one to put on top and shade the map accordingly? Obviously not. We would shade the map according to both results. It's what we've done in the past. The polls are overlapping. We have no reason to privilege one over the other. The map should reflect both. You have explained without justifying. 76.106.47.236 (talk) 22:27, 14 December 2011 (UTC)

If two polls were both conducted, for example, in Iowa on the 18th of December and one of them showed Gingrich and Romney tied and the other showed Gingrich and Paul tied, we'd colour the map to show a three-way tie. Polls might overlap, and they often do but we sort them by which is the most recent and the most recent of all is used to colour the map. We "privilege" the most recent poll because that is the most accurate reflection of opinion in that state.Tiller54 (talk) 14:47, 16 December 2011 (UTC)

California

Gingrich has not picked up California, Romney is within the margin of error. Tiller54 (talk) 18:13, 14 December 2011 (UTC)

South Dakota

With the new poll showing Paul leading in a statistical three-way tie in that state, should it be colored in? Do we do triple-striping, or only two color striping? —Torchiest ^talk_edits 16:30, 16 December 2011 (UTC)

It will now be coloured black, with an annotation at the bottom of the key stating that Paul, Gingrich and Bachman are in a three-way tie. Also I note that someone changed Iowa back to a Gingrich/Paul tie when it's currently a Gingrich/Romney tie, so that needs to be reverted. Tiller54 (talk) 16:35, 16 December 2011 (UTC)

Maybe Iowa should be colored black as well, since the two most recent polls show Gingrich/Paul and Gingrich/Romney statistically tied, and they ended on the same date? Or do we always go by just the latest poll? That would be the one that was done only on Dec 13 and shows Gingrich/Romney tied. —Torchiest ^talk_edits 16:43, 16 December 2011 (UTC)

We always go by the latest one. Still, there'll be a new Iowa poll out in a couple of days anyway. Tiller54 (talk) 16:48, 16 December 2011 (UTC)

Quick as that: Paul 23 Romney 20 [2]Metallurgist (talk) 04:52, 19 December 2011 (UTC)

And even quicker, Paul 24, Romney 18 Tiller54 (talk) 19:54, 19 December 2011 (UTC)

Paul leading Iowa

Can someone redo the map to change Iowa to all purple? I made an edit to that image before and I think I bungled it, so I don't want to mess it up again. Thanks. —Torchiest ^talk_edits 18:30, 19 December 2011 (UTC)

Do you need an image editor to edit the map or is it just code?--Metallurgist (talk) 19:13, 19 December 2011 (UTC)

Last time I saved the map and edited it with an image editor, but I think it was a little messed up after that. If you look at the history of the image it seems as though the striped two-person states are created using either code or a special type of paintbrush. Not really sure. —Torchiest ^talk_edits 19:22, 19 December 2011 (UTC)

And it's already been fixed, huzzah! —Torchiest ^talk_edits 19:34, 19 December 2011 (UTC)

There's no margin of error listed, and given that only 391 people were polled it could be up to 6%, which would put Romney within the margin of error Tiller54 (talk) 19:57, 19 December 2011 (UTC)

I don't think it should be shown as Paul being in the lead. We have two polls ending December 18th, with one showing Romney and Paul tied for first. There are also polls from a week ago showing Gingrich leading or tied for first in Iowa as well. My thinking is we should either have strips for Paul and Romney or maybe even color it black for a three-way tie.--The Devil's Advocate (talk) 21:04, 19 December 2011 (UTC)

I asked about this in the last section. Practice has been to go strictly by the most recent poll only, which would be the Dec 18 poll. —Torchiest ^talk_edits 21:09, 19 December 2011 (UTC)

Yes, the 18th is the most recent one but it has no margin of error. I'd say we should err on the side of caution and leave Iowa striped for a Paul/Romney tie until the next poll comes out. Tiller54 (talk) 21:14, 19 December 2011 (UTC)

In my own opinion polls without margins of error should not even be listed, to list a poll without its margin of error is rather amatureish and smacks of unreliablity.XavierGreen (talk) 21:39, 19 December 2011 (UTC)

Here on InsiderAdvantage's website, it says the margin of error is 5.7%, which puts Romney in a statistical tie with Paul: http://insideradvantagegeorgia.com/ 72.253.114.212 (talk) 22:06, 19 December 2011 (UTC)

Ah, thanks for that. They are tied after all then! Tiller54 (talk) 22:41, 19 December 2011 (UTC)

I went to Commons and reverted to the one showing them tied.--The Devil's Advocate (talk) 23:17, 19 December 2011 (UTC)

Ron Paul in lead

I added Paul to the lead because he actually has won polls in two states, South Dakota and Iowa. He may be statistically tied, but he did win the polls. That seems sufficient to include him with the rest. —Torchiest ^talk_edits 18:45, 20 December 2011 (UTC)

Yes, that's fair enough. I've undone my edit.Tiller54 (talk) 19:04, 20 December 2011 (UTC)

Iowa

Do we want Iowa a 3-way tie because multiple polls ended the 18th? 96.241.30.163 (talk) 16:37, 21 December 2011 (UTC)

Iowa

Do we want Iowa a 3-way tie because multiple polls ended the 18th? 96.241.30.163 (talk) 18:58, 21 December 2011 (UTC)

Virginia

With Gingrich, Huntsman, Bachmann, Santorum and Perry not on the ballot in Virginia,[3] [4] that leaves Paul and Romney as the only candidates who have qualified for the ballot. With that in mind, should Virginia be coloured for Gingrich? He might be ahead in the polls there, but if he's not even on the ballot shouldn't it be coloured grey? Tiller54 (talk) 10:54, 24 December 2011 (UTC)

Yes it should, but I dont know how to code the map.--Metallurgist (talk) 18:35, 24 December 2011 (UTC)

Just because they are not listed on the ballot does not mean that people cannot vote for them, write in ballots exist in virginia and newt gingrich plans on persuing a write in campaign there. Rick perry plans on challenging the state divisions of elections ruling, so he may still be listed on the ballot depending on the outcome of the hearings his campaign files.XavierGreen (talk) 21:38, 24 December 2011 (UTC)

Virginia doesnt allow writeins. Virginia should be gray.[5]--Metallurgist (talk) 22:17, 24 December 2011 (UTC)

Metallurgist is right, write-ins are prohibited. We should probably wait to see if any appeals are lodged first, though. Tiller54 (talk) 23:18, 24 December 2011 (UTC)

//“Gingrich cannot have voters write in, because Va. Code section 24.2-529 prohibits write-in ballots in primary elections,” Carl Tobias, a professor at the University of Richmond School of Law, explained via email.//[6] By that logic, we should wait for the end of the election before updating anything. The map is supposed to be dynamic with changing trends. The poll is no longer representative on several counts, not just Gingrich being booted, there is another 20% from other candidates that arent on the ballot. Altogether, thats 50% of the survey that is inaccurate. This is no different than if a campaign was suspended. The candidate could reenter, but until then, the polls and our map should reflect the status.--Metallurgist (talk) 02:40, 25 December 2011 (UTC)

Well thats the same logic that ive been using to argue that the polling for states like rhode island are irrelevent and should be removed from the map. I was under the impression that write in votes were allowed, but it appears i was mistaken. Regardless i still think we should wait until the appeals process has been completed.XavierGreen (talk) 04:37, 25 December 2011 (UTC)

Well I think this page should be deleted until every poll has come out and the election is over.--Metallurgist (talk) 10:33, 25 December 2011 (UTC)

I agree with the idea that the map is supposed to be dynamic. Currently, the leader of the latest Virginia polls is not going to appear on the ballot. We should change the state to gray to reflect the fact that there are no up-to-date polls. If Gingrich finds a way to get on the ballot, we can change it back. When Cain suspended his campaign, weren't states that he was leading in changed to gray, even though we didn't have more recent polls? He is still going to appear on some of those states' ballots, I believe, so he has more of a shot than Gingrich currently does in Virginia. Finally, I would say leaving Virginia colored for Gingrich is a case of WP:CRYSTAL. He might get back on the ballot, but according to the most recent information available, he isn't on it. —Torchiest ^talk_edits 11:13, 25 December 2011 (UTC)

Doesn't the name of the page mean anything??? This page is for "Opinion Polling"! NOT for who wins, or is on the ballot! Light-jet pilot (talk) 14:43, 25 December 2011 (UTC)'

That is very true, in theory a candidate can win a poll and yet be unqualified for a ballot. This page should reflect only what the polls say nothing else.XavierGreen (talk) 15:14, 25 December 2011 (UTC)

The page certainly does reflect what the polls say, but if the map took no notice of who was or wasn't a candidate then half the states on the map would show Cain, Palin, Huckabee or Chris Christie in the lead which would be ridiculous. Tiller54 (talk) 16:44, 25 December 2011 (UTC)

True Tiller54, but all the candidates you mentioned have really droped out or declined to run. Light-jet pilot (talk) 17:06, 25 December 2011 (UTC)

Exactly. Any poll that included Palin or Huckabee as an option would be purely hypothetical, as would a poll in VA that included Gingrich, given that the only choice VA GOP primary voters will have is between Romney and Paul. On a related note, Gingrich himself says that he "declined" to appear on the Missouri ballot (whether that's true or his campaign just messed up), so any poll there that included him would also be hypothetical, just as if it included George W. Bush or Chris Christie. Tiller54 (talk) 18:12, 25 December 2011 (UTC)

The missouri primary is just for show and doesnt count for anything, to avoid having their delegate cound cut in half missouri's republican party will be having a caucus to elect delegates.XavierGreen (talk) 15:30, 27 December 2011 (UTC)

Alright, back to the point. What the polls say only matters if they are asking a real question. If a poll asked about Sarah Palin and she won, we would not post that poll on the map. A poll might still put Cain as an option and if he won the poll, would he be on the map? No, because he is not on the ballot and the poll is irreflective of the reality. Similarly, Gingrich at present will not be on the ballot in Virginia, so the poll showing him as leader is outdated. Its as simple as that. I dont get what the problem is. Metallurgist (talk) 06:22, 28 December 2011 (UTC)

Well the problem with that logic is that Cain is still on the ballot in many states. One must have it one way or the other not both lol. Either we show on the map polling for only candidates who are on the ballot, or everyone regardless of what their ballot status is. There needs to be firm criteria.XavierGreen (talk) 15:03, 28 December 2011 (UTC)

There's no problem with that logic - Cain's ended his campaign. Tiller54 (talk) 15:27, 28 December 2011 (UTC)

Oh, and for the editors who think Rhode Island should be "greyed out" too - a solution may present itself [7] Tiller54 (talk) 17:44, 28 December 2011 (UTC)

I'm with Tiller54 on this one. The difference isn't that they're on or off the ballot, they DROPED OUT OR DECLINED TO RUN. The way I see it, that's very clear "firm criteria" Light-jet pilot (talk) 18:19, 28 December 2011 (UTC)

Okay, I removed the Virginia poll from the map because it seems that poll isn't representative of the actual primary itself. If there's a lot of disagreement I guess just revert it, but this seems to make the most sense to me. ThrawnRocks (talk) 18:54, 28 December 2011 (UTC)

Delaware is missing

In the "1 Polling" section, Delaware is omitted. Even if there are no polls that are immediately available for Delaware, it should be added as a sub-section in "1 Polling". Otherwise, people from that state when checking this page, may be confused and think that their state will not have a primary. If someone does add it as a section, they could simply put the text "no polling data currently available". I doubt that no polls have been held in Delaware, so there are probably some available online that could be found, via Dover newspapers or other state media.

Stopde (talk) 13:22, 31 December 2011 (UTC)Stopde

That sounds like a good idea, but for the record, Delaware isn't the only state missing. The list is as follows. Arkansas, Delaware, Idaho, Indiana, Kansas, Maryland, North Dakota, and Wyoming. :) Light-jet pilot (talk) 01:50, 2 January 2012 (UTC)

I do not have an opinion if missing states should be added, but if they are added I believe the territories which also elect delegates should be included. They are: American Samoa, District of Columbia, Guam, Northern Marianas, Puerto Rico, and the US Virgin Islands. Tucson Indigo (talk) 05:46, 9 January 2012 (UTC)

Highlighting

Does the pink highlighting behind some table cells or even individual words have any meaning? If so, explain it. If not, remove it.68.110.104.80 (talk) 16:22, 1 January 2012 (UTC)

It indicates a lead or shared lead per the national polling page. It's "pink" per discussion in the "Top three" section above. Hope that clears that up for ya! Happy editing! Light-jet pilot (talk) 01:34, 2 January 2012 (UTC)

Michele drops out

Oregon has some Michele Bachmann orange coloring, and this should be removed (and changed to red for Romney, or gray for data unknown), now that she has dropped out of the race, the way Herman Cain's colors were removed when he dropped out. Also, South Dakota in a potential 3-way tie, should be corrected to a two-way tie, or perhaps colored as gray, with Bachmann now being removed there as a potential winner. Stopde (talk) 17:27, 4 January 2012 (UTC)Stopde

Done. Thanks for the heads-up about that ThrawnRocks (talk) 17:40, 4 January 2012 (UTC)

I don't know...sure she's out, and soon most of the others will be too, but the last poll shows her winning, and I think it should remain that way until a new poll comes out. Look at South Carolina, where all the polls are relatively old and the primary is fast approachingEricl (talk) 16:24, 5 January 2012 (UTC)

I only removed the states where she was tied for the lead (OR and SD). SC was fine, even before the most recent poll came out. 98.228.57.70 (talk) 17:22, 6 January 2012 (UTC)

CBS News claims no second place in NH for Suffolk Jan 2-3?

Is there a proper way to show that, according to CBS News in their report of the Jan. 2-3, 2012 Suffolk University polling in New Hampshire, there was no 2nd place (Ron Paul/14%) winner? (i.e. polling skipped from 1st to 3rd, 4th, 5th, etc.) CBS News is a WP:RS, yes? --Goldfndr (talk) 06:28, 8 January 2012 (UTC)

It's not relevant to this page. Fox News used a picture of Barack Obama on top of Mitt Romney's numbers when discussing another poll, but that's not included either. Tiller54 (talk) 13:16, 8 January 2012 (UTC)

South Carolina

The latest PPP poll is from Jan. 5-7. There is already a PPP poll for Jan. 5-7. Is it a duplicate? Light-jet pilot (talk) 13:39, 12 January 2012 (UTC)

Alphabetic vs chronological

Wouldn't it be better to list the states chronologically based on when they go to the polls? I always come here and end up looking down to Iowa and New Hampshire. — Preceding unsigned comment added by 216.81.8.90 (talk) 00:27, 20 December 2011 (UTC)

That's not a bad idea, really. I do the same thing. —Torchiest ^talk_edits 00:37, 20 December 2011 (UTC)

I agree --24.136.179.55 (talk) 05:05, 20 December 2011 (UTC)

Great idea, but it should be sorted by when the voting ends not when it begins. I think Maine should be 2/11 not 2/4. Tucson Indigo (talk) 15:48, 13 January 2012 (UTC)

South Carolina update

South Carolina is now a tie between Rick Santorum and Mitt Romney in a recent Rasmussen Report it had Romney at 27, Santorum 24 and Gingruch 18. — Preceding unsigned comment added by 50.13.66.210 (talk) 23:57, 12 January 2012 (UTC)

What happens after Iowa caucuses?

What should happen to Iowa after tonight's caucuses? It doesn't make sense to continue to show polls there once it is over. Should the map be coloured for the winner? — Preceding unsigned comment added by 216.81.5.62 (talk) 19:55, 3 January 2012 (UTC)

I've always assumed that it would be colored solid for the winner. Of course this isn't perfect, given that delegates are handed out proportionally, and not winner take all like in the presidential election, but it seems to be the simplest solution. That way by the end of the primary season the map will simply be the results for all the states. At the point showing pre-primary polls shouldn't really matter on the map, though they will still exist in list form on this page. I think we'll probably want to have some way to indicate actual wins vs opinion poll wins in the count under the map. Maybe a non-italicized number? Or perhaps put the poll numbers in square brackets. ThrawnRocks (talk) 23:58, 3 January 2012 (UTC)

Except that tonight's caucuses (along with, among others, Minnesota's February 7th caucus) won't actally apportion delegates; they're non-binding. If they were binding, the RNC would have penalized Iowa by halving its delegates. Does anybody know what actually determines delegate selection in Iowa, if not the caucuses themselves? And, should that affect map coloring (i. e., should the map be colored according to the caucus winner, or the candidate who actually gets the most delegates; they may not be the same). 173.165.239.237 (talk) 00:54, 4 January 2012 (UTC)

I believe this map should show only polling data not results, the main page for the republican primaries should be the one where the results are placed not here. As for the Iowa delegates, they are chosen at the iowa state convention, this caucus only deterines the precinct delegates to the county caucuses.XavierGreen (talk) 00:55, 4 January 2012 (UTC)

So in the end the map will simply show the results of the final polls to be released for each state just before their primaries? I'm not sure what information you think the map would be conveying with that. This is a page about the opinion polls for the primaries. The primary results are inherently relevant information for the article. These are polls with an end goal in mind. To exclude that end goal once reached seems strange to me. Why is it more important to show that Ron Paul was statistically tied with Mitt Romney before the caucus than to show that Mitt Romney actually won the caucus? That information should be kept historically in the table for Iowa for sure, but why should that be front and center? ThrawnRocks (talk) 08:42, 4 January 2012 (UTC)

I'm in agreement with that. It's strange to colour in Iowa for a Romney/Paul tie when Paul came 3rd Tiller54 (talk) 18:19, 4 January 2012 (UTC)

I think the state should be colored in as the winner's color, because the "last poll" is indeed the primary or caucus. Stopde (talk) 09:42, 4 January 2012 (UTC)stopde

Im torn between showing the final poll and the actual result. Theres arguments for both sides. Altho as I think about it more, actual results. But only have one winner even if its a statistical tie like it was in Iowa.--Metallurgist (talk) 19:18, 14 January 2012 (UTC)

South Dakota

What happened to South Dakota on the map? It should be black and striped.--Metallurgist (talk) 19:19, 14 January 2012 (UTC)

Bachmann was statistically tied for the lead and she has withdrawn. Tiller54 (talk) 01:43, 15 January 2012 (UTC)

Candidates that have dropped out

Some polls that include candidates that have dropped out, should have that state colored gray, until a new poll comes out. For example, let's say a person supported Bachmann -- but Romney is leading in the poll for that state. It is not safe to assume that the person will shift their vote to Romney. Let's say a person supported Huntsman -- but Perry is leading in the poll for that state. It is not safe to assume that the person will shift their vote to Perry. Let's say a person supported Cain -- but Gingrich is leading in the poll for that state. It is not safe to assume that the person will shift their vote to Gingrich. I think RECENT polls that include dropped out candidates should have their states colored gray, UNTIL we get a new poll that indicates the current lineup. Take South Carolina for example. It should be colored gray, because the most recent poll (InsiderAdvantage of January 15, 2012) includes data for Jon Huntsman. It should be colored the appropriate color after the newest poll excludes him. Florida's most recent poll excludes Huntsman, so I think it is fair to have it colored red (as it currently is) for Romney, since he is at the top of that poll. Does anyone else agree? I will defer to those that update the page more regularly, but I think this is a way to keep the map more fair. :)Stopde (talk) 09:24, 17 January 2012 (UTC)

Ohio

Ohio is again miscolored. A Quinnipiac poll released on January 18, 2012 has Romney leading all other Republican candidates in Ohio with 27% support. Santorum is second with 18% support. Other candidates have the following numbers: Gingrich = 17%, Paul = 10%, Perry = 4%. Here is the link: http://www.quinnipiac.edu/institutes-and-centers/polling-institute/ohio/release-detail?ReleaseID=1690 — Preceding unsigned comment added by Neodanite (talk • contribs) 05:18, 19 January 2012 (UTC)

Ohio is miscolored. The latest poll had a split to test the effect of Palin running or not. But Perry won both halves of the split. Presumably the state was striped for Romney and Perry because Perry wasn't running when the poll came out. But he is now. So he should get the whole state. 76.106.47.236 (talk) 22:08, 2 September 2011 (UTC)

A Quinnipiac poll released on September 28, 2011 has Romney leading all other Republican candidates in Ohio with 24% support. Perry is second with 20% support. Other candidates have the following numbers: Palin = 9%, Cain = 7%, Paul = 6%, Gingrich = 4%, Santorum = 4%, Bachmann = 3%, Huntsman = 1%. Here is the link: http://www.quinnipiac.edu/x1322.xml?ReleaseID=1652

Due to this new poll, Ohio should be colored solid red to indicate that Romney is the new leader in Ohio. — Preceding unsigned comment added by 71.108.134.163 (talk) 15:39, 28 September 2011 (UTC)

Also, PA should be updated too. The newest polls have Romney in the lead in PA. Bullshark44 (talk) 21:05, 28 September 2011 (UTC)

Then post the new polls under the PA section? I'll update OH now, but I have no idea what polls you're talking about. If you know of polls that took place that aren't shown under the list of poll for their respective states, just add them. ThrawnRocks (talk) 23:33, 28 September 2011 (UTC)

Now that it's actually started...

Shouldn't the states be listed in Chronological rather than Alphabetical order? The stuff everyone's looking for is on NH, SC and FL, and who really cares were the polling is for the later states as they're going to be totally obsolete very soon.Ericl (talk) 16:27, 5 January 2012 (UTC)

I think it is a great idea to have them in chronological order, the way it is currently. I think once a state has had a primary/caucus, that it should be bumped down to the bottom -- because no more polls are being held there. It seems pointless to have Iowa now at the top, when the next situation is New Hampshire. And NH should be bumped below Iowa at the bottom once the results are in later tonight. It seems useless clutter to have to scroll down to see the next chronological poll when states that have already had their "say" are bunched up above it. Stopde (talk) 09:24, 10 January 2012 (UTC)

If they're arranged chronologically, putting the old ones at the bottom would be much easier to navigate. When they're all finished, they'll be back in chronological order anyway.Tiller54 (talk) 18:57, 10 January 2012 (UTC)

Perhaps dividing them into "Upcoming primaries" and "Past primaries" would clear up any possible confusion?Tiller54 (talk) 18:58, 10 January 2012 (UTC)

Totally agree (especially about the "When they're all finished, they'll be back in chronological order anyway"). :) Stopde (talk) 21:06, 10 January 2012 (UTC)

This is insane and confusing. Its easier to find states in alphabetical order. Also, thats what they were in 2008.--Metallurgist (talk) 19:14, 14 January 2012 (UTC)

Question to the more advanced editors: is it possible to consolidate a few of the tabs into one paragraph, for easier reading? For example, tabs 17, 21, 48, and 49 could all be merged into one section on South Carolina. 35 and 50 could be put together in a South Dakota section. 26, 36, 39 and 40 could all be put together in an Iowa section. Etc. Stopde (talk) 05:16, 22 January 2012 (UTC)

Perry drops out & Santorum gets Iowa

Rick Perry has dropped out -- Missouri and Kentucky colored blue for him need to be changed to gray. Also, most news articles I am reading online show that Santorum has won Iowa -- not Romney. The http://en.wikipedia.org/wiki/Republican_Party_presidential_primaries,_2012 page has already been changed to indicate this.Stopde (talk) 17:21, 19 January 2012 (UTC)

This is true. http://www.politico.com/news/stories/0112/71657.html and http://www.politico.com/blogs/burns-haberman/2012/01/iowa-result-santorum-up-officially-a-split-decision-111399.html Light-jet pilot (talk) 17:30, 19 January 2012 (UTC)

Top three

I see we have columns now for the top 3 candidates. Quick question, why is the colour for the first place candidate(s) bright green? Shouldn't it be Republican red, like on every other poll? Tiller54 (talk) 00:21, 9 December 2011 (UTC)

I Agree. Light-jet pilot (talk) 03:24, 9 December 2011 (UTC)

If you mean changing it to the same color as in the nationwide polling article, then I agree. That makes sense. LANTZY^TALK 04:13, 9 December 2011 (UTC)

Ah, much better! Tiller54 (talk) 16:08, 12 December 2011 (UTC)

Different issue: why do we only highlight the person in "first place"? Given our policy on striping states within the margin of error, shouldn't we be highlighting anyone who's within the margin of error on the individual poll results too? 76.106.47.236 (talk) 03:10, 14 December 2011 (UTC)

Generally, anyone within the margin of error is bolded, but sometimes people forget. —Torchiest ^talk_edits 05:53, 14 December 2011 (UTC)

I'm not talking about bolding. I'm talking about the red shading. If two people are within the margin of error shouldn't they both be shaded (as well as bolded)? There is not a single poll result on the current page where more than one candidate is shaded, NOT EVEN when two (or more) candidates were actually numerically tied (e.g. the October poll in Virginia conducted by Quinnipiac where Cain and Romney tied at 21% but only Cain is shaded red). If the shading isn't to indicate who is statistically in the lead according to the poll (which it evidently is not doing at the moment), then what is it for? 76.106.47.236 (talk) 17:20, 14 December 2011 (UTC)

Oh, I understand now. I added shading to almost half of the page. If anyone else feels like finishing it, go for it. —Torchiest ^talk_edits 22:49, 14 December 2011 (UTC)

Thanks. Since we seem to be in agreement about this now, maybe if I have some time later I'll contribute to it also. 184.80.136.186 (talk) 01:12, 15 December 2011 (UTC)

Why did we switch to a "top three" system? It is incredibly confusing. Previously, I could just look down a list of polls and quickly observe trends. Now, I have to read every name in the first-place column for each poll. Furthermore, the highlighting is really redundant. If every name in the "first place" column is highlighted, why bother? In order to show statistical ties, it would be more appropriate to place both those names in the "first place" column. --Nick2253 (talk) 00:48, 26 January 2012 (UTC)

Updated Polls for Arizona and Michigan

http://http://www.rasmussenreports.com/public_content/politics/elections/election_2012/election_2012_presidential_election/michigan/2012_michigan_republican_primary New Michigan poll by Rasmussen Reports released on 2/2/12. (Romney at 38%, Gingrich at 23%, everyone else is below 18%) The state should be colored red to reflect Romney's re-assertion of dominance in the state. It is currently colored to reflect the statistical tie between Romney and Gingrich that existed when Epic-MRA polled the state for the Detroit news.

http://www.rasmussenreports.com/public_content/politics/elections/election_2012/election_2012_presidential_election/arizona/2012_arizona_republican_primary New Arizona poll released by Rasmussen Reports on 2/2/12 shows Romney way ahead in Arizona. (Romney at 48%, Gingrich at 24%)

Both Michigan and Arizona should be re-colored to show Romney's dominance in these states. They are both currently mis-colored. — Preceding unsigned comment added by 71.108.142.18 (talk) 05:51, 3 February 2012 (UTC)

As of 9 p.m. Pacific Standard Time, the State of Arizona and the State of Michigan still remain mis-colored. This should be fixed immediately. — Preceding unsigned comment added by 71.108.142.18 (talk) 05:00, 4 February 2012 (UTC)

As of February 4, 2012 at 5:30 a.m., Pacific Standard Time, both Arizona and Michigan are still mis-colored. Please correct this. Thanks. — Preceding unsigned comment added by 71.108.142.18 (talk) 13:28, 4 February 2012 (UTC)

Updated Poll for Nevada

http://http://www.lvrj.com/news/romney-poised-to-roll-poll-shows-138551749.html The map shows a split lead between Romney and Gingrich. Romney is up by 20% in the Las Vegas Review Journal poll which was conducted on Feb. 1, 2012 and released on Feb. 2, 2012. Please update the map. — Preceding unsigned comment added by 71.108.142.18 (talk) 14:34, 2 February 2012 (UTC)

Old Polls?

Rhode Island should be colored red with diagonal lines across it. Romney had a clear lead in the most recent poll conducted there. The state should be colored the same way Vermont and Utah are colored. — Preceding unsigned comment added by 71.108.142.18 (talk) 06:00, 3 February 2012 (UTC)

Looking at states like Alabama, which have not had a poll conducted in over a year, should the results of these polls be included on the map? Seems to me like it is far too outdated to be considered a useful information at all. — Preceding unsigned comment added by 71.36.70.196 (talk) 19:27, 1 August 2011 (UTC)

Yes 94.172.109.210 (talk) 16:05, 20 August 2011 (UTC)

The latest poll from Alabama is a year-and-a-half old, and only included Huckabee, Palin, and Romney. Huckabee "won," but isn't running, so we're giving Sarah Palin credit for the state? That's not a "best guess" by any stretch of the imagination. It's a formulaic interpretation of irrelevant data. A better "guess" would be to examine demographically similar states (like maybe Louisiana and Georgia, both of which have very recent polling) and extrapolate from there (this, of course, would introduce considerable bias). Alabama should be recolored to indicate that there is no useful polling. It's the only responsible choice that isn't biased. Similar considerations should apply to many of the other states on the map. 76.106.47.236 (talk) 21:07, 26 August 2011 (UTC)

What is the measure for when a poll becomes irrelevant? What criteria do you use to determine that it should no longer be considered at all? Alabama may be a clear case, but how do we define the cutoff point? I agree with what you're saying but having a cutoff point has the potential to introduce just as much bias as having none, and I'm not sure how to go forward. ThrawnRocks (talk) 21:54, 30 August 2011 (UTC)

Alright here's a simple metric for you. Any poll that does not include a field substantially similar to the current field possibly with the addition of candidates who have not yet ruled out running is irrelevant. A poll that includes Huckabee and not Perry or Bachmann or Paul is irrelevant. And if we insist on incorporating old bizarre data we should at least have the decency to be honest about it. Palin did not win the poll in Alabama. Either show Alabama as being for Huckabee or as not having any relevant polling data. If I knew how to edit the image file myself I would be bold and just make these changes but since I don't I have to just hope hat someone else will do the sensible thing. 76.106.47.236 (talk) 21:55, 2 September 2011 (UTC)

I agree with 76.106.47.236. If the most recent poll within a state has a candidate in the lead who has withdrawn or declined to run, the state should be "grayed out" to indicate "no relevant data". Suppose Sarah Palin also declines to run before we get another poll from Alabama -- that shouldn't mean that Romney is shown as the leader in the state by default because he came in third place in the state's last poll, a year and a half ago. --Metropolitan90 (talk) 03:49, 9 September 2011 (UTC)

I would say that at a bare minimum polls that have withdrawn candidates leading in a state be considered irrelevent, since they have no bearing on the current situation.XavierGreen (talk) 14:58, 9 September 2011 (UTC)

I agree with 76.106.47.236, Metropolitan89, and XavierGreen. Polls that don't accurately represent the current field of declared candidates should not be used for the purpose of marking "leaders" on the map. —Torchiest ^talk_edits 18:27, 22 September 2011 (UTC)

What about Montana? Palin has dropped out, and the poll used included data without Palin. Should this not be Solid-bach? Utesfan100 (talk) 03:45, 22 October 2011 (UTC)

The poll done without Palin is being used. Bachmann and Romney are statistically tied in that poll, which is reflected on the map. ThrawnRocks (talk) 18:16, 22 October 2011 (UTC)

Yes. Since I posted that, I discovered that I have downloaded an old version of the SVC file where this change had not yet been made. P.S. Now that I have learned how to update the SVC file, where can I find a good outline for how to replace the map with an updated version? Utesfan100 (talk) 02:27, 23 October 2011 (UTC)

black and gray stripes

Shouldnt the cutoff date for that be changed by now? Sep 19 is eons ago in this race.--Metallurgist (talk) 09:46, 31 January 2012 (UTC)

Totally agree. I suggested above in the "Candidates that have dropped out" section that the regular/advanced editors update the map, because it has states with colors based on polls that still include Bachmann, Perry, Huntsman, Cain, when it cannot be assumed that the people that supported those people will support the person that was leading that very old poll. The map should be nearly all white, until recent polls only have Santorum, Paul, Gingrich, Romney, and any other non-withdrawn candidates in that poll (i.e. Karger, Roemer, etc). Stopde (talk) 13:14, 31 January 2012 (UTC)

Ohio

Ohio should be a three way tie as Rick Romney and Gingrich are all within the same margin of error for first. — Preceding unsigned comment added by 50.13.66.210 (talk) 02:42, 3 February 2012 (UTC)

Based on the recent poll, Rick is 4 percentage points from the lead. The margin of error is 3.9 points, so it is not quite a 3 way tie, but it is near it! Stopde (talk) 12:27, 3 February 2012 (UTC)

Updated Poll for Michigan

Michigan is again mis-colored. It is colored solid blue as of 9:45 p.m. PST. It should be blue and red to reflect the statistical tie between Santorum and Romney. The Rasmussen poll released today indicates that Romney and Santorum are in a statistical tie with Santorum at 35% and Romney at 32%. http://www.rasmussenreports.com/public_content/politics/elections/election_2012/election_2012_presidential_election/michigan/2012_michigan_republican_primary — Preceding unsigned comment added by 71.108.142.18 (talk) 05:46, 15 February 2012 (UTC)

Missouri

Missouri should be blue and yellow--Metallurgist (talk) 08:38, 1 February 2012 (UTC)

I think you are correct. I am new(ish) however, so I don't know how to change it. Light-jet pilot (talk) 20:35, 1 February 2012 (UTC)

Yea Ivent a clue either how to change them.--Metallurgist (talk) 23:15, 1 February 2012 (UTC)

Wait a minute. Im completely wrong. Gingrich isnt on the ballot.--Metallurgist (talk) 05:26, 4 February 2012 (UTC)

MISSOURI IN WRONG PLACE IN CONTENTS--92.12.208.75 (talk) 11:26, 6 February 2012 (UTC)

No, the actual caucus is in March.--Metallurgist (talk) 00:38, 7 February 2012 (UTC)

New York

Needs Brown and Yellow per new poll.--Metallurgist (talk) 00:33, 7 February 2012 (UTC)

New poll(s) for New York

Quinnipiac, February 15, 2012 Location (talk) 16:31, 15 February 2012 (UTC)

Wins on map

I wonder if the map could show a distinction between polls and a win. Maybe a star to indicated that the colour is the winner and not a pol result. I would not be able to make this change myself. — Preceding unsigned comment added by 216.81.21.194 (talk) 21:31, 12 February 2012 (UTC)

I like this idea!!Light-jet pilot (talk) 03:12, 13 February 2012 (UTC)

I can see this being useful, but at the same time, the map reflects who is winning the state. Once its been one, its settled. I think after the primaries are over it might be worth having the final poll winner displayed on each part of the map.--Metallurgist (talk) 17:36, 13 February 2012 (UTC)

Using Newer Polls

Shouldn't the date for no longer using polls be made more recent for the graphic. Polls before November and even December are really useless now, so i think the graphic should only use polls from Dec, Jan, and Feb. Thoughts? Zwilson14 (talk) 19:20, 11 February 2012 (UTC)

Yea, this was mentioned above. I dont think the mapmakers check the talk page here much.--Metallurgist (talk) 03:23, 14 February 2012 (UTC)

Updated Poll for Tennessee

Tennessee is mis-colored. It is colored solid blue as of 5:35 p.m. PST. It should be blue and red to reflect the statistical tie between Santorum and Romney. The Rasmussen poll released today indicates that Romney and Santorum are in a statistical tie with Santorum at 34% and Romney at 30%. The MoE is 4%. http://www.rasmussenreports.com/public_content/politics/elections/election_2012/election_2012_presidential_election/tennessee/2012_tennessee_republican_primary — Preceding unsigned comment added by 71.108.142.18 (talk) 01:37, 5 March 2012 (UTC)

That is just one poll. The polling numbers are different depending on which poll you look at. In the We Ask America poll that's on there right now, Gingrich, Santorum, and Romney were all within the MOE.- 6 March 2012, 1:32 PM CST

Updated Poll for Florida

The poll from Quinnipiac that was released on 1/8/11 is now included but Florida is still colored yellow and red (stripes). http://www.quinnipiac.edu/institutes-and-centers/polling-institute/florida/release-detail?ReleaseID=1687. The map should be changed to show that Romney now leads all other candidates in Florida by 12%, well outside of the 4.1% MoE.

Date for MiszaBot--Metallurgist (talk) 14:11, 6 March 2012 (UTC)

Map timeline?

Since having one final map that matches the results article would be a bit dull, would having a map timeline be helpful? By this, I mean a map of polls from Jan-Jun 2011, Jun-Aug 2011, etc. (the time periods can be chosen more carefully). Then, it would be possible to see the changes over time. Thoughts? Hamiltonian (talk) 01:55, 15 February 2012 (UTC)

Perhaps. That is an interesting idea...--Metallurgist (talk) 21:02, 15 February 2012 (UTC)

I think that's a good idea. Light-jet pilot (talk) 04:08, 16 February 2012 (UTC)

I love the idea. It would make this article great. If they was made month by month back in time it would illustrated how different candidates have surged in media and poll attention. I really wish I knew how to make maps Jack Bornholm (talk) 14:37, 27 February 2012 (UTC)

South Dakota

Shouldn't South Dakota be filled in with Ron Paul's color (whatever that may be) per the newest poll listed which was from this past December? He led Newt 22-19. Go Phightins! (talk) 16:19, 19 February 2012 (UTC)

No. Because Bachmann left the race, so the polling data is no longer relevant.Light-jet pilot (talk) 17:42, 19 February 2012 (UTC)

Why is Texas still filled in? Rick Perry is no longer in the race but finished third in the poll? Does it have to do with whether a candidate was within the MOE (the SD poll had a 5% margin and Paul was ahead 22-19-18 with Bachmann at 18)? Go Phightins! (talk) 21:15, 19 February 2012 (UTC)

Yes, I think that's why.Light-jet pilot (talk) 13:44, 20 February 2012 (UTC)

Two questions

1. Can everybody change the map? 2. How is it done? - link? Abe 92.192.81.220 (talk) 08:19, 27 February 2012 (UTC)

I would like to be directed toward an wiki article or program or something else that teaches you to make maps too. Any help? Jack Bornholm (talk) 14:39, 27 February 2012 (UTC)

1. Yes, anyone can edit the map with relative ease. That's kind of the point of svg use on Wikipedia as far as I can tell. In something like a jpg (and I think a png) people would just open it up in an image editor and use some sort of paint bucket tool to paint the region they want. But this can color too much or too little, and re-saving the image in those formats can lead to deterioration of image quality. On the other hand, you can just edit the code of an svg to change the color, so it's complete and slick, and pretty quick.

2. This image was based on the sam-sex marriage in the US map, which has a brief explanation of how to edit: [8]. The g ids used in this map are different but follow the same naming convention. All you need to do is download the latest version from Commons, open it up in plain text (notepad or textedit should do), find the block near the bottom that lists the state abbreviations followed by a line of text, and change the g id currently in use to the appropriate one. You should then be able to open it up in a browser to check that it worked, and then you can upload a new version of the file in Commons. ThrawnRocks (talk) 23:10, 28 February 2012 (UTC)

Thank you 92.192.37.72 (talk) 12:10, 29 February 2012 (UTC) Abe

Thanks. Now I have made some update to the map, how do I upload it? I only seem to be able to upload a new image, not simply update the current map? Jack Bornholm (talk) 00:21, 2 March 2012 (UTC)

This is my problem too.Light-jet pilot (talk) 00:40, 2 March 2012 (UTC)

You have to upload a new version on the wikimedia commons site. The link you want to click is at the bottom of the file history. —Torchiest ^talk_edits 03:39, 2 March 2012 (UTC)

What link? Light-jet pilot (talk) 04:15, 4 March 2012 (UTC)

It's at the bottom of the history of images. click this to upload. —Torchiest ^talk_edits 04:12, 5 March 2012 (UTC)

Iowa

Question: shouldn't Iowa show up as a tie? That's as close to a statistical tie as I've ever seen, and wasn't there something about not being able to find all the ballots? — Preceding unsigned comment added by 128.187.97.20 (talk) 16:33, 27 February 2012 (UTC)

Nope. Santorum is the certified winner. It may technically be a statistical tie, but its the "poll" thats taken as the results.--Metallurgist (talk) 10:38, 29 February 2012 (UTC)

Wyoming

Since Wyoming is being counted as a Romney win, shouldn't it have the results under "Polling for completed primaries"?Light-jet pilot (talk) 17:49, 2 March 2012 (UTC)

Except there weren't any polls for Wyoming that we know of. And the results are already in the results article. —Torchiest ^talk_edits 18:38, 2 March 2012 (UTC)

But they did it with the Missouri section. Light-jet pilot (talk) 22:50, 2 March 2012 (UTC)

True, but there have been polls conducted there. I don't have any objections, I'm just guessing at to the reason. But it does seem a little redundant with Results of the 2012 Republican Party presidential primaries. —Torchiest ^talk_edits 22:58, 2 March 2012 (UTC)

Yea, there have been no polls on Wyoming.--Metallurgist (talk) 12:59, 5 March 2012 (UTC)

Updated Poll for Illinois

Illinois should be colored red. It is not colored at all as of 3/11/12 at 2:35 a.m. PST. It should be colored red to reflect the results of a new Chicago Tribune poll released on 3/10/12 that indicated that Romney is now leading in this state by 4%. The MoE is 4%. http://www.chicagotribune.com/news/local/ct-met-illinois-poll-20120311,0,1617767.story — Preceding unsigned comment added by 71.108.142.18 (talk) 09:37, 11 March 2012 (UTC)

Alabama Poll

Alabama shouldn't be solid red. Romney's lead is within the MoE of Gingrich and Santorum.-13 March 2012 6:55 CST

The MoE is no longer being considered in that way. See the discussion here. Light-jet pilot (talk) 02:34, 14 March 2012 (UTC)

Are we just doing who is in the lead now? (Because that's even worse.) Hamiltonian (talk) 06:16, 14 March 2012 (UTC)

I guess so. :( It doesn't help that the only (right) option can only be figured out by someone with a political science degree. I still don't understand it. Sigh... Light-jet pilot (talk) 13:54, 14 March 2012 (UTC)

It's actually not too complicated if it's explained (it rarely is, though). Let's see if I can do it. (1.) When you take a poll, you have a margin of error. It is usually reported with 95% confidence (basically all polls are). If we accept that and move on, it makes the explanation a bit simpler. (2.) The reported margin of error means that 95% of the time the reality of the situation will be somewhere between the poll's value, plus or minus the margin of error. (3.) Let's say we have three candidates, Al, Bernice and Carla. We have a poll that says Al 37%, Bernice 32%, Carla 21% (10% undecided/other). The margin of error is reported to be 4%. (4.) Who is in the lead? Well, the polls says that Al is 37 +/- 4%... so he's somewhere between 33 and 41. Bernice is between 28 - 36. Carla is between 17 and 25. What can we say? We know that either Al or Bernice is in the lead. Why? Al could be 33, and Bernice could be 36. We know that Carla isn't in the lead... even if she is at 25, that still lower than the lower value of Al or Bernice. (5.) What if we were highlighting who could be in the lead? Well, you'd highlight Al or Bernice, but not Carla. Does this make sense? Hamiltonian (talk) 16:27, 14 March 2012 (UTC)

Highlighting was removed from the charts. The map was a separate issue, and it was settled upon just marking the most recently individual poll leader with the map colors. I'd prefer doing multiple color striping as was being done before. Since changing the map is relatively easy, and is done frequently, I see no reason why it couldn't be done using the correct, wider MoE rules to determine it. —Torchiest ^talk_edits 17:43, 14 March 2012 (UTC)

Good explanation. Thanks!! Light-jet pilot (talk) 21:15, 14 March 2012 (UTC)

Margin of error

How are we interpreting margin of error? Are the candidates highlighted if they could be within the margin of error of the lead? Let's take the latest Arizona poll as an example: Romney 38 Santorum 31, MOE 4 (only Romney highlighted). Anyhow-- with this margin of error Romney could be anywhere 34-41, and Santorum could be 27-35. That is, Romney could be leading 41-27... or Santorum could be leading 35-34. Should both be highlighted/coloured in? (This isn't nearly the only example.) Hamiltonian (talk) 01:39, 16 February 2012 (UTC)

Well, yes, the candidates are highlighted if they are within the margin of error of the lead. Thus, Santorum, even if givin the 4 point MoE, would only be 35% to Romneys 38%. If he were able to catch Romney with the MoE, we would highlight them both. In your example, you counted the MoE twice...Santorum up 4 AND Romney down 4. But it doesn't work that way. At least, that's how it was explaned to me. Light-jet pilot (talk) 04:15, 16 February 2012 (UTC)

I've actually had some confusion about this myself. But if you look here, it seems to indicate that the MoE is for each percentage. Thus, Hamiltonian's reading of the poll is correct. Each number can be 4% above or below the reported value. We've been counting it as half that value for all intents and purposes. Correcting this would be a massive undertaking at this point, and a lot more polls will be statistical ties. —Torchiest ^talk_edits 04:50, 16 February 2012 (UTC)

MoE isnt even technically calculated correctly here in samples with more than 2 choices. Hamiltonian is technically correct that both can range that way, but generally the way statistical ties are determined are as it is now (if they difference between the totals is less than the MoE).--Metallurgist (talk) 06:20, 19 February 2012 (UTC)

I don't mean to push the point-- but I'm not only technically correct, but I'm also actually correct. It's calculated by looking at the range, and seeing if they overlap. I've never actually seen it done the way you describe, and I work with polls all day. --Hamiltonian (talk) 17:05, 19 February 2012 (UTC)

I cant point to anything in particular, but Im pretty sure Ive seen polls like I described as not statistically tied. Maybe on 538. Saying a poll that is 50-42 with 4.4 MoE is a statistical tie is just crazy talk. Pretty much everything would be a statistical tie then.--Metallurgist (talk) 01:04, 21 February 2012 (UTC)

Sadly, all sorts of polls are statistical ties. "Statistical significance" really relies on how confident you want to be. The margins of error are reported for 95% confidence. A poll with a 50-42 margin, and a 4.4 MOE at 95% confidence sadly can't be said to be different. Now it's unlikely that they are the same.. there's something like a 90% chance that they are different. Unfortunately, that's not the standard confidence interval-- it's 95%. The candidate at 50% can only be estimated to be between 45.6 and 54.4% 19 times out of 20, while the candidate at 42 can only be said to be between 37.6 and 46.4 ... so there is a reasonable chance that they are in fact tied. Does it mean more ties? Yes. Is that a bit boring? Yes. Alas... --Hamiltonian (talk) 01:11, 21 February 2012 (UTC)

Yea, I know the statistics, but thats not generally how its applied. Perhaps in a math wing of a university, but to say 50-42 is tied is absurd and nonsensical.--Metallurgist (talk) 09:10, 22 February 2012 (UTC)

"to say 50-42 is tied is absurd and nonsensical" No, it's not-- if the margin of error does not say so. A poll cannot be said to show a lead, if the margin of error is 4% and the lead is 8%. This isn't an obscure statistical point, or something dreamed up by academics-- it's basic political science. Do newspapers and TV networks get this wrong all the time? Yes. It is maddening. Does it mean that this page has hundreds of errors? Yes. Does it mean that there are candidates colored as being the in lead when they cannot be said to be in the lead? Yes, sadly. --Hamiltonian (talk) 17:49, 22 February 2012 (UTC)

Proposal: Remove shading

Based on the above discussion, I think the simplest solution would be to remove the background shading completely instead of hashing out which method of highlighting to use. The candidates are already listed 1st, 2nd, 3rd, and other. Coloring is somewhat redundant anyway. I agree with Hamiltonian entirely about how the MoE works, but I think that since there is a dispute, the easiest thing would be to the let the reader make their own interpretation based on the reported numbers and what the polls' MoE are listed as. —Torchiest ^talk_edits 18:16, 22 February 2012 (UTC)

I agree with what you say(no shading), but the MoE is used to color the map, so we still have to decide.Light-jet pilot (talk) 03:19, 23 February 2012 (UTC)

For the map, We could just go with the 'winner' in each poll unless there is an actual tie.--Metallurgist (talk) 10:14, 23 February 2012 (UTC)

Sure...that's awesome. It might be good to count an automatic 1% MoE, like "43% to 44%" and "40% to 40%". But if that's too complicated, than your idea is still good.Light-jet pilot (talk) 13:59, 23 February 2012 (UTC)

I removed the shading-- it continues to be done incorrectly, providing readers with misleading info. --Hamiltonian (talk) 17:07, 2 March 2012 (UTC)

Does this mean we should now reinstate old polls on the map where there was a statistical tie with a dropped out candidate? For example, the last poll in Mississippi from November had Newt Gingrich in the lead. But on the map, Mississippi is grey because Newt was statistically tied with Herman Cain, who dropped out. South Dakota is in a similar situation. --Noname2 (talk) 19:52, 3 March 2012 (UTC)

The map shading is a separate issue from the chart shading. I don't think it makes sense to color states according to polls with candidates who aren't in the race now. —Torchiest ^talk_edits 23:06, 3 March 2012 (UTC)

I think the map should be consistent with the article. So if the article isn't going to highlight statistical ties then neither should the map. --Noname2 (talk) 23:26, 3 March 2012 (UTC)

Shading should not be removed. It makes skimming the page a lot easier --Lolthatswonderful (talk) 18:49, 17 March 2012 (UTC) So are we settling on keeping the shading in now? --132.208.24.181 (talk) 15:37, 18 March 2012 (UTC)

Proposal: Fix shading

No offense but polls do contain actual information. As someone else said, they're misinterpreted A LOT but Wikipedia does not have to misinterpret them or effectively make the article less useful if you want to just scroll down the screen and see who the winners are. What I do find strange is that we're using red shading for the winner by state and other colors for the overall map shading. I'd suggest we use the various colors for the candidates instead--again it makes it easier to just browse the list visually. As far as statistical ties I think we should just not shade when that is the case, although that would leave relatively few of the races shaded, but that is actually more true to reality. Theshibboleth (talk) 00:06, 25 February 2012 (UTC)

WIN vs. win (map question)

Is there a simple solution to determining at a glance a "primary win" vs. a "recent poll" on the map? Light-jet pilot (talk) 00:13, 22 February 2012 (UTC)

I was thinking about this. Not sure how to handle it. I think we are planning to have a timeline of maps once the primaries are over.--Metallurgist (talk) 09:07, 22 February 2012 (UTC)

It's just hard to double check the numbers at the top, the way it is now.Light-jet pilot (talk) 03:21, 23 February 2012 (UTC)

Yea I know. Not sure how to resolve this. Maybe stripe it or something.--Metallurgist (talk) 10:10, 23 February 2012 (UTC)

Stripes seem to cluter a lot, someone suggested putting a star on the state when a color is final...which would be fine with me.Light-jet pilot (talk) 15:16, 23 February 2012 (UTC)

A star would be good I think. Go Phightins! (talk) 22:23, 23 February 2012 (UTC)

Or that same flag symbol used in the body of the article. —Torchiest ^talk_edits 22:47, 23 February 2012 (UTC)

I vote for a star.--Metallurgist (talk) 07:35, 26 February 2012 (UTC)

I vote for a star too Jack Bornholm (talk) 14:31, 27 February 2012 (UTC)

Go for the stars 92.192.126.161 (talk) 07:04, 2 March 2012 (UTC) Abe

Is someone doing this? Tiller54 (talk) 01:07, 4 March 2012 (UTC)

I have tried, but can't upload the new version.(see question below)Light-jet pilot (talk) 04:14, 4 March 2012 (UTC)

Has there been any progress on this? Tiller54 (talk) 20:27, 18 March 2012 (UTC)

I believe I can... Light-jet pilot (talk) 21:58, 18 March 2012 (UTC)

I'd rather this be done in such a way that you can add stars by editing the file in notepad, just like you can currently change the color of states by editing in notepad. --Noname2 (talk) 00:10, 19 March 2012 (UTC)

Me too. I wish I knew how. All I did was edit it in "InkScape", which is free to download. Light-jet pilot (talk) 00:34, 19 March 2012 (UTC)

That's probably why no one has changed it till now. Light-jet pilot (talk) 00:37, 19 March 2012 (UTC)

Maps

I can now edit the maps, so if something isnt being updated, leave a message on my talkpage.--Metallurgist (talk) 01:38, 25 March 2012 (UTC)

Me too. Light-jet pilot (talk) 02:24, 25 March 2012 (UTC)

Awesome!--Metallurgist (talk) 14:48, 25 March 2012 (UTC)

Pennsylvania and North Carolina are both Mis-colored

A poll released two days ago on April 3, 2012 indicates that Mitt Romney now has a 6% advantage over Santorum in that state. According to the rules of this wiki, any lead now justifies the coloring of that state for that candidate only. Link to poll: http://www2.highpoint.edu/articles/show.php?id=2857

Also, as is indicated by a PPP poll that was added to this wiki yesterday, Pennsylvania is now outside the MoE for Romney. Therefore, Pennsylvania should be red rather than the solid powder blue that it is today. Link to poll: http://www.publicpolicypolling.com/pdf/2011/PPP_Release_PA_404.pdf

Please fix both problems today. — Preceding unsigned comment added by 71.108.142.18 (talk) 14:24, 5 April 2012 (UTC)

You were correct about Pennsylvania, however, a newer poll has been released having them both within the MoE. As for NC, the MoE for that poll is 6.5%, so the map is correct. "According to the rules of this wiki" is a POV phrase, since there has not been concensus on the MoE issue. (See the former discussion here) Light-jet pilot (talk) 19:32, 5 April 2012 (UTC)