Jump to content

Talk:Statewide opinion polling for the 2008 Democratic Party presidential primaries/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1

Polls and dates

When you add a poll that would gives or adds a state to candidate make sure to note the change in the Current Leaders sections. Also, under that section, post the date of accuracy.Rougher07 21:44, 2 February 2007 (UTC)

Map

I reverted back to the DEM PRI MAY 1.JPG image. The Demprim.PNG image takes the strange approach of including middle names for some candidates (Obama and Edwards), but not for others (Clinton and Richardson). It could easily be construed as being biased: using Edwards' full first name , "Johnny" or Obama's middle name "Hussein" as if they mean something. Perhaps emphasizing youth and a Muslim parent respectively. There's obviously nothing wrong with either name, youth, being Muslim, and/or having a Muslim parent, and use of full names might be reasonable if it was applied to all candidates, but I think the most commmonly used name and/or the name applied to the wikipedia article is most reasonable. --Aranae 23:16, 1 May 2007 (UTC)


To suggest that using somebody's formal name is biased is stupid. "Rodham" is included as is "Reid" as is "Hussein". I didn't know Bill Richardson's middle name was Blaine, I'll include it next time. Johnny is John Edwards' *real* first name... that is why it was being used. I will next time include Blaine for Richardson, but using a far uglier map because it uses Barack Obama's middle name is stupid. If you're offended by his middle name, you don't get to rewrite it or rewrite history or facts.

Isaiah13066 23:41, 1 May 2007 (UTC)

Why don't we just use what the candidates use when they register their names on the ballots? Behun 02:39, 2 May 2007 (UTC)

I agree with Behun on using names on the ballot. My point was that using full names for Edwards and Obama, but not writing Hillary Diane Rodham Clinton and William Blaine Richardson III is just weird. Full names for everyone is perfectly fine, but I don't think half and half is NPOV. Do people care how Clinton dealt with her name after marriage? Do people care that Richardson is "the third"? The answer should be no, but may not be. As this page pertains to an election and is rather high profile, treating the candidates equally is particularly important and simple NPOV. --Aranae 04:11, 2 May 2007 (UTC)
I have edited the graphic for correctness, aesthetics, and use of common name. Let me know if this works for everyone. --JakeZ 23:03, 2 May 2007 (UTC)


JakeZ thanks for the new map which has a higher quality; in future modifications I will change the color scheme to make it more colorful and contrasting, unless you do that. I also ran into a problem: changing color with the map you have is virtually impossible... my previous map enabled you to easily change the fill of each state.

The middle name "biased" argument was absolutely stupid.

Isaiah13066 00:03, 3 May 2007 (UTC)

I thought common name or ballot name was more accessible for everyone. It's what the media uses and is what the public identifies with. I don't really have a stance on whether middle name should be there from an NPOV perspective. I care about people knowing and understanding who's who. As for the colors, I can boost up the contrast. Actually, what would help is if you could pick out a color scheme from http://kuler.adobe.com. JakeZ 02:22, 3 May 2007 (UTC)

These maps have not been updated in months and do not accurately reflect the current political climate. I have removed them. Jkhall (talk) 19:49, 10 December 2007 (UTC)

The map needs to be updated: the most recent polls show Iowa, New Hampshire, and South Carolina to all be a close tie between Obama and Clinton. 64.119.13.45 (talk) 19:00, 1 January 2008 (UTC)

Someone has removed the map and pie graph completely, so I have removed the precis about them until they are replaced (if they are replaced). 64.119.13.255 (talk) 04:48, 4 January 2008 (UTC)

Could you put a date on the map. This is not updated and seems not only pretentious, but also self indulging. 75.161.117.4 (talk) 04:12, 1 February 2008 (UTC)


  • Over the past four days there has been four polls released for California. Yet when one single poll today is released with Obama ahead the state gets changed over to him? Try averaging out the polls or assume that the other three maybe more accurate over just one. —Preceding unsigned comment added by 69.123.81.12 (talk) 17:46, 3 February 2008 (UTC)
  • Such arguments were not voiced when Alabama switched from Obama to Clinton earlier. Who precisely would determine which poll is more accurate, and by what qualifications. The current map, not to be confused with the map seen earlier, has always been updated to the latest poll. If you would like to change this method, provide a specific and detailed rubric on how we should measure previous polling, for how far back, how we determine which polls are most accurate, and why such a system is any more reliable then one it is replacing. —Preceding unsigned comment added by Chester-S (talk • contribs) 20:04, 3 February 2008 (UTC)
  • Who ever makes the map should average out the lastest polls from within say a week of each other. Its just curious that since polling started in California last year it's been Clinton in every poll, and as soon as one poll comes out that has Obama in the lead that poll all of a sudden becomes the the poll to base the map on.
  • I don't think it is curious in the least, since thats how the map has always operated whether or not it has moved in Clinton or Obama's direction. In the fast moving news and media cycle surrounding the political cimate right now, why would polls a week ago be as relevent as the most recent poll? I think the New Hampshire polls are an excellent example of why an average, while sounding good, is not. —Preceding unsigned comment added by Chester-S (talk • contribs) 21:05, 3 February 2008 (UTC)
  • The idea of averaging recent polls is that no one polling company is the most accurate. If you average for example the most recent Rasmussen, Field, Mason-Dixon and Reuters/CNN polls, I think you have a fairer look at the data then to just go with the one that maybe was released 2 hours later then another, it doesn't make it the most accurate. People may come here see the map and assume Obama is the clear leader of the state when he's not.
  • Why does the map not show Michigan and Florida? The polls below include the states. I know that at this time the delegations from the two states wont be seated but that could change. —Preceding unsigned comment added by 69.123.81.12 (talk) 21:41, 3 February 2008 (UTC)
  • Every single news media outlet includes Michigan and Florida when they show which states have been won by the different candidates. In addition, the title of this entry contradicts not showing these two states, since both had extensive polling done.

I move to delete the map -- several polls are conflicting, esp eg California. Thewebthsp (talk) 16:23, 4 February 2008 (UTC)

  • I think the map is a decent enough feature - it's certainly useful for anyone who doesn't want to wade through the data to see what the overall picture is! However, I agree with the poster above that the "last poll" methodology is somewhat crude - it makes the map too volatile and is open to too much interpretation (if two polls come out on the same day, which is the last one?) I think a rolling average would be too difficult to keep up to date and too prone to errors - so what I'd suggest is the following: Each poll is scored as Obama, Clinton or Tie (if the lead is less than the margin of error). Then every time a new poll is added, the person adding it does a simple count of the number of "wins" each candidate has in the last three days' polls. If either candidate has a majority, they are awarded the state. If not, it's scored as a tie. If there are no objections, I am happy to go through and do this for the existing data. Umeeksk (talk) 20:06, 4 February 2008 (UTC)
  • Sorry but the map is off, all polling in Oklahoma has shown Hillary consistantly ahead by large margins and the most recent was taken 2/2-2/3. Also the most recent Mass polls show Hillary clearly in the lead the only poll showing Obama ahead is w/in the margin of error. Maybe it is just me but the map seems highly biased toward Obama when OK is not shown and Mass is colored for Obama with one poll vs every other poll. What is worse is DE is colored for obama yet NOT ONE poll has shown him leading, I'd change the map right now if I knew how.Tulsaschoolboard (talk) 23:45, 4 February 2008 (UTC)
  • There is extreme Obama bias going on here. As stated a recent poll not including Edwards shows Clinton easily in the lead in Oklahoma as well as new SurveyUSA polls showing Clinton with solid leads in Massachusetts and California yet the maps show Obama as the clear leader in both states. These two states should atleast be colored like New Jersey. Furthermore, Ohio should be colored with Clinton based on a new poll there, even though the poll was partially completed while Edwards was still a candidate even if Obama got all the votes in the poll that Edwards had he would still be behind Clinton. (Tjliles2007 (talk) 00:41, 5 February 2008 (UTC))
  • I truely apologize. Any bias was unintentional. Some of the numbers you are refering to had not been added when I set off on my editing journey in GIMP (graphics programme) this morning. I hope my latest update resolves your concerns. Infoporfin (talk) 01:16, 5 February 2008 (UTC)
  • All the numbers I used were correct and the latest according to this version of the article. That excludes the mistake I made with Missouri, where I copied the mistake of the last person who updated the map, and Deleware, which I just goofed up on. Sorry again. I added Oklahoma for HRC. The latest poll was not in the version of the article I was looking at. Infoporfin (talk) 01:33, 5 February 2008 (UTC)

I think the map would be better after tomorrow, with established results. It looks really messy at the moment. Thewebthsp (talk) 02:10, 5 February 2008 (UTC)

Quite a bit of this controversy could be avoided by not coloring in states with statistical dead heats or conflicting polls. It is a blaitant misuse of polling data to indicate a leading candidate when the two are within the margin of error of one another. At the very least some distrinction should be made between states were the candidates are very close vs. those in which there is a clear leader. —Preceding unsigned comment added by 134.173.90.52 (talk) 05:44, 5 February 2008 (UTC)

  • There's a dead-heat in MA? You have to be kidding me. The map should go down altogether for today, because its too prone to manipulation. Either that or making changes to it should be locked once its finalized. —Preceding unsigned comment added by 24.215.168.121 (talk) 08:30, 5 February 2008 (UTC)

I'm finding it difficult to easily read the map's encoding of state-wide caucus results (center dot either purple or brown) on top of popular votes and polls (background either purple, light purple, brown, or light brown). I think the caucus results and popular votes should be left out of the map. After all, this is the Opinion Polling article; the popular votes and the caucus results have their own maps at Results of the 2008 Democratic presidential primaries. Wdfarmer (talk) 10:10, 5 February 2008 (UTC)


Thanks for adding the "no clear leader" category to the map! It really does a more accurate job of reflecting the poll data. I would add that Arizona and New Mexico should also be colored "no clear leader" because the latest closing poll in AZ is within the margin of error, as is the only poll available for NM. Thanks again! Szu (talk) 17:28, 5 February 2008 (UTC)

  • The map states the shaded colors indicate popoular vote but then the key at the bottom says it's delegates that count? Whoever is making this map is obviously biased in favor of Obama. This is the only map on any respectable news outlet that doesn't show the results of Michigan and Florida and shows the winner of the states according to delegates instead of popular vote. Someone that knows how to make maps--and is not biased toward either candidate--should replace whoever made this one and depict honest and unbiased results.


Hi again! The polls for Oregon, Maryland and Ohio include John Edwards... I'm not sure these should be included in the map, as they do not reflect the current two-way race that the other polls do. Thoughts? Szu (talk) 23:24, 8 February 2008 (UTC)

  • I used data from recent polls in Maryland that do not include John Edwards, but you are right about Oregon and Ohio; that is why I did not include them in my map (proportional) and removed them from the winner-take-all map. Infoporfin (talk) 05:04, 10 February 2008 (UTC)
I see the MD poll data has been added. Thanks for all your work on this map! Szu (talk) 20:44, 10 February 2008 (UTC)
  • John Edwards suspending his campaign is more impactful than simply shifting 100% of his support to Obama or Clinton or his support being split between them. It can seriously impact the way those 2,156 people who answered that poll OVER A WEEK ago are thinking about voting ALMOST A MONTH FROM NOW. This poll was taken before super Tuesday and the Feb 9 contests. This is an extremely fast-changing race. That poll isOLD! Infoporfin (talk) 05:04, 10 February 2008 (UTC)
  • A new map was added showing piechart distributions of the delegates awarded to each state. I moved it from the top to the last section since the top of this article is opinion polling. (Tjliles2007 (talk) 04:16, 10 February 2008 (UTC))
  • Interesting justification in that the map that you replaced it with also shows election results and not strictly opinion poll results. The map you removed clearly distinguishes between election results and opinion poll data. A Proportional map was also requested in discussion on this very page. Infoporfin (talk) 05:06, 10 February 2008 (UTC)
  • Okay, I changed the color scheme of both maps to heavily deëmphasize primaries and caucuses and emphasize opinion poll results. Please do not revert my edits without developing consensus on this talk page. Infoporfin (talk) 07:35, 10 February 2008 (UTC)
  • Can you change the coloring to what was used on the original map? I know its not much of a big deal, but like if you look at the very small states on the east coast its hard to see the color difference...like Connecticut... with the other colors it was easier to see. (Tjliles2007 (talk) 17:22, 10 February 2008 (UTC))
  • In response to this suggestion I put boxes in the water of the Caribbean to represent DC and the smaller east coast states. I changed the colors in the first place because you thought the image was not specific to opinion polling (the subject of this article). I made polling data stand out more and the results of primaries and caucuses that have already happened less noticeable. Infoporfin (talk) 20:54, 10 February 2008 (UTC)
  • Why can't we leave the second map on the right side as well.... the formatting looked much better that way then it does now with it set to the left side. Also why do you guys keep leaving the second part of the caption off every time you alter it? (Tjliles2007 (talk) 17:20, 10 February 2008 (UTC))

Polls are contradictory for Wisconsin -- This poll "esitmator" map cannot at all be considered accurate. The color coded map would be better to include a color for both Clinton and Obama.Thewebthsp (talk) 19:47, 13 February 2008 (UTC)

  • Can someone please color Texas both yellow and purple? Some recent polls go for Obama, some for Clinton, and some are within one point or tied. While this is not the case for Ohio (no poll, to my knowledge, has shown Obama as the winner) I think Texas on the map should reflect the mixed poll numbers. MoChan (talk) 21:27, 2 March 2008 (UTC)

Keep Obsolete Data off The Map

Please stop adding data to the map for states that have not had recent polls. The poll for Oregon includes John Edwards. As a local, I can assure you that statewide popular opinion has changed dramatically since then for two reason: John Edwards has suspended his campaign and Super Tuesday has taken place. In the caption of the map it says only polls reflecting the suspension of John Edwards' campaign are used in the map. If you change the map, at least change the caption to be accurate. Infoporfin (talk) 19:56, 9 February 2008 (UTC)

Map Colors

In case anyone is interested, I chose the current colors of the maps (yellow and orange) to have as little racial/gender/partisan implications as possible. They are also distinguishable to people with several types of color blindness according to this Color Blindness Simulator. Infoporfin (talk) 05:23, 11 February 2008 (UTC)


  • Your colour scheme is awful and confusing. Yellow and orange are so similar that it is hard to read the map. I think more people would object to having a poorly coloured map than they would to having a colour scheme that might have partisan implications. There are a lot more options for colour schemes than red/blue (connotes republican/democrat) or black/white (connotes race). 140.180.25.247 (talk) 20:41, 11 February 2008 (UTC)
  • To me, it's more complicated than avoiding red, blue, black, and white. There's a Green Party in the United States and the color green suggests someone has greener politics. Red and pink have traditionally stood for leftist beliefs while the late 20th Century meme in the US is Red for Republicans (right) and Blue for Democrats (left). Pink and purple have long been considered "girly." As for black and white, while people are shunted into theses racial categories, noöne's skin is actually paper white or ink black. This being said, colors (such as the yellow and purple that I have substituted for yellow and orange) that tend to approximate the actual candidates' (Clinton and Obama) skin colors are problematic to me. Then there is the light=good/dark=evil question. Infoporfin (talk) 23:08, 11 February 2008 (UTC)
  • I have changed the colors to purple and yellow; is this an improvement? I'm in my twenties and found the yellow and orange easy to distinguish, but I guess it depends on the person and xyr monitor. Likewise, for the dark version of each candidates' color, I chose a yellow and an orange that, to me, looked equally dominant on my monitor. I gave great importance to the colors looking equally "strong" and "weighty" but again, I'm sure it depends on the person and monitor. If you have suggestions for other colors, please state them using RGB numbers or upload an example. :) Infoporfin (talk) 23:08, 11 February 2008 (UTC)
  • I think the colour choices are good choices for equality in "strength". I think it's strange, however, that you've given the dominant darker colour to undecided states. Wouldn't it make more sense to give more "weight" to states that have actually been decided than to those that are just projected? Also I would like the decided states to be more distinguishable from the grey states. I'm sure you've put a lot more thought into this than I have, but I would've made the decided states in the dark colour and the upcoming ones in (dark)colour-and-(light)grey stripes. ErikRydbert (talk) 23:41, 13 February 2008 (UTC)

Population-Proportionate or Electoral Vote-Proportionate Map?

It might be fun and interesting to include results on a population-proportionate map such as this: http://www.georgehernandez.com/h/aaBlog/2004/media/11-09_Election-StateCartogram.png
or an Electoral Vote-proportionate map such as this: http://files.blog-city.com/files/A05/141484/p/f/statecartogramelectoral.jpg --138.88.90.41 18:44, 17 May 2007 (UTC)

Wrong manner of representation

The grouped bar graphs are a bad choice for the representation of the data in this article. Something like a line graph, or even an area graph over time would be better. The grouped bar graphs are illegible and I suggest they be changed. 69.252.188.131 16:47, 20 May 2007 (UTC)

[copied from image talk page:] This graph is worse than useless, a rank, sour shit splattered across the face of an otherwise informative article. I hesitate to offer suggestions of any potency to such a clearly communications-impaired dope as whoever would claim to have "designed" this graph, but even a cursory glance at any of Edward Tufte's writings—not that I necessarily agree with his every dictum—might have spared the world from this horror of visual noise that you've inflicted on an innocent, unsuspecting readership. Good fucking grief. Tickletaint 02:26, 31 May 2007 (UTC)

I have removed the bad bar graphs. Jkhall 04:22, 2 December 2007 (UTC)

Mississippi

As with the Republican map, Mississippi is colored for a candidate (Clinton), but we have no polls for Mississippi. Are we missing a poll or is there an error on the map? --Aranae 04:30, 22 May 2007 (UTC)

Can We Please Remove Datamar Polling

Visit the following: http://www.pollster.com/blogs/poll_datamar_florida_primary.php

"I looked at the pdf and nothing was divulged. A "proprietary algorithm" of a "likely turnout model based on election cycles and other factors." Very strange. "Barak" Obama was misspelled--not promising from someone who supposedly deals with political names on a constant basis.

Democratic numbers are especially "eyebrow-raising."

The few demographics it did divulge were simply hilarious ... This is polling malpractice if anything is. Sounds like Edwards needed a poll to show donors, even if it was utterly fraudulent."


Pollster.com is one of the most respected websites online.

Isaiah13066 23:50, 25 May 2007 (UTC)

I think including the poll and then citing that it is controversial with a reference is probably the best to deal with that sort of problem. I think the same would be true with the anon user who keeps deleting the Insider Advantage poll in South Carolina. If there's a problem with a poll, someone notable besides wikipedia editors should have called it suspect and we can then list it and mention the controversy (and probably not include it in our tally at the bottom). --Aranae 03:52, 26 May 2007 (UTC)

Pollster.com may be respected, but anyone can leave comments. The quoted portion above is from the comments by someone who seems to claim that Edwards is behind these 'fraudulent' results.

Proportional Representation and the "Current Leaders" section

The current leaders section creates the false impression that the race for delegates in each state is "winner take all". Rather, any candidate who garners a 15% or larger share of the votes in each state primary is entitled to a share of the delegates proportional to the votes earned.

As the section stands now, it shows Clinton having a substantial lead in potential delegates earned, when no poll shows that to be the case.Blwarren713 11:48, 6 June 2007 (UTC)

If this is true that some or all states will allow their delegates to vote as they please or in proportion to the vote in the primary/caucus, then I wholeheartedly agree that the map should change to reflect this reality, and the "results" at the bottom of the page should be adjusted accordingly.--Robapalooza 17:46, 7 June 2007 (UTC)
I've corrected the Current Leaders section to update the number of delegates assigned to each state. Specifically, based on various sources including The Green Papers, California will have 440 delegates, Florida 210, Nevada 30, South Carolina 54, and Iowa 56. All other states were correct.--Robapalooza 20:43, 14 June 2007 (UTC)

Addition of Proportional Representation (DNC's 15% Threshold) Sub-Sections in the "Graphical Representations" Section and in the "Current Leaders" Section

In response to some of the astute concerns raised above, and in accordance with the 15% Threshold rule as set forth in "DELEGATE SELECTION RULES FOR THE 2008 DEMOCRATIC NATIONAL CONVENTION," Issued by the Democratic Party of the United States, Governor Howard Dean, Chairman, As adopted by the Democratic National Committee, August 19, 2006 Delegate Selection Rules, I have added two new charts to the "Graphical Representations" section, and a new Table to the "Current Leaders" section, which apply the 15% rule. That is, if a candidate receives less than 15%, he or she will not be eligible for delegates from a particular state. For those candidates who achieve the 15% threshold, the delegates will be apportioned proportionately. As a result, the previously-observed over-representation of Hillary Clinton due to a "winner takes all" approach has been mitigated and, I believe, these charts and tables provide a more accurate picture of the current race. Please note the large number of "undecided" voters and the number of states that lack polling data altogether (which I've included in the "undecided" category). I believe this data to be objective and unbiased. However, please note, I've retained the "winner takes all" graphical data and table, because there will be 3,515 pledged delegates which will be selected by primary voters and caucus participants (which might be predicted by poll results), and there are 852 unpledged delegates, colloquially known as superdelegates, which are DNC members, Democratic members of Congress and Governors, and other important figures in the party. Therefore, it's not clear whether "winner takes all" or "proportionate" is more accurate. Over time, I plan to modify this analysis with the best available data. For example, some analysits and bloggers are predicting the votes of various superdelegates based on their endorsements. In any event, "predicting" the result is difficult and any attempt to graphically summarize data is going to be filled with assumptions. I hope by providing an analysis based on both "winner takes all" and "proportionate" (subject to 15% threshold) analysis, I've at least improved what we're presenting here. Any suggestions for formatting or overall improvement are welcome. --Robapalooza 22:19, 7 June 2007 (UTC)

Fantastic work, Robapalooza! You've performed a great public service, and you deserve due appreciation. Black Regent
Updated to include new polling data from NH, FL, NC, TX and PA. Also, correction of OH.--Robapalooza 20:24, 14 June 2007 (UTC)
Updated--Robapalooza 18:19, 6 August 2007 (UTC)

moveon.org straw poll

User 01001 added a moveon.org straw poll showing Obama and Edwards leading Clinton. While this is certainly relevent to the subject of the present article (opinion polling), moveon's poll has almost nothing to do with the actual election process, which, whether we like it or not, is a state-by-state affair, where the rules differ widely from state to state (and in the case of Florida, where an entire state's results may be voided by the Democratic National Committee, because Florida isn't playing by the national rules in setting the date for their primary [1]) and where 19% of delegates are "superdelegates" that aren't bound by the outcome of the vote on a given primary day. I vastly prefer state-by-state opinion polling, because, in theory, it gives you a good idea of which candidate is likely to carry some or all of the state's delegates to the nominating convention. Should we keep the moveon section? I say just add a link at the bottom of the article to the poll but not include it as the #1 item in the article. --Robapalooza 17:25, 19 June 2007 (UTC)

There are no other national polls on this page, nor DailyKos straw polls etc. The title of the page is Opinion Polling for the Democratic Party (United States) presidential primaries, 2008. Rather than ask for presidential preference, it asked 'which candidate would be best able to lead the United States out of Iraq.' It doesn't seem like the MoveOn straw poll belongs here.

Idaho poll

A poll was conducted in Idaho and the conclusion was that a candidate who is not running received the most votes. Our job is not to interpret, but to report. This poll should remain, particularly since the same poll (without Gore) is included. --Aranae 02:32, 1 August 2007 (UTC)

If we look at the Republicans, Fred Thompson is not an official candidate but he is included. Therefore, I feel that Al Gore should also be included. Behun 02:42, 1 August 2007 (UTC)

Gingrich and Gore are also included in many polls. The only thing I can manage to come up with as to why objections rae raised is that he actually won in that poll. --Aranae 03:33, 1 August 2007 (UTC)

The poll has been removed again. --Aranae 22:01, 1 August 2007 (UTC)

New Chart - First Five States (Iowa, Nevada, New Hampshire, South Carolina and Florida)

Given the relative importance of the first five contests in Iowa, Nevada, New Hampshire, South Carolina and Florida, I've added a chart showing Clinton v. Obama v. Edwards v. Undecided/Other using the DNC's 15% threshold. In particular, I think this illustrates the fact that Obama has closed the gap in these states and that this race is not a yet a coronation for Clinton.--Robapalooza 01:22, 8 August 2007 (UTC)

Added Chart for Iowa - Edwards v. Clinton v. Obama

I added this.



In the near future, I will try to prepare similar charts for the other four early primaries, i.e. NV, NH, FL and SC. --Robapalooza 22:28, 6 August 2007 (UTC)

Updated data 8/18/07.--Robapalooza 23:00, 18 August 2007 (UTC)

Proposed Chart Revisions

The charts for the first four states only includes Clinton, Obama, and Edwards. This is clearly biased toward those three canidates, and perpetuates the notion that this is a race between only three people. Please remove them or update them to include all the candidates in the polls.Mathnsci 23:13, 25 August 2007 (UTC)

Thank you for your input. In defense of the charts in their current form, (1) The DNC has a 15% threshold rule, so unless a candidate can achieve at least 15% of the vote in any one state, they will not be able to earn delegates in that state (other than superdelegates). I have only included three candidates, because the remaining candidates have less than 15% of the vote on average. (2) Given that, in the main article, the complete poll data is shown below each of the charts, I do no think the reader will misunderstand or assume that there are only three candidates. (3) If one were to include all the candidates in the chart, the data would be nearly impossible to discern. However, time permitting, I will consider adding an "Other Candidates/Undecided" category.--Robapalooza 13:35, 26 August 2007 (UTC)
I ran a trial chart for the Iowa data (see below) using nearly all candidates in the polling. I think that the results are difficult to read.--Robapalooza 14:08, 26 August 2007 (UTC)


I ran a trial chart for the Iowa data (see below) using the three major candidates in the polling (Edwards, Clinton and Obama) and a fourth category including all other candidates and undecided voters. Now, the total = 100%. I think the results are much easier to read. What do others think about this subject? Keep just the 3 major candidates? Graph all candidates? Graph the three major candidates and add an Other/Undecided category?--Robapalooza 14:20, 26 August 2007 (UTC)


Side-by-Side Comparison of Old Iowa and Proposed Iowa

[[2]]
Thoughts, suggestions, additions, deletions or corrections?--Robapalooza 17:55, 27 August 2007 (UTC)

I think the other/no opinion is the better option. I do, however, still think that a linear regression is of only limited value as these things oscillate based on news, forums, and other factors. For example, a candidate with high name recognition will often show a long term decline vs. a low name recognition. A candiadte who jumps in late (such as Thompson) will show a steep (but perhaps meaningless) increase. I do appreciate the work you're doing on these and find them interesting, I just question the value of the statistical metric you've chosen. --Aranae 21:33, 27 August 2007 (UTC)
I agree that the linear regression may be misleading with a late-arriving candidate such as Fred Thompson, but I think that the linear regressions are good for candidates such as McCain, to show an overall trend in the data. Removing the linear regression would, I think, make the data difficult to interpret or understand. Would, say, a moving average be more interesting/useful? Time permitting, I'll do a sample Iowa chart with a moving average for the sake of comparison.--Robapalooza 18:05, 28 August 2007 (UTC)
A moving average is definitely more informative, but I suspect it would be a lot more work. Isn't that why you dropped it in the first place? I think the regression is definitely informative for candidates like McCain and Romney (steady decline and rise respectively), but would miss the subtleties of Obama vs. Clinton (fast climb and leveling off vs. leveled off and then climb). What about fitting a polynomial? For example, all data (in the general) for Obama will fit a rapidly increasing line. Plot January 15th into the equation of that line and he'd be predicted to poll at about 30+%. Plot a quadratic (2nd order) and he's predicted to be at about 24% (about the same as now). Considering his general election numbers have been crudely flat since about March/June, the latter is probably a more realistic prediction based on the data at hand. Now that I look at it again, as long as the "before they were running" data are excluded (such as Obama's 2005 polls or all but the most recent polls for Thommpson), the long term trend lines actually seem in line somewhat with the short term trends. Based on short term trends, the long term redictions seem halfway reasonable. I don't personally think they will hold up as quick shifts are apt to happen once the general electorate starts tuning in in November through early 2008, but they don't really seem that bad to me. For now, I think that moving averages are more meaningful but the graphs as is are fine if you don't want to do all that work. --Aranae 23:21, 28 August 2007 (UTC)
I saw simlar graphs for Fred Thompson, and I will say the same thing here. Use of regression lines on this data without context of upper and lower confidence limits along with constricted y-axis gives the wrong impression to the reader. Furthermore, regression lines shouldn't really be used with this kind of data anyway. The best way to describe this data would be to use the Stock Ticker Graph where the +/- margin of error related to the poll would be used as the high and low mark. Any kind of regression, linear or polynomial or whatever is likely to be used by the reader to predict out into the future, and these statistics are not to be use for prediction purposes. Arzel 13:36, 30 August 2007 (UTC)

state-by-state polling data on candidate pages

The same set of charts providing state-by-state polling data has recently been added to the articles about several presidential campaigns.

I propose that the detailed state-by-state comparisons remain here at the main article, Opinion polling for the Democratic Party (United States) presidential primaries, 2008. Including the same information in each candidate's article is duplicative and makes the articles too long. It also makes maintenence and updating much more difficult: instead of updating one page, many pages must be updated whenever a change is made. Each candidate's article would have a summary of the state-by-state data and a link to the main article. Thoughts? -Fagles 20:45, 28 August 2007 (UTC)

On a purely maintenance front, the current set up is rather unmanageable. However, that being said, the state by state comparison could be turned into a template (or similar) and transcluded into each candidates campaign article. That way only one page has to be updated and any change made to the "state by state" page would automagically cascade to the campaign article pages.. That being said, the state by state should really only be in the Opinion polling article, not every candidates campaign article. Even if the state by state comparisons were turned into a template, it make significant portions of the candidate's campaign articles identical and the state by state comparison is something that is better carried out in a centralized location of the opinion polling article. --Bobblehead (rants) 21:03, 28 August 2007 (UTC)
I strongly feel the state-by-state comparisons are best left on a centralized page and not be included on individual candidate pages. It's redundant to include the information on multiple pages and clutters up the candidate pages. A link at the bottom to opinion polling would be sufficient. - PoliticalJunkie 18:56, 2 September 2007 (UTC)
I think the advantage of keeping the charts in each candidate's presidential campaign page is that it shows a side-by-side comparison of that particular candidate's support such that a reader can quickly ascertain the status of the particular candidate in one article. That is, it's efficient. Whereas, the present page includes data for all states and a reader cannot quickly assimilate the data. --Robapalooza 16:44, 9 September 2007 (UTC)

Winner Takes All?

I think the "Winner Takes All" section should be moved below the "15% Rule" section, because otherwise I think it could be misleading. The 15% rule is what matters in the primaries. Davemcarlson 10:49, 8 September 2007 (UTC)

Unfortunately, the rules for selecting delegates are not simple and uniform. In the Democratic party, while most states are "proportional," some are caucuses, which may result in a winner-takes-all result. See, for example, [3] —Preceding unsigned comment added by Robapalooza (talk • contribs) 16:51, 9 September 2007 (UTC)

At least the "winner" header in each field should be changed. Perhaps "Delegates Selected" with the number of delegates each canditate is awarded. —Preceding unsigned comment added by 159.233.99.136 (talk) 18:44, 7 December 2007 (UTC)

I think we should delete this section altogether. Of what use is it to tally up what would happen if all states were winner-take-all, when we know that's not the case? --Allen (talk) 18:50, 15 January 2008 (UTC)
I removed the section. --Allen (talk) 17:48, 17 January 2008 (UTC)

To address the concerns some of you have over the use of linear regressions

To address the concerns some of you have over the use of linear regressions, I've added links to sites that summarize polling data in a similar manner, but use different methods for plotting averages. I will keep maintaining the charts on this site, since it includes more data than the images in the attached links. Thoughts, questions, suggestions, etc. are welcome. This example is from the Republican California primary section, but similar links appear on this page. [4] --[User:Robapalooza|Robapalooza]] 00:46, 10 September 2007 (UTC)

Date format

Not a big deal, but it's been bugging me, so I thought I'd ask. In the present page, all dates are provided with a link to a wikipage about that date and year, e.g. October 14, 2007, appears as October 14, 2007. The wiki page Republican counterpart of the present page, i.e. Opinion polling for the Republican Party (United States) presidential primaries, 2008 has not been including such date formatting. Personally, I think the information is not particularly helpful or useful, as the dates of the primaries are in the future, and the dates of the polls are not significant in and of itself. I am inclined to delete the formatting for the dates (i.e. just the double-bracketing, not the dates themselves) in the present article (except for maybe the dates of the primaries), but wanted everyone's input, if possible. Three possible options: 1. Keep the date formatting here and add similar date formatting to GOP counterpart page. 2. Get rid of all date formatting. 3. Keep date formatting for primary dates, but eliminate date formatting for poll dates. --Robapalooza 22:03, 14 October 2007 (UTC)

I think all of the date brackets should be removed. They serve no real purpose. Michael 134.84.96.142 23:35, 5 November 2007 (UTC)
The date brackets allow users to customize the appearance of dates (so US users see January 1, 2001 but UK users see 1 January 2001, etc.) WP:MOS says full dates should always be linked, for this reason. 151.199.53.48 19:02, 11 November 2007 (UTC)

Latest Nevada Poll

The Nevada poll that linked to the Elko paper was taken out as it was highly unscientific. It was simply a web poll in which a person or group could flood the website with votes. While I wish Obama was winning Nevada with 36% as the poll suggests, its not a random sampling of the population in Nevada. 70.247.188.137 18:25, 29 October 2007 (UTC)

New poll out for January: http://www.rgj.com/blogs/inside-nevada-politics/2008/01/new-poll-democratic-race-in-nevada-dead.html Jayavarman1 (talk) 19:46, 14 January 2008 (UTC)

Updated Iowa Graph Image

I wasn't sure how to update the image, but, if someone wants to do that for me, I compiled a new graph for the Iowa results, copying as closely as I could the style of the previous one. I only used the results of the past three months. Here is a link to a .png. http://download.yousendit.com/6D836BC6274F3FA9 —Preceding unsigned comment added by 159.91.112.4 (talk) 20:45, 6 December 2007 (UTC)

Updated New Hampshire Graph Image

I updated the New Hampshire Graph, again from September to December. There's a link to the older one on the image page. —Preceding unsigned comment added by Luckydevil713 (talk • contribs) 20:16, 9 December 2007 (UTC)


Graphics Consistency?

Should the color scheme of the map and the graph be the same -- green and red seem to be switched...Dawginroswell (talk) 16:49, 12 December 2007 (UTC)

Updated the map and pie chart to be more consistent with the Iowa and New Hampshire graphs. Also added a short description to the All Data pie chart, explaining what data it is actually representing.Luckydevil713 (talk) 04:22, 13 December 2007 (UTC)


Thanks!Dawginroswell (talk) 13:11, 13 December 2007 (UTC)

Estimating the proportional distribution of delegates

Until now the delegates were distributed using the following scheme:

  • If the candidate received 15% or more, assign the candidate his (percentage * total number of state delegates (pledged + unpledged)) delegates, rounded to the nearest integer.
  • If the candidate received less than 15%:
    • If when ignoring the undecided respondents (not Other/Undecided), the candidate receives 15% or more, assign him delegates according to his original percentage.
    • else don't assign the candidate delegates.
  • Other/Undecided receives whatever remains.

This is the scheme I have used when updating the numbers (this is simply the scheme that was in place when I started editing the article).

Today an anon user has updated the distribution of candidates in some of the states using what appears to be the following scheme:

  • Distribute the pledged candidates between the candidates that received 15% or more.
  • Assign the superdelegates to Other/Undecided.

This scheme, IMO, has the following advantages/disadvantages:

  • Advantages:
  • More accurately represents the results one would expect in the primaries (based on current polls results).
  • Disadvantages:
  • Not by much, as (if I understand the delegation selection process) most (if not all) states don't pledge the delegates according to (or only according to) the state-wide voting results, but rather according to district level results (for which no polling data is available).
  • The results of this delegate distribution scheme (in the article) are less ambiguous than what the polling data suggests, as Other/Undecided polling respondents are completely ignored.

Based on these disadvantages, I believe that the first scheme is better. Please discuss. Rami R 17:44, 18 December 2007 (UTC)

So, does anyone actually care how the numbers are calculated, or can I just replace the results with random numbers? Rami R 11:12, 22 December 2007 (UTC)
Personally, I don't think it makes any difference. We all know that after Iowa, New Hampshire, and South Carolina, the field will be reduced to 2 or 3 candidates. And well before any delegates go to Denver, we will probably know who the candidate will be. Do the South Dakota Delegates make any difference?--Appraiser (talk) 16:33, 23 December 2007 (UTC)
I think the way you're doing it is fine. And I disagree with Appraiser that no one cares... because I do, haha. Luckydevil713 (talk) 21:51, 24 December 2007 (UTC)

I have recalculated the proportionate distribution of the delegates, and added to the article an explanation of the calculation. Rami R 13:54, 27 December 2007 (UTC)

Kudos

The effort gone into this page and keeping it current is truly commendable. This is the best place to keep up on the latest poll data and beats the hell out of the big "news" outlets. Thanks for your hard work everyone. 193.129.64.154 (talk) 10:48, 21 December 2007 (UTC)

Margin of error

This is truly a fantastic article, although I would love to see MoE included for each poll. I also believe it makes sense stylistically to bold all candidates who are within the margin of error for first place. Polls are not exact, and it's important to get that point across. 151.204.231.247 (talk) 21:37, 26 December 2007 (UTC)

I have added MoE and sample sizes to the recent polls (where available). Perhaps I'll add for other polls later. As for bolding all condidates within the MoE: I don't think it's critical, and readers may find it confusing ("Why is Clinton/Obama/Edwards bolded when he's/she's not leading!?" and then "fix!!!!" (good case) or "POV!!!!" (bad case)). Rami R 19:05, 27 December 2007 (UTC)
Perhaps underline or italicize those within the MoE? --Aranae (talk) 20:17, 27 December 2007 (UTC)
  • I have tried a solution to make it more obvious which polls are within the MOE, as well as a new system to determine how the map should be shaded - see table for Alabama. Are people happy with this? If so I will turn it into a template and roll it out for the whole article. Umeeksk (talk) 01:09, 5 February 2008 (UTC)

ARG Polls

I Look at polls constantly. the American research group polls are simply imposable unless they are slanted in favor of Hillary Clinton. I suggest these polls be removed or labeled as such. —Preceding unsigned comment added by 69.227.179.208 (talk) 22:04, 26 December 2007 (UTC)

These numbers are, in my opinion, crazy compared to the others. Nevertheless, without solid support from citable sources we are in no position to say which polls we like or dislike. --Aranae (talk) 22:20, 26 December 2007 (UTC)
Real Clear Politics(the site that I have been using to update all the Iowa and NH polls) has just stopped including ARG results in their 5 poll average. Also, I'm not sure if this is a reputable source, but this blog did a hefty job of research about them. I think everyone was a little skeptical because of those ridiculous Iowa numbers that had hillary as high as 34%.Luckydevil713 (talk) 20:55, 28 December 2007 (UTC)
I agree. I took a look at all the ARG polls and in every one of them they favour Clinton (or, rather, against Obama) compared to other polls. This does not look like an unintentional bias to me.Herunar (talk) 03:36, 8 January 2008 (UTC)
After reading through the research [5] and doing some more validation myself, I've come to the conclusion that the American Research Group estimates cannot be trusted and should not be included in the article. They formed a significant part of the article, but then, I believe in quality over quantity. They simply form too much of a bias - some of their polls contradict other polls by over 20 percent. In an article that focuses more on numbers than on notability, I believe our contributors have the right to remove obviously inaccurate sources, though notable (not much now - RealClear has stopped using them). Herunar (talk) 04:01, 8 January 2008 (UTC)
I'm not sure how the article works, but is it necessary for some of the data to be corrected now that the significantly biased ARG estimates are removed? Or do these data come from other sites? Herunar (talk) 04:03, 8 January 2008 (UTC)
While I agree that ARG polls don't seem quite right, our collective opinions regarding ARG are not a sufficient reason to remove their polls from this page. Only if a reliable source is provided that explicitly claims that ARG polls are unreliable, then can we discuss the option of removing the polls.
Also, while you're arguing to remove the ARG polls from this article, the wikipedia article on ARG remains clean of any accusations regarding their credibility. Rami R 10:41, 8 January 2008 (UTC)
It is not our collective opinion. It is the fact that the CNN is now questioning the polls' reliability, that RealClear has already removed them from their poll estimates. A reliable source disputing another source is not needed when the source itself has no way to prove its reliability. Herunar (talk) 04:59, 9 January 2008 (UTC)
CNN is questioning ARG polls? Really? Where? When? Did RCP mention why they removed ARG from their poll estimates? Is the reason of the removal not just speculation? As for not needing a reliable source to dispute another's reliability: from WP:V#Questionable sources: "Questionable sources are those with a poor reputation for fact-checking". Not "without a good reputation". A reliable source is needed to establish a poor reputation. Rami R 15:54, 9 January 2008 (UTC)
No, I am not clear as to whether RCP removed them from their poll estimates. But one thing is certain - they do not want to use them, and thus either their notability and reliability is questioned. Since RCP made the ARG sources notable, it can easily be concluded that the matter is about reliability. This is not speculation. This is clear common sense.
I never said ARG was "without a good reputation." Facts: they have a poor reputation. And if you bothered to look at the blog quoted above, this would be apparent. The blog itself could not be called a reliable source, but since it made its assumptions based on logic of its collected data, I can foresee no dispute. The facts it quoted are solid - a quotation from [6]: "The New Hampshire-based American Research Group's tracking poll ended up buried deepest in the snow bank: They had Bush winning by two the day before the primary, merely 20 points off the mark." Another quote from the same article: "It was the second debacle for ARG in as many New Hampshire Republican primaries." This seems to me a straight accusation of bad reliability. And we know ARG had tracked New Hampshire polls by, at most, 5 times up to now. 2008 would count as another incident. If they messed up 3 out of 5 by a wide margin, I don't really see any reliability in ARG. The ARG themselves offered this very convincing explanation: "We did not have a polling problem, we just ran out of time." [7] Nothing else need to be said. Herunar (talk) 04:58, 10 January 2008 (UTC)
Some points regarding your comment:
  • Commonsense is just the most reasonable speculation. Another reasonable speculation is that ARG polls did not support RCP's preferred candidate.
  • I actually had read the blog and the Washington post article. The Washington post article is about polls reliability in general, and as an example it cites ARG's 2004 and 1996 New Hampshire polls. It also cites as examples "Voter Research and Survey", Quinnipiac and Gallup polls. Should we remove their polls as well?
  • Every single polling institute got New Hampshire wrong this year. ARG getting this one wrong too doesn't really mean anything.
  • ARG's explanation actually is convincing: At Election day there was major pro-Clinton trend (possibly as a result of Clinton's emotional reaction a day earlier). From the start of this trend until it's peak there wasn't even one whole day.
Rami R 17:51, 10 January 2008 (UTC)

Differentiating between primaries...

On the "Based on Winner Takes All" section, I added a new 'asterik' that would be a way to distinguish between primaries that have occured, and those that haven't. I would also suggest, in addition to removing the '¤' that the editor move that state whose primary just occured to the top of the column, regardless of alphabetical order. Or perhaps there is a clearner way?Luckydevil713 (talk) 09:21, 3 January 2008 (UTC)

This article is supposedly about opinion polls, so I'm not sure it is appropriate to summarize the actual results (a new page for the results was recently created). Perhaps the "current leaders" section should be renamed "last polls leaders" or something of the sort? Rami R 11:08, 3 January 2008 (UTC)
I think the results are interesting on this page, because it shows how much of a disparity there can be between the polls and "reality." Case in point: Obama in Iowa.--Robapalooza (talk) 21:46, 5 January 2008 (UTC)
It seems to me that the just polls show this (the last 2 polls, conducted around the same time, show extremely different results). Not that it really matters... Rami R 23:26, 5 January 2008 (UTC)

Montana

December poll found for Montana: http://billingsgazette.net/articles/2008/01/02/news/state/16-polls.txt

Clinton - 29% Edwards - 19% Obama - 17% Jayavarman1 (talk) 03:51, 7 January 2008 (UTC)

Added-Luckydevil713 (talk) 19:00, 10 January 2008 (UTC)

I know it was nominated before but...

What is the point? These polls are very old and outdated. And even if we do put the new ones up, it would make the article gigantic. Can someone tell me why this article exists? The information is now useless besides as a footnote in history. It would have to be updated every month practically to be of use. Contralya (talk) 14:28, 9 January 2008 (UTC)

The article is updated as new polls come in. The information is very useful for guaging candidate support in each state, as well as statistically following a candidate's campaign history in correlation to the numbers. For example, I have used the numbers to design scenarios for an election game modeling the election at different points in time. Relax. Jayavarman1 (talk) 09:47, 12 January 2008 (UTC)
Personally, I find this article very useful. I don't know if it's just me, but it's a great collection of up-to-date information that is not so easy to find in other websites. Herunar (talk) 14:00, 16 January 2008 (UTC)

Should the recount be mentioned?

Since New Hampshire is holding a recount, should there be a note in the New Hampshire section that Clinton is no longer the definite winner? --Crushti (talk) 19:05, 12 January 2008 (UTC)

Linking dates

I think some articles shouldn't really have all dates linked. In the case of other entries, the date, whether it's someone's birth or the release of an album, is relevant, but on this page, we're likely at this stage to have all dates in 2008 linked at some point or other until the end of all primaries. William Quill (talk) 15:44, 18 January 2008 (UTC)

The links are there to enable autoformatting: see WP:DATE#Autoformatting and linking. As such, they should be retained. Wdfarmer (talk) 23:01, 20 January 2008 (UTC)
I had restored the autoformatting for Alabama, but in the process discovered that autoformatting seems to require that the year also be linked in each date. Is this correct? If so, about 95% of the dates in this article will have to be fixed. Is autoformatting too high a standard to apply to the article in this case? I've decided to remove the years except where needed for clarity, although this doesn't work for users who set their Date preferences to YYYY-MM-DD. What't the right solution? Wdfarmer (talk) 23:41, 20 January 2008 (UTC)

Okay, I've re-read WP:DATE, and as an example I have reformatted the dates in Opinion polling for the Democratic Party (United States) presidential primaries, 2008#Alabama to conform to these rules I found there:
In WP:DATE#Dates:

  • Date ranges are preferably given with minimal repetition (5–7 January 1979; September 21–29, 2002), using an unspaced en dash. If the autoformatting function is used, the opening and closing dates of the range must be given in full (see Autoformatting and linking) and be separated by a spaced en dash.
  • Yearless dates (5 March, March 5) are inappropriate unless the year is obvious from the context. There is no such ambiguity with recurring events, such as "January 1 is New Year's Day".

In WP:DATE#Autoformatting and linking:

  • Full dates, and days and months, are normally autoformatted by inserting double square-brackets, as for linking.
  • Do not autoformat dates that are:
    • in date ranges within the same calendar month.

Is the result an improvement for this article? Or should we just not try to be WP:MOS compliant in this case? Wdfarmer (talk) 02:34, 21 January 2008 (UTC)

Still out of date

MOST of the polls are from mid last year, back when Hillary was "inevitable". This page is NOT useful any more because now Obama is about the same as Hillary nationally. Someone looking at these charts would think that Clinton is the un-deputed victor, but it is really a two way race with Edwards just behind them. I mean come on, almost every single poll has her winning on this page, and that does NOT reflect the current reality. This page is currently HISTORICAL. Contralya (talk) 17:37, 19 January 2008 (UTC)

How about deleting or archiving all polls before the beginning of this year, or at least not using them to calculate leaders. William Quill (talk) 23:56, 20 January 2008 (UTC)
I am saying that the polls are very much out of date. This article includes less than 10% of the polling from 2007, so it isn't a good historical page either! This page is useless as it is now. It is NOT I repeat NOT kept up to date. It isn't even just a month behind, it is half-a-year behind. Contralya (talk) 02:35, 22 January 2008 (UTC)
It's not useless. Drop it. Polls from 2007 take a small percentage simply because there aren't so many polls in 2007. If you believe you could find more, which is possible, go on. 11:10, 22 January 2008 (UTC)Herunar (talk)
What use is there? The numbers are SO different now that a number of states have voted. People started actually caring about the election in January. It seems like there have been 50 national polls since then, and more for particular states. Someone would look at the current numbers and think that the race is practically already won by Clinton. And the fact is that even if this was kept reasonably up to date, it would take a lot of work to make it valid, as a week is a long time for primaries. GO ahead, tell me of a practical use for this information. Contralya (talk) 14:18, 29 January 2008 (UTC)
This article has already been a candidate for deletion and consensus was that this article IS useful. If you feel different, I would direct you to WP:GAFD#If_you_disagree_with_the_consensus. Polls are kept up to date as they are made available, and the page is undergoing constant clean up. Maybe you could help clean and make it better. Thanks. -Luckydevil713 (talk) 09:18, 30 January 2008 (UTC)
A practical use of this poll is that I feel it is useful. Now you can bugger off and actually start contributing. Herunar (talk) 08:41, 4 February 2008 (UTC)

Table for 15% threshold

The table showing the delegate count according to polls, the 15% threshold, and superdelegates is a textbook example of WP:SYNTH and so it constitutes original research. If some one has a source which actually takes into account the 15% threshold than by all means cite it. If there is no source for the information it should be removed. While the number crunching is nice and helpful it also is original research unless a source can be provided.--The Devil's Advocate (talk) 03:55, 30 January 2008 (UTC)

All states use the 15% rule in awarding proportional delegates... which is why John Edwards recieved no delegates in Nevada, since he only received 4% of the vote. As for a citation... See the year 1988 where it talks about the threshold, and how it was lowered from 20% to 15%. -Luckydevil713 (talk) 08:40, 30 January 2008 (UTC)
I'm not talking about the threshold, but the data in the table determining delegates based off the threshold. That is not specifically cited, meaning it is snythesis.--The Devil's Advocate (talk) 18:39, 30 January 2008 (UTC)
Yeah, the table also seems to take superdelegates into consideration when applying the polls, which isn't how superdelegates support candidates. I think the whole table is misleading for the purposes of showing delegate estimations, since it treats polls as if they've already assigned the delegates. At most it should show the actual delegate figures for states that already held primaries, no estimates based on polls unless there's a source for it. Leebo T/C 18:00, 31 January 2008 (UTC)
  • I've taken the table out. I think it's borderline original research and it should at least only make use of recent polls and leave out the (unelected) superdelegates. --78.22.223.177 (talk) 22:38, 31 January 2008 (UTC)
    • So is anyone actually in favor of keeping the table? --Sloane (talk) 17:59, 1 February 2008 (UTC)
      • I'm for leaving a table that only shows actual results. The fluctuating and often out-of-date poll-delegate estimates are inaccurate and misleading. Leebo T/C 18:32, 1 February 2008 (UTC)
      • I see your point about estimates that are based on older polling data, but I have been updating estimates based on new polls as they come in. I would take suggestions on a cutoff date for "older" polls not being included, but I think it is a useful table otherwise, and does not make any biased assertions. —Preceding unsigned comment added by Jl3307 (talk • contribs) 19:51, 1 February 2008 (UTC)
        • Can you explain why the estimates include superdelegates? Superdelegates aren't assigned by voting, so polls should have no effect on them. For instance, Connecticut has only 48 pledged delegates, but the 44/48/8 poll estimates seem to include all 61 total delegates. Leebo T/C 20:14, 1 February 2008 (UTC)
          • How are these delegates estimates calculated anyways? By popular majority? That's not how they work, just look at the Nevada primary, where Obama won the majority of delegates but lost the popular vote. I think we should at least take out the superdelegates, as the table is way too confusing at the moment. --Sloane (talk) 20:30, 1 February 2008 (UTC)
            • The people editing the table, such as Jl3307, are estimating the totals incorrectly, leading to the misleading figures. If one is to update the estimates, they should at least be done accurately and not include superdelegates. But this goes back to the original point of the estimations being systhesis. They should be removed for that reason, as unsourced data. Leebo T/C 20:33, 1 February 2008 (UTC)

color="#2A8E82">C]] 20:33, 1 February 2008 (UTC)

            • As far as the way Nevada worked out, i'd love to see polls based on every congressional district, which is how they assigned their delegates. A few states work things that way, but most don't. There is a clear link above the table to the superdelegate "polling" conducted by CNN, which is updated often. Those superdelgates that don't have a preference for a candidate obviously have not been included in the totals. I will add an additional link for superdelegates. Jl3307 (talk) 20:44, 1 February 2008 (UTC)
              • The fact of the matter is that if you are doing math to figure out these numbers, and you don't have a source for the resulting estimates, you are doing original research. This is more than adding and subtracting, as such simple math is allowed. Leebo T/C 20:47, 1 February 2008 (UTC)

If anyone is still interested in seeing a table with delegate estimations according to polls, I compiled a couple here: --Sloane (talk) 22:37, 3 February 2008 (UTC)

As well, I am still tallying my estimates on my talk page. --Jl3307 (talk) 04:51, 4 February 2008 (UTC)

Why is this POV?

Why is the neutrality of my update of the map disputed? You addad the nonpov template, but you have not said why. Infoporfin (talk) 00:40, 5 February 2008 (UTC)

Thanks, I'll look up there. I thought this was cronological. Infoporfin (talk) 00:52, 5 February 2008 (UTC)

There are a few problems w/ map Delaware is clearly in the Clinton column as is Oklahoma based on polls and Mass should be as well based on the fact that all but one poll show her ahead including the most recent. Also Missouri is just as much a toss up as New Jersey probably even more so based on polling. Tulsaschoolboard (talk) 00:58, 5 February 2008 (UTC)

  • Hopefully I fixed all these mistakes and omissions. I made stupid mistakes for Missouri and Deleware, but the others were a result of looking at a version of the page from earlier today. I explained this more further up on this page in the Map section. Infoporfin (talk) 01:47, 5 February 2008 (UTC)


  • Also for Nevada the way its worded and then shown it suggests that Obama will win the state at the State Convention because it has the yellow dots instead of purple dots. (Tjliles2007 (talk) 01:14, 5 February 2008 (UTC))
  • My understanding is that he will win the state convention. That is what multiple television channels have been saying and that is what is says in this wikipedia article about the 2008 Nevada caucus. Infoporfin (talk) 01:47, 5 February 2008 (UTC)
  • Clinton has nearly 600 more state delegates then Obama, she will win the state convention. Obama may end up with 1 more delegate to the National Convention then Clinton, but that doesn't mean he won Nevada. (Tjliles2007 (talk) 02:50, 5 February 2008 (UTC))
  • The 600 delegate margin in favor of Clinton that you describe is for county convention delegates. The county conventions will elect delegates to the Nevada State Democratic Convention. The state convention will elect the delegates that will represent Nevada at the Democratic national convention in Denver. Because of the way Clinton's county delegates are distributed throughout the state, will probably win one more national delegate than her at the state convention. Infoporfin (talk) 07:06, 5 February 2008 (UTC)
  • She won the popular vote and my version of the map shows as much. Obama will likely receive one more delegate than her from the Nevada state convention. That is why he is shown as the projected winner of the state-wide caucus. Infoporfin (talk) 07:09, 5 February 2008 (UTC)
  • This isn't very hard to solve, just change the description to "most national convention delegates won" or something. --Sloane (talk) 03:20, 5 February 2008 (UTC)
  • Here we go again... two Wisconsin polls from different orginizations within 4 days each showing a different leader and of course the map is immediatly changed to show it as an Obama state. We really need to lose the bias on this page. Theres conflicting polls either leave it neutral or give it its own color. (Tjliles2007 (talk) 19:40, 12 February 2008 (UTC))

Shouldn't states with conflicting or inconclusive poll numbers be colored as such?

File:08 Democratic Primary Polling.png

Hi folks, I'd suggest that only those states where the polling shows a clear leader be marked as such on the map. States where very recent polls disagree on a leader, or where the polls show candidates within the margin of error, should be colored "inconclusive"... such as gray on the map at right.

Remember, the purpose of the map is to represent the poll data, and the poll data for those states does not indicate a clear frontrunner.

(BTW the lower map shows states by their electoral size.)

Thoughts?

Szu (talk) 02:21, 5 February 2008 (UTC)

Article split

Shall we split the article up between states that have already held their primaries and states that have yet to come? Or are we going for number one?--Sloane (talk) 23:28, 6 February 2008 (UTC)

  • Uhh... the states beginning with letters A-L appear to have been removed. Why? Many of these have not had their primaries yet. I came here in search of Louisiana. Xezlec (talk) 19:52, 9 February 2008 (UTC)
  • Louisiana never had any polling numbers that were posted thus was not listed in the article. Polling from January and Super Tuesday were placed in their own articles to cut down on the size of this article. (Tjliles2007 (talk) 05:37, 10 February 2008 (UTC))

Florida and Michigan Delegates

I added into the chart the Michigan and Florida delegates with a strikethrough to show the value should the DNC allow the delegations to sit. I also at the very bottom added a alternate total including Michigan and Florida.

I think this is important to show because there is a very good possibility that the two delegations will end up seated at the convention. (Tjliles2007 (talk) 23:27, 7 February 2008 (UTC))

At the moment, though, they remain irrelevant. Keeping the table straightforward is difficult enough given the status of these two states. Adding data that might be relevant someday doesn't mean that the data, as it stands now, isn't counterfactual and therefore a detriment to the integrity of the article. Leaving it off for now would be best.
—GrittyLobo441 (talk) 02:36, 8 February 2008 (UTC)

I removed the total at the bottom including Michigan and Florida. It is one thing to include them in the table for reference but it is quite another to allow the chart to culminate in a total to include these primaries when neither candidate campaigned in the states and Barack Obama did not have his name on the ballot in Michigan. I suggest that these 2 states are left in the table but the total including these 2 primaries should be excluded from the end of the table as misleading. —Preceding unsigned comment added by 144.26.117.1 (talk) 17:47, 20 May 2008 (UTC)

Pie map innaccurate?

The pie chart map at the top of the article seems to be inaccurate WRT Louisiana. CNN reports Obama won 33 delegates with 57% of the vote to Clinton's 23 delegates with 36% of the vote there.[8] — Swpbtalk.edits 16:52, 10 February 2008 (UTC)

I used numbers from CBS's election results page. They seem to have been update more recently for most if not all states. They show Obama with 34 Louisiana delegates and Clinton with 22. The Sum of that is 56, the total number of pledged delegates for LA. Infoporfin (talk) 20:48, 10 February 2008 (UTC)

Ok - I realized I saw the chart wrong. When I glanced at it, it looked like the bottom third was gray. You might want to make the colors a little bolder. — Swpbtalk.edits 12:40, 11 February 2008 (UTC)

Display problem with proportional chart

The proportional chart of polls/results overlaps the contents table on smaller monitors so that you can't use it - can this be fixed? Umeeksk (talk) 21:52, 12 February 2008 (UTC)

Polling

I really have to wonder about the legitimacy about some of these polls.Like Wisconsin, where so recently Hillary Clinton had quit a lead, and suddenly, a few days later he's winning largely?.Maybe its just how eacch was conducted, but variations like that shows how unrealiable they are, especially for an encyclopedia. Rodrigue (talk) 18:23, 13 February 2008 (UTC)

  • That's the nature of polls, I'm afraid. All we can do is give the methodology, the margin of error, and the polling organisation, and then people can form their own interpretations. Anything else would be decidedly POV. In the case of Wisconsin, by the way, the last three polls have shown Obama ahead (although I think the Public Policy Polling result is an outlier).Umeeksk (talk) 00:44, 16 February 2008 (UTC)
  • Rodrigue, you say "variations like that shows how unrealiable they are, especially for an encyclopedia". This seems to imply that Wikipedia should not contain polling results because of their unreliability. I disagree with you; all this article is doing is presenting the polls, it's not suggesting any more weight on them, and the opinion polling article discusses all of this in the event that someone wants to find out for themselves. It would indeed be inappropriate for Wikipedia to state, based on polling, that one candidate would beat another in a particular state. It is not inappropriate to simply state that a poll was conducted and show the findings of the poll. The article implies nothing by showing the results. Leebo T/C 18:31, 19 February 2008 (UTC)

LV = only Landline Loters?

Some polls say "LV" behind them. I have no idea what this means, but since some polls say "(includes cellphone voters)," does "LV" mean "only landline voters"? —Preceding unsigned comment added by SteveSims (talk • contribs) 06:45, 19 February 2008 (UTC)

If you go to the links for each of the polls with the "LV" you'll see that they didn't contain any info about "landline voters." I think it just means "likely voters." Leebo T/C 18:36, 19 February 2008 (UTC)

Polling of completed primaries

Why are states that had their primaries a while ago being removed? I think we should be keeping all polling data for historical reasons. I think we should either keep the data here or move it to the relevant state primary articles. Thoughts? Andareed (talk) 12:08, 23 February 2008 (UTC)

Why are the February 5 results on a different page than the rest of the February results? It would make more sense to me to have all the completed elections on one page. I was looking at them to compare accuracy of polls to results. Senatorpjt (talk) —Preceding comment was added at 02:17, 4 March 2008 (UTC)

Decision Analyst polls

Theyre clearly anomalous results (probably more than 2 SD away from the mean)... should they really be in there? It's seriously misleading to put in results showing Obama having shot up 14% in Ohio and 12% in Texas in the space of 24 hours 79.74.113.97 (talk) 18:39, 24 February 2008 (UTC)

I agree they're probably outlier polls, but they do exist. If the polling method was fair, you can't cherry pick which poll data to keep just because the results seem to be contradict a few other polls. --WTStoffs (talk) 19:23, 24 February 2008 (UTC)

Actually the same polling firm seems to have 2 contradictory polls for ohio and texas, which seemed to have been removed, I don't think its cherry picking as I think there is some serious outlier here. Rodrigue (talk) 19:29, 24 February 2008 (UTC)

I don't think it makes sense to just arbitrarily decide they are way off. Is there evidence that they are off, or that Decision Analyst is a poor polling organization, or something? If not, then why not just wait until more polls come along, which either will or won't correc them.

fair enough, maybe Obama is just surging that much, though I don't know, have to wait for more polls. Rodrigue (talk) 19:57, 24 February 2008 (UTC)

As I understand it, Obama is nevertheless the leader of the latest poll in Texas, and the state's color should be changed accordingly. Ohio, since polls taken within the same timeframe conflict, should be deisgnated as "no clear leader." Arbitrarily discounting polls that are inconsistent with previous polls, yet quite consistent with prevailing trends is both dangerous and unwise. —Preceding unsigned comment added by 128.135.226.70 (talk) 00:11, 25 February 2008 (UTC)

"dangerous and unwise" seems a bit strong =). These are the sorts of words we could use to describe, for instance, John McCain mockingly singing "Bomb bomb bomb bomb bomb Iran" and then responding to questions about it with "Insensitive to what? The Iranians?!?" —Preceding unsigned comment added by Benjaminady (talk • contribs) 03:00, 25 February 2008 (UTC)

  • There were 5 polls released within a 24 hour period with four showing Clinton ahead and one showing Obama ahead. I think theres some serious questions about the accuracy of the Decision Analyst polls, while I don't think they should be excluded here, I think we shouldnt let these polls significantly influence trend of the races. (Tjliles2007 (talk) 03:07, 25 February 2008 (UTC))
  • I would also cast these results in doubt. Go to the Decision Analyst website to see the method they used for polling. The polling method involved using another website called American Consumer Opinion® (www.acop.com). American Consumer Opinion pays people for there opinion. It may be just me, but it seems a bit odd to use data from a paid online survey panel. 70.58.158.33 (talk) 08:43, 25 February 2008 (UTC)

I think the main thing with those polls are the noticeable lack of an 'undecided' option, which may legitimize the polls. However, someone really should look into them paying people for their opinion. This would certainly adversely effect accuracy. However, a poll is a poll, and I think it should be posted here.

They're mathematically anomalous compared to the other; it's not cherrypicking. It's extremely misleading. I'm sure there are plenty of polls nationwide, but not all can be included if they are horribly outlying. Note the media hasn't bothered aknowleging them. The race is close; Barack Obama is not 10 points ahead in Texas. 79.74.24.29 (talk) 18:24, 25 February 2008 (UTC)

These polls should not be removed unless someone can find something more legitimate than, "They're not showing Clinton ahead or in a tight race". After all, as concerns the polls in Texas, it wasn't the first showing Obama ahead. We have to remember that Obama has been seriously outperforming the polls. Look at the polls for Wisconsin and the actual results. The undecided seemed to almost all go for Obama and very few voters going for someone other than Clinton and Obama. Just because a poll shows a vastly different result doesn't mean it should be discounted, indeed, it could be the more reliable poll. One thing to remember is margin of error, in Ohio their margin of error could mean 51% Obama and 49% Clinton, a fairly reasonable possiblity. In Texas it would make it a 54% to 46% win for Obama, which syncs up pretty well with a shift of undecideds to Obama in other polls.--The Devil's Advocate (talk) 19:25, 25 February 2008 (UTC)

On another point I'd suggest people look at their polls for the Republicans which are very consistent with other scientific polls except showing a larger lead for McCain, but when one considers the undecideds it's perfectly consistent. If anything these polls simply forced people who would otherwise say undecided to give a decision and their decisions seem to fit in with the votes of previous undecided, that is, going with the person with perceived momentum.--The Devil's Advocate (talk) 19:36, 25 February 2008 (UTC)
  • Three polls have come out since the Decision Analyst poll. Two have Obama slightly ahead, one has Clinton ahead by 1 point. Even if the DA poll is a bit of an outlier, it isn't contradicting other poll results anymore. WTStoffs (talk) 22:14, 25 February 2008 (UTC)

Poll

  • decision analysist poll for ohio was removed, as its 54-46 lead for Obama was really contradicted by the 2 latest polls, as well were quit a shift from trend of previous ones, and they're pollnig methods were suspicious.
    Similiar situation for they're texas poll, likely will also be contradicted by subsequent polls the same way it differs from previous ones. Rodrigue (talk)
  • Im not sure we should delete the poll. People can make their own assumptions about the polls from that company. But for basis of how the maps are colored the polls from that group probably shouldnt be used. (Tjliles2007 (talk) 18:16, 25 February 2008 (UTC))
  • It's kind of ridiculous to remove a poll based on the fact that it doesn't match others. Nothing is "contradicted" simply because polls don't agree or show trends. Focus on describing why their polling methods are bunk if you think that's the case. Leebo T/C 18:17, 25 February 2008 (UTC)
  • Those are compelling reasons to remove it. Not these reasons about averages and outliers and trends. Those should have no impact on the polls presented, because we're only presenting polls not manipulating the data or interpreting them. Leebo T/C 18:41, 25 February 2008 (UTC)
  • Wow, some ppl don't know how to read.Click on the link where the poll is, the page says under methodology:
    Results based on scientific sampling of registered voters from American Consumer Opinion® (www.acop.com), one of the largest online research panels in the world.
    See, its not a random land-line phone survey, its an online survey, therefore unreliable. Rodrigue (talk) 19:55, 25 February 2008 (UTC)
  • There is still nothing saying this was an online survey. They specifically say, "Results based on scientific sampling of registered voters from American Consumer Opinion." This could mean they simply used it to get their sample and then conducted a phone survey. For instance, according to the ACOP site, "Occasionally, these consumers are asked to participate in mail surveys or telephone surveys, but generally the surveys are all conducted over the Internet." How do we know this wasn't one of those occasions? However, even if it was online does this automatically mean it's unreliable? It's not like this would have been some poll plastered on their page for all to answer, according to the site they took a scientific sample of registered voters in Ohio adjusted for the demographics and conducted the survey. Also, on your point about payment, going by their site this is hardly a matter of people being paid as professional opinion givers, more like they get a small reward for their participation. Finally, the site seems to conduct surveys on products, not elections, so there's no indication ACOP's participation is relevant to Decision Analyst carrying out the poll.--The Devil's Advocate (talk) 20:00, 25 February 2008 (UTC)
  • Well, if it was online it would be unlike most/all the other polls on the page, questionable to use one with different methodology.
    wether or not it was in fact online, the fact that both its polls differed highly from others suggests bad methodology. Rodrigue (talk) 20:30, 25 February 2008 (UTC)
  • That's a logically falacious argument. The difference in results doesn't imply anything about the methodology and nothing can be determined from the results. Likewise, a poll being close to previous polls doesn't suggest good methodology. It suggests nothing. Leebo T/C 20:31, 25 February 2008 (UTC)
  • But multiple odd polls from the one company that likely conducted them online?.Whatever, let them stay, I'm sure subsequent polls will show how odd the results are. Rodrigue (talk) 20:39, 25 February 2008 (UTC)
  • The recent ARG poll in Texas also shows a strong lead for Obama and when you consider the fact undecideds seem to be shifting almost entirely to Obama both ARG and the latest Rasmussen would be showing results close to the Decision Analyst poll. It's also not questionable to use it even if it's online if all other things are the same. If the whole list of demographics is adjusted for in the sample and it includes only registered voters in Ohio who are likely to vote in the primary it's no different from any other poll. However, since you have no factual basis for saying this is an online poll it's a moot point.--The Devil's Advocate (talk) 20:43, 25 February 2008 (UTC)
  • I can see this is going nowhere so I'm done, but just that from what the site said it may or may not have been online, but was the only polling firm to suggest so, but yea just moot. Rodrigue (talk) 20:59, 25 February 2008 (UTC)
  • They didn't say anything about how they conducted it. They just said where their sample of voters came from. How they contacted them isn't relevant as long as they were selected in a scientific manner. Leebo T/C 21:07, 25 February 2008 (UTC)
  • I don't understand why Texas is not in Obama's column right now. This "no clear leader" nonsense seems to have little basis in reality. Three of the last four polls place him ahead, each of them by margins outside the statistical margin of error. The only poll where Clinton leads shows a statistically insignificant one percent lead. Please explain. —Preceding unsigned comment added by 128.135.226.192 (talk) 21:51, 25 February 2008 (UTC)
  • Just changed it. Since we have five polls to work from reliably for the present situation it can't be ignored that two of the recent polls show a huge lead and those two are of the most recent, with the only other one more recent showing a one-point lead for Hillary with a large enough undecided vote to sway it either way, most likely to Obama going by past primaries.--The Devil's Advocate (talk) 22:06, 25 February 2008 (UTC)
  • All of these recent Texas polls have the two of them in a statistical tie though. Just because Obama is leading in a poll doesn't mean he actually is, since they all seem to be inside the margin of error. A poll needs to be outside the margin of error in order to say one candidate is "leading." Leebo T/C 15:25, 26 February 2008 (UTC)

"Margin of error", never understood that, don't polls technically have a %100 margin of error by basically asking ppl something they can lie, or not really know and guess, and so not really representing real opinion. Rodrigue (talk) 17:21, 26 February 2008 (UTC)

  • See Opinion polling. It refers to the amount of sampling error that occurs when you don't ask every single likely voter. Larger sample = lower margin of error. And if you're going to assume that people are lying, you shouldn't even care what the polls say in the first place. It also seems a bit odd for you to go on and on about the methodology of certain polls when you aren't aware of how margin of error factors into them. Leebo T/C 17:41, 26 February 2008 (UTC)

Well it may still have been withing the margin of error, but previous polls had clinton at the higher-end, now its reversed, I assume thats why image of US map now says he's winning texas by latest poll, otherwise maybe its wrong. Rodrigue (talk) 18:53, 26 February 2008 (UTC)

  • I'm still not clear on how that map is being created or why poll data is being used that way. I am of the opinion that we shouldn't be assigning frontrunner status in specific states based on our own amateur analyses of poll data. Leebo T/C 18:59, 26 February 2008 (UTC)
Well, one could argue there shouldn't be an article like this at all, however, since such an article does exist a map is only natural as it serves as a nice summary for the information contained in the article.--The Devil's Advocate (talk) 00:22, 27 February 2008 (UTC)

But the interpretation of who's the so-called front-runner seems questionable, many polls are still with-in margin of error, so does that count?. Rodrigue (talk) 01:06, 27 February 2008 (UTC)

The question is who is leading, which is certainly Obama. Only one poll shows a Hillary lead and it's a one-point lead and there's another showing a tie. Reasonably Obama is the leader in the latest polls as his lead in other polls is larger.--The Devil's Advocate (talk) 05:09, 27 February 2008 (UTC)
  • My point is that no matter how large a "lead" one candidate has (1 point, 2 points, 5 points), or how many polls show him or her in the "lead", they are statistically tied if they are inside the margin of error. On your other point, about there being an argument against having an article like this, it would be an exceedingly weak one. This article is for showing poll results, that's it. The map attempts to analyze these results. That's inappropriate, unless the map itself is coming from a reliable source. Leebo T/C 15:46, 27 February 2008 (UTC)
The map isn't based on an analysis, but a simple reading of the numbers. In other words, leaving out the undecided factor and margin of error.--The Devil's Advocate (talk) 04:32, 28 February 2008 (UTC)
  • You can't leave out the margin of error. It means nothing without it. And I'm in discussion with the person who created that map to inform them that it's original research to combine polls (he performs weighting calculations on them) and then present them in the article. Leebo T/C 17:51, 28 February 2008 (UTC)
Except the map was updated by more than one user and the point is to sum up what the polls are saying. Reasonably we shouldn't leave out margin of error, but at the same time applying margin of error is more like original research than just a plain viewing of the polls and listing of who is ahead according to them. Margin of error should only be a factor if significant differences arise. Most of the recent polls show Obama ahead, making it reasonable to put him as the leader. If we factor in margin of error, no one is ahead in the three important states and the map becomes pointless. If anything there should be a disclaimer on the map.--The Devil's Advocate (talk) 00:12, 1 March 2008 (UTC)

The question that needs to be answered for including the poll is whether or not the poll is encyclopedic. If the poll was performed by a major polling firm then it probably is encyclopedic--even if they made a major error. If the poll was some random website poll then it probably isn't encyclopedic. Some people may find having summary graphs of up-to-date average polling data useful so I won't complain about it. I will note that it technically isn't encyclopedic since it borders on original research and the methodology for what polling dates and values are valid is a little fuzzy (and it is of course destined for irrelevancy as the primaries continue and the up-to-date graphs are updated). This is an encyclopedia, not a news site. I say this not to try to push the enthusiasm for updates off to Wikinews, but to note that it isn't crucial to update every day for whichever candidate is the front-runner or leading in a graph or map at some point in time since there is a date stamp on all data. Cheers.--Burzum (talk) 09:33, 28 February 2008 (UTC)

I think your missing the point devil.Margin of error is part of the science of polls, if someone is winning or loosing just by margin of error, it means nothing, they're statistically tied.However many polls show the person ahead within that margin simply means they're tied. Rodrigue (talk) —Preceding comment was added at 00:27, 1 March 2008 (UTC)

Except then the three crucial states are all tied and then it's really just unhelpful and pointless to have a map. I think adding a disclaimer noting the map does not account for margin of error would be sufficient.--The Devil's Advocate (talk) 07:55, 1 March 2008 (UTC)

You were right in the first part of your statement.You can't argue that margin of error shouldn't be used because it would make the map pointless, if thats the case the map should just go, you can't cherry pick fact and methodology to make things interesting.

forget the disclaimer, they'd just be tied, don't mislead by saying one is winning withought margin of error. Rodrigue (talk) 16:23, 1 March 2008 (UTC)

This was, of course, not an issue when Hillary was leading but now it's suddenly an issue because Obama's leading in one of the states? I don't think so. The map provides a nice summation of the polls. We already highlight the leader in each poll and the margin of error isn't as big an issue as some might make it. The polls have proven fairly accurate outside of caucus states the undecideds have been the major swing in each case. I suggest you look at some of the polls for primary states that have already gone. All of them were close, within one point or so usually, to the actual results because the undecideds ended up going to him.--The Devil's Advocate (talk) 19:03, 1 March 2008 (UTC)

whatever, not like I'm the one who even brought this up. Rodrigue (talk) 22:58, 3 March 2008 (UTC)

Delegate Count Correct?

  • If I total the columns for the delegate counts I get Obama at 1183 delegates (1250 with FL & MI) and Clinton at 1025.5 delegates (1203.5 with FL & MI). Is this correct? If so then the current totals are not correct. —Preceding unsigned comment added by Jabberjoe (talk • contribs) 23:13, 25 February 2008 (UTC)
  • I went throught to verify your count and indeed it was off a bit (likely not updated totals when someone added in delegate updates) I also checked the numbers against CBS and added the final Illinois delegate to Clinton and inserted the final totals from the Washington Caucus a few weeks ago. I also added below that table the remaining delegate total from locales that already held its events (17.5) and the total delegates of upcomming events (931). (Tjliles2007 (talk) 23:46, 25 February 2008 (UTC))
  • Great catch. The total including superdelegates is 134 as is stated on the article about the North Carolina primary. According to The Green Pages North Carolina has 115 pledged delegates. I have correctted the number to 115. (Tjliles2007 (talk) 18:28, 27 February 2008 (UTC))

Changing Texas

  • According to the latest opinion poll (February 28, 2008), Hillary Clinton is leading Barack Obama, 47%-43%. However, Texas is still shaded as if Obama were leading there. Could someone please change the color from light purple (Obama) to light yellow (Clinton) so we can keep this page accurate? unsigned comment added by Jellyfish101 (talk • contribs) 16:22, 28 February 2008 (UTC)
  • The image considers the "Latest polls" and this is only the second poll of seven showing Hillary ahead in the past few days. As such it is accurate right now unless more polls start coming out showing her ahead.--The Devil's Advocate (talk) 23:37, 28 February 2008 (UTC)
The stripe was primarily because the latest poll, when the map still went by just the latest poll, showed a massive lead outside the margin of error while others showed a tight race. Presently nothing like that has come up and it's going by latest polls, not latest poll.--The Devil's Advocate (talk) 04:42, 29 February 2008 (UTC)
  • The "latest polls" show Clinton leading in Texas. An average of all the recent polls shows Clinton leading. I agree that Texas should go back to being striped. In fact, it should have stayed striped the whole time since EVERY poll, no matter who was leading, was strictly within the margin of error, meaning both Clinton and Obama were in a virtual tie--as they have been for the past two weeks. --Mbenzdabest (talk) 18:29, 4 March 2008 (UTC)
Honestly, now I think we might as well not do anything. The primaries in Texas and Ohio are today so we have a good chance of knowing who win in just a few hours. However, I don't think it should have been striped the whole time, because arguing margin of error all important states couldn't show a leader and the whole purpose of the map would be lost.--The Devil's Advocate (talk) 18:57, 4 March 2008 (UTC)

poll

Consistency in Full vs Last Names

Sometimes "Gravel" is listed (last name only) and other times "Mike Gravel" is listed. This happens with Richardson, and others as well. I think we should be consistent, and either use the full name or last name only. I'm personally in favour of using last names only (e.g. Obama, Clinton, Gravel, etc...), as we end up with smaller tables and a smaller article. Thoughts? Andareed (talk) 08:15, 3 March 2008 (UTC)

I went ahead and made the change. I'll change the other opinion poll pages if nobody objects or reverts my change after a few days. Andareed (talk) 06:05, 4 March 2008 (UTC)
The other pages are now done. Andareed (talk) 20:49, 6 March 2008 (UTC)

Clinton Lead In Texas Confirmed By Many Polls

Could someone please change the color of Texas to reflect Hillary Clinton's lead? Several polls have shown her surging in Texas and Ohio. I think that this page should be as accurate and up-to-date as possible. Jellyfish101 (talk) 17:26, 4 March 2008 (UTC)

I think by NOW it is obvious that Obama bias is rampant and disgusting on this page and will remain this way until the nominee is chosen. As soon as one poll, no matter how obscure it is, has Obama leading the map QUICKLY changes to Obama. But even if the THREE MOST RECENT polls show Clinton leading in Texas, it remains purple. Obviously, Clinton's win in Texas proves the map should have indeed been changed to reflect this as was repeatedly stated in this discussion. If you want something done, you have to do it yourself. --Mbenzdabest (talk) 08:54, 5 March 2008 (UTC)
The last time the image was updated before the election was on 21:46, 26 February 2008 (UTC). You have nobody to blame but yourself if the image wasn't updated. Be bold!--Burzum (talk) 12:57, 5 March 2008 (UTC)
Your accusations of bias would be more convincing if you weren't a self-declared Hillary supporter. Fact is, those polls were released the very day of the primary so there was not much point in changing it since we could just wait for the actual results.--The Devil's Advocate (talk) 20:33, 5 March 2008 (UTC)
Yes, I am a Hillary supporter. But I wish you would be less rude about my enthusiasm. You'd think I was asking you to delete the page. Jellyfish101 (talk) 20:24, 6 March 2008 (UTC)

How do we color Texas if Obama wins the caucus and Clinton the primary?

  • We can't do it by delegate count since then we'd have to change Nevada from yellow to purple. Making it striped makes it seem like they tied. Any ideas? SteveSims (talk) 05:14, 5 March 2008 (UTC)
  • Seems like popular vote is the best way to go... those involved with the caucus already voted once, and so as a matter of actual people in support or against, the popular vote best describes it. I recommend following the Nevada standard. -- Kallahan (talk) 05:20, 5 March 2008 (UTC) for
  • Texas is 2/3 primary vote for delgates, 1/3 for caucus delegates. If there is indeed a split, perhaps 2/3 of the map striped for Clinton, 1/3 striped for Obama? --WTStoffs (talk) 07:01, 5 March 2008 (UTC)
  • The map shows the POPULAR VOTE. All reputable news sources (including newspapers and television news) follow the same format of showing popular vote for the "winner." So delegates should be ignored. The delegate selection process is completely out of sync with how the state votes anyway. A footnote stating this discrepancy, as it is now, should suffice. --Mbenzdabest (talk) 08:51, 5 March 2008 (UTC)
  • While it's true that caucusers probably voted in the primary, it cannot be said that most primary voters voted in the caucuses. Using the Texas primary vote to override the caucus votes (unreleased) and paint it all one color is misleading. Since Obama had more success in organizing people to attend caucuses, and with the close race overall in Texas, it is just as likely more people at the caucuses voted for Obama than Clinton as it is that the system (like in Nevada for Obama and Alabama for Clinton) favored one candidate disproportionally over the other. Since the vote totals are not released, I think it's fair to show that there were two separate elections one the map and color the state like the Democratic primary results page is, with 1/3 shaded for the caucus and 2/3 for the primary. Krazy19Karl (talk) 19:24, 5 March 2008 (UTC)
  • It's already noted that popular vote does not mean that person is the winner as the person could win more delegates. This already happened with Nevada, but it's colored for Hillary.--The Devil's Advocate (talk) 20:34, 5 March 2008 (UTC)
  • Caucus states reflect the popular vote in the caucus, even if they have a meaningless primary (like Washington State). So, it seems best to not discount the popular vote in the caucus and stripe Texas 2/3rds for Hillary and 1/3 for Obama, if he wins the caucus. SteveSims (talk) 22:30, 6 March 2008 (UTC)

Changing the Average Support in Recent Polls Box

  • This box has not been updated in quite a while. I don't have a clue about how to change it, so who would like to do the honors? Call me a "self-declared Hillary supporter" but her North Carolina margin is way off (the latest poll shows Obama 47-Clinton 43). Could someone make this more accurate? I think that Texas, Ohio, Rhode Island, and Vermont can be removed, since they have occurred already. Jellyfish101 (talk) 20:30, 6 March 2008 (UTC)

Update Mississippi Map; New Recent Poll

  • I added a new poll for Mississippi from InsiderAdvantage. Someone else also put up a new Mississippi poll from American Research Group (I don't know if we're using them as a source). Whoever knows how to update the map please do. --Mbenzdabest (talk) 18:31, 7 March 2008 (UTC)

Changing the Average Support in Recent Polls Box (Again)

  • I posted this earlier, but no one has seemed to pick up on it:

This box has not been updated in quite a while. I don't have a clue about how to change it, so who would like to do the honors? Call me a "self-declared Hillary supporter" but her North Carolina margin is off (the latest poll shows Obama 47-Clinton 40). Could someone make this more accurate? I think that Texas, Ohio, Rhode Island, and Vermont can be removed, since they have occurred already. Jellyfish101 (talk) 21:03, 8 March 2008 (UTC)

Delegate Source

I updated the delegate counts in the table to use numbers from each state article, and added a note to this effect. This is what Democratic Party (United States) presidential primaries, 2008 and Results of the 2008 Democratic Party presidential primaries are doing so it makes sense to do it here too. Andareed (talk) 11:07, 9 March 2008 (UTC)

Michigan and Florida re-dos?

At least one set of polls (Rasmussen Reports) has been released for Florida and Michigan in the case of re-do Democratic primaries? Should a section be added for polling in states with potential primaries? PNDebater (talk) 18:31, 9 March 2008 (UTC)

  • I think they states should be readded to the maps and saded for the leader on current polling for a revote and the polls can be added with a note that its for a possible revote. (Tjliles2007 (talk) 18:39, 9 March 2008 (UTC))
  • Then should the polls for unscheduled potential primaries be in a separate section or with other future primaries with a note naming them as potential?PNDebater (talk) 18:48, 9 March 2008 (UTC)
  • I think making a new section is unnecessary. Making a note that about a possible revote seems sufficient. I will add the two polls to this page and people can move them to a new section if they really want Andareed (talk) 21:34, 9 March 2008 (UTC)

Texas color

Kentucky poll

Does Kentucky have a poll conducted yet for its primary comming up in May to see whos leading? OHWiki (talk) 09:19, 10 March 2008 (UTC)

  • There hasn't been a solely Democratic nomination survey done for Kentucky, however, based on these results which show McCain winning in a head-to-head match against Obama by +36 points but only defeating Clinton by +10 points, you can assume for yourself that Clinton has the advantage over Obama. Also, in the Rasmussen Markets, Clinton is favored to win in Kentucky by a good margin. Many have argued that the futures market is even better than polls at predicting who eventually wins. Also, based on the history of Kentucky and the Clintons, it seems logical that she would have an advantage. --Mbenzdabest (talk) 02:49, 21 March 2008 (UTC)

Merge polls into state primary/caucus articles

I would like to propose that polls be merged into their respective state primary/caucus articles. I don't believe that having giant lists of polls is a very useful way to convey this information. As well, future polls would be added directly to the state pages and not on this page. Any remaining content (e.g. graphs, etc...) could be moved into Democratic Party (United States) presidential primaries, 2008#Opinion polling. Andareed (talk) 02:44, 12 March 2008 (UTC)

Florida Revote

I removed the Florida section because Florida democratic congressional members said tonight the state will not have a revote. (Tjliles2007 (talk) 04:19, 12 March 2008 (UTC))

I haven't heard about that. Do you have a source that Flordia definitely will not have a revote? Yeldarb68 (talk) 12:56, 12 March 2008 (UTC)

I dont think it's really that definitive. [10] --Siradia (talk) 14:07, 12 March 2008 (UTC)
  • I agree those polls should never have been included in the same section as the other polls until a definitive revote is announced for both states. However, I still think they should be included somewhere in the article, maybe under a different section. The reason stated above is not enough to remove the polls; you are relying on only one delegation group in Congress. The congressional representatives of Florida cannot single-handedly decide what Florida can and cannot do. Senator Bill Nelson is still actively pushing for a Florida revote. And Howard Dean and the DNC are willing to look at multiple ways of seating the delegates. Besides, the Florida Senate and Governor Charlie Crist, both of which have more power in terms of deciding whether Florida will or will not have a revote versus the Congress delegation, are still in favor of a revote. Therefore, the Florida Senate trumps the U.S. Representatives and their voice is more powerful, in favor of leaving the polls. Until an official statement by the DNC and/or the STATE (and not the representatives in Congress) is given the polls should remain somewhere. If no one can find a more compelling reason for deleting the polls they should be added again.--Mbenzdabest (talk) 15:09, 12 March 2008 (UTC)
  • The Florida Democratic Party has just announced that there will not be a revote in Florida. I have again deleted the Florida polling section from the article as it is no longer relevent. (Tjliles2007 (talk) 22:04, 17 March 2008 (UTC))

Totals don't add up

Well, I once before tried to make an edit to make certain numbers match up with reality - and was aggressively reverted. So this time I figured I'd just post here and see what you have to say. See I had this crazy idea that totals were supposed to actually add up, like say in the delegate table on this page. Having applied addition to the best of my abilities, I make a total of 1243.5 delegates for Clinton and 1406 delegates for Obama (not counting Michigan and Florida, although obviously the numbers that include these are wrong too). Can anyone give a reason why hundreds of delegates should not be included in these totals (and need I say in a way that favors a certain candidate)? —Preceding unsigned comment added by ErikRydbert (talk • contribs) 02:32, 17 March 2008 (UTC)

  • I will go and add them up and insert new totals... You mention that the actual numbers used to get a total were wrong... The numbers used here are generally from CBSNews unless otherwise sourced. If you had changed a states total as part of your edit for a new grand total and those numbers were not from CBS or were not sourced otherwise then thats probably why your edit was reverted. (Tjliles2007 (talk) 14:49, 17 March 2008 (UTC))
  • Iowa recently held their second round of the process which changed the totals some, California recently certified their results which included a minor change, Colorados has been incomplete all along those results could have changed recently and it wasnt updated on here or someone changed it using another source recently and I didnt notice yet to revert it back... I Will check them all later and either source the numbers or revert them back to CBS News totals. (69.123.81.12 (talk) 20:22, 25 March 2008 (UTC))

North Carolina Polls/Map

Will someone please change North Carolina to striped? Barack Obama and Hillary Clinton are in a statistical dead heat there (44%-43%). (Jellyfish101 (talk) 15:31, 20 March 2008 (UTC)

We should wait until at least one other survey confirms the result. 83.227.141.19 (talk) 21:51, 20 March 2008 (UTC)
Agreed. It has been discussed before (see above; discussed prior to Texas and Ohio primaries) that that whole striped state deal is pointless. Because so many states are/were tightly contested between Obama and Clinton, so many states would be striped that it would defy the purpose of even having a map that shows one candidate having an advantage over another. If Obama is winning by one point in the latest poll and by even more in the older ones, it should remain for Obama. --Mbenzdabest (talk) 02:53, 21 March 2008 (UTC)
A new poll indicates that Obama has regained the lead. Andareed (talk) 17:13, 25 March 2008 (UTC)

Should there be a separate map with the leader in delegates?

I ask because the popular vote, while important to look at, is not as important as the delegate count for each state. For example, the map shows that Clinton won Texas, and she did -- the popular vote, anyway. In terms of delegates, I think Obama beat her there, and ultimately, that's what's more important. Perhaps a separate map for actual delegates would be better, perhaps showing the percentage each candidate won by. Or maybe it doesn't need to be that specific yet, but Missouri, New Hampshire and New Mexico would definitely be stripped, and Texas and Nevada would be a different color. -- Frightwolf (talk) 01:56, 26 March 2008 (UTC)

There's already an entire section with tables dedicated to delegate count. --Mbenzdabest (talk) 01:58, 26 March 2008 (UTC)
True, but the table also has popular vote count, and we still have the map for popular vote, anyway. I see no reason why we shouldn't have a map for, really, the more important factor: delegates. The pop. vote map is there for a quick, easy "who-won-what" kind of deal. Might as well have another one for a quick, "who-won-what-delegates" overview, with the tables giving a more in-depth look at it. -- Frightwolf (talk) 02:02, 26 March 2008 (UTC)

Map colors for states without recent polling data

The gray and the light purple are very close to each other such that all the gray states (like Oregon, for example) look like they are Obama poll winning states. Should the gray be darkened/changed so that this is not misleading? —Preceding unsigned comment added by 71.140.67.191 (talk) 14:01, 26 March 2008 (UTC)

I got confused with that also... but then I tilted my laptop screen back and the colors became clearer... try adjusting your monitor settings. --Mbenzdabest (talk) 02:11, 28 March 2008 (UTC)
I darkened the gray so that this differences were more evident. Should there be a gray block added to the color key? —Preceding unsigned comment added by Chimera16 (talk • contribs) 13:11, 30 March 2008 (UTC)
No. There really isn't a point to doing that. I wouldn't say that it is confusing right now. Anyways, it will be completely color coded in a couple of months so we should just leave it be. Cheers.--Burzum (talk) 19:41, 2 April 2008 (UTC)

New Pennsylvania Poll

I think that the new Pennsylvania poll showing Obama leading Clinton 45-43 is a little shaky. I find it hard to believe (and yes, I am a Hillary supporter) that Obama can suddenly take the lead from a 12-point deficit. Anyone agree? —Preceding unsigned comment added by Jellyfish101 (talk) 15:18, 2 April 2008 (UTC)

Sorry, this isn't a discussion board. This is only an article to tabulate polling data from notable polling firms, whatever values they find. It is irrelevant to this article whether those values are in error or not as long as they come from well known polling firms. Cheers.--Burzum (talk) 19:44, 2 April 2008 (UTC)

either way, the numbers weren't fabricated and the state should be shaded for Obama. --128.205.116.44 (talk) 19:46, 2 April 2008 (UTC)

This is the only PA poll that has ever shown Obama in the lead. At best, it should be coloured as being tied/too close to call. Selecting outliers clearly fails WP:NPOV Andareed (talk) 20:01, 2 April 2008 (UTC)

Andareed, I'm not sure that I agree with you. Another poll shows that Clinton has a 9-point lead over Obama. I don't think that is it tied or too close to call, so I propose leaving it the way it is. Again, don't get angry with me just because I support Hillary. I know how some of you are quick to judge. —Preceding unsigned comment added by Jellyfish101 (talk) 16:17, 2 April 2008 (UTC) —Preceding unsigned comment added by 74.70.95.48 (talk)

I completely agree with you. Andareed (talk) 20:33, 2 April 2008 (UTC)

Non-issue. Not a toss up.[11] Changing the color at this point would be quite premature. --Kallahan (talk) 20:26, 2 April 2008 (UTC)

It doesn't matter whether or not you agree with the poll. What matters it the poll exists. Now, one poll isn't enough to say that Obama is now leading (I think it was agreed that we need at least three consecutive ones or something to determine that). I think this is a fair assessment for all sides. =) -- Frightwolf (talk) 21:28, 2 April 2008 (UTC)

Puerto Rico poll

[12]

I don't speak Spanish, but from what I can tell it's 50% Clinton to 37% Obama, surveying 800 people, with a margin of error of 4.4%. Andareed (talk) 18:26, 10 April 2008 (UTC)

I added the poll. Thanks for the tip.--Patrick Ѻ 20:14, 10 April 2008 (UTC)

How about "new" tags?

As some one who often visits this page to see if there are new polls: It would be great if new additions to this page could be tagged as new. Maybe a template which appears to readers as a small label saying "new", while editors would see the date when it was added, and could remove the "new" label after a certain time (e.g. after three days). CuriousOliver (talk) 19:22, 14 April 2008 (UTC)

Indiana poll

To the IP user that continues to revert the new poll for Indiana, it doesn't really matter what the numbers are. This is an article that shows all the polling, whether it shows a huge jump or not it's still a poll from a reliable source. It's not some crazy whack-job poll. It could also have a huge difference depending on where the polling data was received from. The poll should remain. While I have the same bias towards Obama, I think the IP user's bias towards Hillary is effecting the user's choice to remove the poll (that shows Obama leading). HoosierStateTalk 22:35, 15 April 2008 (UTC)

  • I am a Hillary supporter and I agree with you all polls should be shown. I do however have questions about the LA Times poll (I assume thats the one you refer to), but it is a poll, we will know soon enough how accurate it is. (Tjliles2007 (talk) 17:04, 16 April 2008 (UTC))

Map Changes

To whoever made those changes in the map, I think it would be best if you first discussed it with the Wikipedia community before making such drastic changes. Anyone agree? Jellyfish101 (talk) 23:27, 16 April 2008 (UTC)

Obama is ahead in the most recent Indiana and Pennsylvania polls. Why are those states yellow? —Preceding unsigned comment added by Bop me two times (talk • contribs)

In regards to this and the Indiana polls, for the last time, this is not a page where we judge the merit of the polls, but rather list them and update accordingly. Indiana was switched to Clinton after just one poll showed her winning, following polls later agreed, but that didn't stop the initial switch. If you can't be consistent, stop editing. Pennsylvania and Indiana both should be colored for Obama. 134.197.66.126 (talk) 17:49, 16 April 2008 (UTC)
  • I didn't revert it because Obama was highlighted as leader in Indiana. I reverted it because someone decided to go and totally redo the map from popular vote to delegates. Please feel free to go change Indiana to purple using the popular vote layout. (Tjliles2007 (talk) 17:59, 16 April 2008 (UTC))
I will tell you why Indiana and Pennsylvania are colored yellow. We had this discussion way back when Texas and Ohio were holding their primaries. Look at #46: Changing Texas, and you will find your answer. One poll is not enough to claim that one candidate is leading over another. If another poll is released saying that Obama is leading in either of the states, it can be changed to purple. And by the way, I disagreed with the change of Indiana to yellow when just one poll showed her leading. I later agreed with it when more polls showed her leading. (Jellyfish101 (talk) 14:01, 16 April 2008 (UTC)) —Preceding unsigned comment added by 74.70.95.48 (talk)

South Dakota

someone should color south dakota for Obama. A recent poll came out with him leading. Marshie71 (talk) 01:04, 17 April 2008 (UTC)

Pennsylvania & Indiana (Map)

The most recent poll in Pennsylvania and in Indiana suggests that Obama will beat Clinton. So why hasn't the map been changed?

~Willwal. Posted at 16:10 on Thursday 17th April 2008. —Preceding comment was added at 15:11, 17 April 2008 (UTC)

One poll said that. We should wait until at least a few polls show similar results. Andareed (talk) 17:52, 17 April 2008 (UTC)


-This is my first post on Wikipedia, so all apologies if I end up doing something disasterous when I edit this.

I'd just like to point out that if we're going to only color states if there are multiple polls showing a canidate in the lead, shouldn't Oregon and Puerto Rico be shaded back to grey? —Preceding unsigned comment added by Xenomorphious (talk • contribs) 20:10, 17 April 2008 (UTC)

That sounds reasonable to me. Andareed (talk) 20:16, 17 April 2008 (UTC)
  • It's not a big deal, In a state that has say 6 polls in the past week and just one shows a candidate leading then yes you should wait until a second poll confirms it, but theres no other polling data to suggest Obama isn't leading in South Dakota. I'm a Clinton supporter and I concede that Obama will probably win South Dakota, unless he keeps making elititst quotes while dissing small town america. (Tjliles2007 (talk) 13:32, 18 April 2008 (UTC))

!!!!!!! (Tjliles2007 (talk) 03:16, 18 April 2008 (UTC))

Please keep opinion polling and add actual vote results

It would be very useful to have the opinion-poll charts remain online after the primary in Pennsylvania, and to add to them the actual vote results, so readers can compare polling data with real returns. In this year of unprecedented turnout and unusual demographics, the polls are widely divergent, probably in large measure because pollsters can't get a good handle on who's going to actually turn out and vote. It would be good to see poll data compared with election results in order to make some judgments about which polling company produces better indicators of what will happen in the voting. —Preceding unsigned comment added by 98.207.113.79 (talk) 08:53, 18 April 2008 (UTC)

  • All polling data from this article from previous primaries remains, they are now in a seperate article of past elections and seperated as Pre supertuersday, Super Tuesday, February, March to keep down the article sizes. Links to all these articles can be found on this article. (Tjliles2007 (talk) 13:28, 18 April 2008 (UTC))

Texas Being Striped On The Map

Whoever keeps changing Texas to that gold-purple striped look needs to stop, or at least discuss it with the Wikipedia community first. Anyone agree? unsigned comment added by Jellyfish101 (talk) 15:31, 18 April 2008 (UTC)

That, and that it is also revolting to the eyes. Xenomorphious (talk) 19:46, 18 April 2008 (UTC)

  • Any particular reason we decided to not use the state for the primary and the seperate square with the others for the caucus? We had it that way for several weeks with no complaints. (Tjliles2007 (talk) 23:01, 18 April 2008 (UTC))
  • Alright, now someone has decided to change the map according to delegate results. I thought it was pretty much decided that the popular vote was how we colored the map? 4.246.159.161 (talk) 01:15, 19 April 2008 (UTC)

Map edit war

There is obviously a map edit war going on so I'd like to say two things:

1. give the two maps different names, and
2. discuss which map is applicable on this page.

Can we discuss this issue and try to reach a consensus? Perhaps we should start with a RFC.--Burzum (talk) 01:18, 19 April 2008 (UTC)

Belligerents
On the side of popular-vote map: Clinton supporters
On the side of delegate map: Obama supporters, logic—Preceding unsigned comment added by Bop me two times (talk • contribs)

Okay, we don't need to get into any heated argument about the maps. Earlier on this page I asked about a delegate-count map because I thought it was far more applicable to who was winning. But I just thought it'd be neat to have an extra map since they barely take up space and both tell us a lot in a compact visual.
If we have one, I say it should be the delegate count, but really, would there be a harm in two maps being next to each other? No. -- Frightwolf (talk) 03:42, 19 April 2008 (UTC)

Maybe we could add new colors to show a Clinton popular vote win/Obama delegate win? Although you'd also have to add colors for Obama/Clinton popular vote wins/delegate ties, and all the colors may be confusing...

Just a thought. Xenomorphious (talk) 04:19, 19 April 2008 (UTC)

Completely New Idea
What do you think of this idea which might resolve this debate once and for all:
This is a page all about statewide polling. If people want to see results they can go look at the results page which has a lot of nice maps. How about just showing the polling results for states where elections are still to come. All the other states where elections have already occurred could be some neutral color.
Here is an example:


File:Democratic presidential primary polls.png

What do you think?--Chimera16 (talk) 14:51, 19 April 2008 (UTC)

I think the problem with that is most primary pages have a map showing who won what state, so we'd have to bring it back eventually. So I wouldn't recommend switching to this (even though I understand what you're saying). -- Frightwolf (talk) 17:56, 19 April 2008 (UTC)
Yes, possibly. The Democratic Party (United States) presidential primaries, 2008 article includes two maps but ignores the polling. Thus all of the effort here to make the "One True Map" is wasted. I would argue that there is no need of a map at all for this article since this article is just a collection of opinion polls. And if a map were to be made for the opinion polls it would certainly be only of a temporary nature. But if people want a map, yours is as good as any and it would be appropriate for an article on opinion polls. I again don't think it is necessary since it will be obsolete in six weeks or so, but it would be a better alternative than wasteful edit wars on redundant maps. Cheers.--Burzum (talk) 18:03, 19 April 2008 (UTC)

Pledged Delegate Count
I'm interested in fairness in reporting the results. Shading Nevada, New Hampshire, or Texas for Clinton, or shading Missouri for Obama, is inaccurately reporting the results of the primaries or caucuses from those states. We can not claim that Obama has won Missouri or that Clinton has won New Hampshire, because each candidate received the same number of delegates, which is what will be used to determine the nominee at the Democratic National Convention. We also cannot claim that Clinton has won Nevada or Texas, since each of these states will be sending more Obama delegates to the convention. Here is the delegate count from each of these four states. This count is what determines the Democratic nominee, not the popular vote.

  • MO: Clinton 36, Obama 36 (TIED) [13]
  • NH: Clinton 9, Edwards 4, Obama 9 (TIED) [14]
  • NV: Clinton 12, Obama 13 (OBAMA WINS) [15]
  • TX: Clinton 65 + 29 = 94, Obama 61 + 38 = 99 (OBAMA WINS) [16]

Whether we have one map or two maps, I don't care. It seems there are three different maps being used. One uses popular vote exclusively, one uses pledged delegates exclusively, and the other uses popular vote except for striping texas. The pledged delegate map is the one that accurately represents how these state delegations will be voting at the convention. People also feel the popular vote is important, although it plays no role in the determination of a nominee except perhaps with the Superdelegates' decisions. If this is included as a secondary map, I wouldn't object. But we must not misrepresent the results by giving the impression that any candidate has won a state that will not send more delegates for that candidate to the convention. Vir4030 (talk) 19:42, 19 April 2008 (UTC)

Actually, I do like Chimera's map. This article IS about opinion polling, not results. Xenomorphious (talk) 21:06, 19 April 2008 (UTC)

I agree the article isn't about results, but I think until this primary is over and the map is removed, it undoubtedly will be and this article will most likely be renamed or split, we have to include results alongside polls as it allows a person to put the polls in context without having to jump between articles. However, on that note there should absolutely not be a map for delegate results as the polls do not measure that. Opinion polls measure the popular vote and so it should only include popular vote, though noting that delegate count could be different. If someone wants to track delegates they can look at the results article.--The Devil's Advocate (talk) 21:13, 19 April 2008 (UTC)

Following this argument, I put the map down in the results section, where the delegates table is. I think this is consistent with how the popular vote map is treated. I understand your point about this being an opinion poll page, and I apologize for not realizing that sooner. Vir4030 (talk) 00:11, 20 April 2008 (UTC)

Now that it seems decided to include the delegate map, perhaps we should remove "Projected" states(OR, SD, IN, KY, WV, PA, NC) as these reflect popular support, as opposed to delegate results. Xenomorphious (talk) 20:42, 20 April 2008 (UTC)

Do we have any other way of projecting delegates in upcoming states aside from the popular vote? I don't think so. The delegate allocation rules are complicated and change from state to state. Vir4030 (talk) 21:06, 20 April 2008 (UTC)

Why not show all Zogby Poll results?

I think it is inconsistent to show only the latest Zogby poll result, while different results from other polls are shown. CuriousOliver (talk) 14:49, 20 April 2008 (UTC)

So add them. 24.13.141.176 (talk) 15:32, 20 April 2008 (UTC)

Some one keeps removing them. CuriousOliver (talk) 10:21, 21 April 2008 (UTC)

Zogby is a tracking poll which lists the polling results for every day on each press release. PollShark (talk) 09:05, 21 April 2008 (UTC)

Every poll is a tracking poll. CuriousOliver (talk) 10:21, 21 April 2008 (UTC)
How about creating a table (with dates as columns) for tracking polls? Andareed (talk) 15:49, 21 April 2008 (UTC)
Which ones do you consider tracking polls and why? CuriousOliver (talk) 23:12, 22 April 2008 (UTC)

I think we can continue as we are, and that the Zogby polls can be included. Rasmussen Reports is also a tracking poll, and it's used. There are others. I don't see why there's a problem as long as we're properly reporting the date range for the poll and including it in the list according to the dates. We shouldn't remove factual sourced data. CuriousOliver, whenever the data is removed, re-add it and put a comment here so we can address it. Vir4030 (talk) 16:32, 21 April 2008 (UTC)

Ok, every poll is not a tracking poll, that's ridiculous and Rasmussen's only tracking poll is for the general election, their statewide polls do not track. So I believe a definition of a tracking poll would be helpful here. A tracking poll is a continuous poll run over several days with results posted daily. Zogby's is the only tracking poll for Pennsylvania like they also did for Ohio and Texas and several Super Tuesday polls. If you would like to view the results from previous days you can see them in each day's press release. PollShark (talk) 21:43, 21 April 2008 (UTC)

So tracking poll is a poll that lists results of previous polls? By that standard, most polls here are tracking polls, not just Zogby. CuriousOliver (talk) 23:20, 22 April 2008 (UTC)
A tracking poll is a poll that generally polls the same people and asks the same questions daily (or every few days) and reports these results. Andareed (talk) 01:27, 23 April 2008 (UTC)
Where do you get the idea from that they ask the same people again? I don't see that mentioned anywhere, and as far as I know it is impossible to do for these sample sizes. The only difference I see between Zogby and a few others is that Zogby have recently been daily, with some other polls being weekly. CuriousOliver (talk) 23:11, 22 April 2008 (UTC)
I could be wrong about asking the same people; I could see people getting irritated by pollsters calling everyday. Andareed (talk) 11:42, 23 April 2008 (UTC)
I think the issue is more about the phone ringing and no one picking up, or not the same person. For polls like these, trying to track down people is way too costly, and there is no big benefit. For these polls, it is all about who picks up a phone or bumps into a poll staff and volunteers. CuriousOliver (talk) 14:56, 23 April 2008 (UTC)

One other comment, if one was to keep every result for a tracking poll like Gallup and Rasmussen for the general election, the site would be so immense that there would be no holding it. I believe that limiting the size of the page is a key factor for moving this site towards an A rating.PollShark (talk) 21:56, 21 April 2008 (UTC)

I grouped the Zogby tracking polls together for each day of tracking and left the unaffiliated Zogby poll for the 9th and 10th alone. This ensures that each day's tracking is recorded on the page without excess links. I believe this is what's great about this site is that people with differing views can work together to create one finished product. This is the spirit of America. Thanks for letting me work on this site as well as the other ones that I have created for the Senate elections and the electoral college elections. Your friend. PollShark (talk) 05:42, 22 April 2008 (UTC)

Mr. Smith didn't like the layout so I changed it in regards to a suggestion by one of my collegues and put a seperate table to show the tracking polls. Your friend PollShark (talk) 06:09, 22 April 2008 (UTC)

I believe your distinction between tracking polls and other polls is arbitrary, ill-defined and serves no purpose I can identify. To move the results from one source to an extra table on the sole ground that they happen to do polls more frequently than others is confusing and has no benefit I can think of. First, you say there are too many results, then you say there are too many links, but links take very little space, so I don't see your problem. CuriousOliver (talk) 13:46, 22 April 2008 (UTC)
Like I said earlier, it's not that they do more polling but that they do a different brand of polling. A tracking poll's use is that you can see the changes over a period of time. Having results of a tracking poll scattered through a table with other polls is, to borrow your term, confusing and time consuming. --PollShark (talk) 17:36, 22 April 2008 (UTC)
Why is it a different brand of polling? We publish the number of people they asked, the dates on which they ask it, and the results they get. I don't understand the fundamental difference between this and any other poll. Is it just a "tracking" poll because they do it every day? The Rasmussen Reports poll is done every three days, is that not often enough to be a "tracking" poll? And why does it even need to be separate? I'm sorry, but I don't understand this at all. Vir4030 (talk) 18:48, 22 April 2008 (UTC)
It is sometimes hard to compare different polls by different pollsters because the question asked is different and because the demographics are different. Tracking polls are useful because they typically ask the same people the same questions on different days. Hence these variables are controlled and you can get a better picture of how the polls are actually changing. Andareed (talk) 19:01, 22 April 2008 (UTC)
Where do you get the idea from that they ask the same people again? I don't see that mentioned anywhere, and as far as I know it is impossible to do for these sample sizes. The only difference I see between Zogby and a few others is that Zogby have recently been daily, with some other polls being weekly. CuriousOliver (talk) 23:11, 22 April 2008 (UTC)
Tracking Polls are polls done over night, partly using results from the previous day's polls. So, in Zogby's tracking polls, for example... They would start their polling, release the results, then continue that same poll with another set of voters. They let these results build up for a few days, and then they start to get rid of the last day when they add the new data. So if they are using 4 days of polling data, they would start on the 10th and poll until the 14th... then the next day they would use only the results from the 11th to the 15th, then the next the 12th-16th, and so on. This is a common method used to find momentum. -Luckydevil713 (talk) 07:26, 26 April 2008 (UTC)

West Virginia

I do not believe that the poll from March 13th in West Virginia is still valid. It is from over one month ago, and as we all know, the polls change daily. I think it should become gray until a new poll comes out. Does anyone agree?Marshie71 (talk) 17:08, 21 April 2008 (UTC)

I agree. It is over a month old, and I would support shading it to gray until a new poll is released. Vir4030 (talk) 18:26, 21 April 2008 (UTC)
Part of the problem is that there is little motivation to run polls in WV, since it's expected that Clinton will easily win there. Andareed (talk) 18:29, 21 April 2008 (UTC)
Well, I don't think that we can shade the map in her color only based on this assumption and a poll that is over a month old....Marshie71 (talk) 04:07, 22 April 2008 (UTC)
It's probably mostly a pointless argument anyways, since there will be more polling soon. Vir4030 (talk) 04:38, 22 April 2008 (UTC)
Good point Marshie71 (talk) 05:57, 22 April 2008 (UTC)

I see double daggers behind the most recent two WV polls, but no explanation what they mean. Explain or remove? CuriousOliver (talk) 14:52, 23 April 2008 (UTC)

April and May polling pages

I've decided to combine both April and May into one article, since PA would be the only state on the April article. PA, IN, and NC have been linked together in the media, and April 22 is near the end of April, so it makes sense to me to group them as such. Andareed (talk) 03:19, 23 April 2008 (UTC)

It seems someone beat me and created Statewide opinion polling for the April and beyond Democratic Party presidential primaries, 2008, which seems fine too. Andareed (talk) 03:21, 23 April 2008 (UTC)

Indiana Democratic Primary

I think Indiana needs to be changed to striped. If you look at the last two polls, Obama and Clinton are in a statistical dead heat (48%-47%) in the first one, and Clinton leads Obama in the second one (50%-45%). Anyone agree? Jellyfish101 talk 14:48, 25 April 2008 (UTC)

  • Nope. If the next poll shows Clinton ahead it will be shaded back to her. I think we agreed that colors be changed after multiple polls confirm a leader. I expect SurveyUSA will have a new poll out early next week, if shes ahead in that we can reshade it then. (Tjliles2007 (talk) 20:56, 25 April 2008 (UTC))
Absolutely not Jellyfish. I think people should STFU if they haven't read the existing talk page.71.94.5.225 (talk) 08:41, 26 April 2008 (UTC)

Now someone has colored Indiana for Clinton (no stripes, just yellow) without any new polls, and also Montana, where there have been no polls conducted since December. Surely this should be reverted.125.239.165.140 (talk) 07:05, 27 April 2008 (UTC)

The same editor also replaced the map showing results by delegate count with a red link.125.239.165.140 (talk) 07:13, 27 April 2008 (UTC)

Pennsylvania On The Map

Whoever fixed the map to shade Indiana for Clinton also made a big error. The map now shows Clinton leading Pennsylvania, where she has already won. I'm not quite sure how to fix it. Could someone help me? Jellyfish101 talk 16:38, 28 April 2008 (UTC)

  • Someone went back and reverted an old map to show Clinton leading in Indiana. That person should have just left it alone until someone updated the map fully. (Tjliles2007 (talk) 21:46, 28 April 2008 (UTC))
A new poll released today by SurveyUSA confirms a Clinton lead in Indiana, so it is correct. However, the map currently displayed shows Clinton leading in Pennsylvania. She won there last Tuesday. Could someone change it, because I don't know how. Jellyfish101 talk 18:40, 28 April 2008 (UTC) —Preceding unsigned comment added by 74.70.72.142 (talk)
No it is not correct. We agreed, three polls are necessary to switch a shading. Why why why why why why why is this still not getting through to people. I am requesting mediation.Chester-S (talk) 00:21, 29 April 2008 (UTC)

Bad polls

A recently posted Insider Advantage poll[17] used exit poll data from the 2004 general election[18] to select their respondents. It would be bad enough if they ised 2004 Democratic primary demographics, but general election data is even worse, since North Carolina's November electorate tends to be older, wealthier, and whiter (though not more male as in many states) than its Democratic primary electorate. Should we even post polls this bad, or should we put notes next to them like I did to this one? SteveSims (talk) 06:53, 1 May 2008 (UTC)

  • A poll is a poll and Insider Advantage shows what other polls in NC show, a trend towards Clinton. In two weeks Obama has gone from a 15 point lead to a 2 point deficit in the IA polls. (Tjliles2007 (talk) 13:10, 1 May 2008 (UTC))
  • Making any sort of claims about whether or why a poll is "bad" would constitute original research. If you can reputable source that rips the poll apart, then I see no problem with noting it. Andareed (talk) 04:06, 2 May 2008 (UTC)

Changing Map

  • Please stop changing the map unless a poll shows a new leader and then a second poll confirms it with no polls between that contradict it. One poll shows Clinton ahead in North Carolina but that doesnt warrant a recoloring of the map until a second poll confirms it. (Tjliles2007 (talk) 21:03, 1 May 2008 (UTC))


Tracking and Graphs

I thought that since Zogby is doing its tracking again, that I would once again seperate the poll and this time I prepared Gallup style graphs for the poll. I think this gives a good idea of how the races are moving and whatnot. Let me know. —Preceding unsigned comment added by PollShark (talk • contribs) 09:23, 3 May 2008

My problem with this is that these polls are not any different than any others except that they are run every day. We should definitely put these in the table with the rest of the polls, sorted by date like the rest of the polls. Putting them above and giving them their own graph gives them too much significance. They are no more correct than any other polls. If you want to make a graph of ALL the polls, that's different - go ahead. Vir4030 (talk) 16:46, 3 May 2008 (UTC)
Your atill thinking of a tracking poll as a buch of different polls, it's not. It's basically one long poll. It's the Zogby tracking poll not tracking polls. For this reason results should be grouped together. We reached this consensus in Pennsylvania so this rule still stands.PollShark (talk) 18:00, 3 May 2008 (UTC)
We didn't reach this consensus, and there's no rule. The first poll was taken from 4/30 - 5/1. The second poll was taken from 5/1 - 5/2. The results are no different or more important than any other poll. The results should not be separated out like that because it makes them appear more important, when they're not. Results should all be grouped together in one table. If a tracking poll is a separate entity and needs to be charted differently, why is there no Wikipedia entry for tracking poll? I'm reverting your changes. Vir4030 (talk) 01:18, 4 May 2008 (UTC)
I concur with Vir4030, there is no consensus here. In fact, the assumption that a tracking poll yields different results from another poll is not established. In that light, until such time that the information of a tracking poll and a solitary poll are proved to yield different levels of accuracy, this discussion should be over. Remember to cite sources and avoid original research.Chester-S (talk) 02:45, 4 May 2008 (UTC)
Check the Pennsylvania results, there is a tracking section. Other than that all I can say once again is that a tracking poll is one poll, not a series of polls. The results are simply tabulated every two days in this case, as much as three or four in others. Say I went out and told you to conduct a poll from April 30 to May 4 and report back to me at the end of the process, this would be a run of the mill poll. If I asked for results every two days, this would be a tracking poll. Either way your conducting A poll from April 30 to May 4. I don't know how better to explain it. Once more, if any other pollster conducted a tracking poll, it would be included in the section. If it would help I will put the table at the end of the others so as not to bring politics into a non political act such as research and data analysis like some others have done. PollShark (talk) 07:01, 4 May 2008 (UTC)
I don't understand the difference between one poll and a series of polls. If I ask 300 people every day what they think, and I release a poll every day counting the 300 people from that day and the 300 people from the day before, and I sell this as a 600-person poll, there is no problem. This poll is equally as valid as a poll which only asked 600 people. The fact that I'm doing it every day, and that I'm only taking the most-recent two days' worth of data doesn't make it any more or less correct or important. There are no Pennsylvania results, they have been removed from the page. If they look like that in the archive, it's because they happened to be that way when they were archived, and now it's irrelevant because we have the results. Zogby is releasing poll results every day, and these results should be entered into the table just like all the other results. Vir4030 (talk) 17:19, 4 May 2008 (UTC)
I reviewed the archived Pennsylvania results, and these tracking polls were counted in the table and also separately. In addition to the compromise I offered in the article of marking the tracking polls, if we were to repeat these polls in a table after the main table, I would not be opposed to that. As long as they appear in the table in chronological order with the rest of the polls also, this would be okay. Vir4030 (talk) 18:03, 4 May 2008 (UTC)
Oppose I really think that separating tracking polls causes more bad than good. It tends to give the impression that these polls are somehow special (they're not really), and breaks the chronological ordering that we previously had. I originally thought they might make the articles shorter (which they generally due), but I don't believe this benefit outweighs the negatives mentioned above. Andareed (talk) 12:01, 4 May 2008 (UTC)
Even putting this AFTER is a problem. Now it appears that they are less important. The poll that is the most-recent does not appear at the top of the list. This is making the results confusing for the visitors to this page. There are now three editors in opposition to this, and clearly there is no consensus. I am reverting these changes and putting the polls back into the table in chronological order. Vir4030 (talk) 17:15, 4 May 2008 (UTC)

Zogby

Hey guys, I see this page on a daily basis at least 5 or 6 times. Honestly, I have been comparing the Zogby polls from the past and present this year and find them to be utterly misleading. I think Zogby shouldn't get the importance it is getting. As an Obama supporter, I still think it would be fair if the poll reflected results closer to what reality is. —Preceding unsigned comment added by 71.193.13.142 (talk) 23:29, 3 May 2008 (UTC)

Misleading or not, I think our job here is to display all poll results in one table organized by something neutral. As far as I have been aware, we have been sorting them by final date, initial date, then decreasing sample size. As long as all polls are listed in the table in order according to this neutral criteria, then we are being fair and doing a service to readers. Vir4030 (talk) 01:30, 4 May 2008 (UTC)
For the last time: it does not matter whether you think a poll is more valid then another. The purpose of this site is to list the polls, not pass judgements on their validity. Why is this such a difficult prospect for people to grasp? It makes it exceedingly difficult to assume good faith when people purposefully ignore discussion before commenting or editing. Or in this case, ignore the very principles of the enclyopedia they are participating on (no original research.) This isn't rocket science people.Chester-S (talk) 02:41, 4 May 2008 (UTC)

Indiana Democratic Primary

Indiana definitely needs to be changed to striped. If you look at the last two polls, Obama and Clinton are in a statistical dead heat in both. Jellyfish101 talk 13:39, 4 May 2008 (UTC)

The last two polls are both Zogby. We need something else to confirm. Don't worry, the color of this state on Wikipedia doesn't have any bearing on the outcome of the race. It will be dark yellow or purple in a couple days. Vir4030 (talk) 17:43, 4 May 2008 (UTC)
Yeah, shouldn't this be sripped, or kept light yellow? The tracking polls act as a compilation. Shouldn't they be counted as one? Real Clear Politics shows an average advantage for Clinton, meaning that they count the compliation Zogby polls as one. BTW, for anyone that wants to know, Zogby endorsed Obama. --Criticalthinker (talk) 06:49, 5 May 2008 (UTC)
Yeah, it should be kept light yellow until we get two polls verifying that Obama has moved ahead of Clinton. The way we have the table now, it's much more clear that there's only one result so far. Vir4030 (talk) 07:26, 5 May 2008 (UTC)

What to do with Zogby?

Jellyfish's comment has made me think that there is something we need to mark. The thing about these "tracking polls" is that the sample used in each poll overlaps with the previous and next polls. If you were to look at a table of polls, add up the sample size of each one, and try to use it to create a "poll of polls", you would be counting people twice. I still do not think these should be separate from the main table, because the table lists them in chronological order without bias. It seems wise, though, to mark these somehow and say "the sample used in this poll overlaps with other polls in this table". This is the distingiushing characteristic, is it not, of one fo these "tracking polls"? Vir4030 (talk) 17:53, 4 May 2008 (UTC)

It is so and I appreciate your attention to both website and statistical integrety. I believe the comment clearly demarks the nature of the poll. PollShark (talk) 23:37, 4 May 2008 (UTC)
I removed everything but the most recent results from the main table for the tracking poll. I then created a link to the tracking section and placed it with the results in the main table. That way it reduces the clutter, puts the poll on the same plane with the others and makes the previous results clearly visable. I think we have reached a good conclusion to this debate. Your friend PollShark (talk) 06:00, 5 May 2008 (UTC)
We have a winner. It's everybody! Vir4030 (talk) 07:25, 5 May 2008 (UTC)


WP:OR issue with "projected"

In the map at the top of the Poll summary and current pledged delegate count section, I believe the use of the word "projected" may run afoul of WP:NOR, especially Synthesis of published material serving to advance a position. There is nothing in an opinion poll that explicitly states that the winner of the poll is predicted/projected to win the actual contest. When the article states that an outcome is "projected" based on an opinion poll, who is doing the projecting? It seems that the article itself is, which constitutes original research. At any rate, the map seems to be a complete duplication of the map at the top of the "Polling" section, with the exception of the wording issue! I suggest taking the second map out entirely, or at least using "likely to win" as an alternative to "projected", though any language that shrinks the assumption made about the significance of opinion poll results would work. — Swpbtalk.edits 17:37, 5 May 2008 (UTC)

The two maps are indeed different by more that just that wording. The first map colors states won by popular vote, which is the equivalent of the polling found on this page. The second map, in the delegates section, colors states won by total number of delegates earned. Since there is no source on the internet for projecting the number of delegates won, this map projects states based on the latest polls, which is what the text in the map states. This whole page is about collecting the opinion polling for each state which has not yet voted. Coloring a map based on these polls for upcoming contents allows for an at-a-glance understanding of the latest statewide opinion polling. Vir4030 (talk) 05:39, 6 May 2008 (UTC)
Thank you for not even attempting to address why this wording is not problematic. 68.55.89.59 (talk) 19:34, 27 May 2008 (UTC)

Maintenance of the Maps - Wikipedia vs Wikipedia Commons

Until the primaries are over, the maps on this page will continue to change. These maps are being maintained by the editors which maintain this article as a way of illustrating the results in an easy-to-view manner. If the image is moved to the Wikipedia commons, it will make it more difficult to do this, requiring two accounts. Until it is finalized, any other wiki which wishes to use this image may simply use it from Wikipedia instead. Please do not move these maps to the Wikipedia commons until the primaries are over and the map has been finalized. Vir4030 (talk) 05:01, 7 May 2008 (UTC)

New Montana Poll

someone should color in Montana for Obama. A new poll shows him leading. 68.45.9.206 (talk) 14:58, 25 May 2008 (UTC)

Delegate Map Error

Shouldn't the delegate map show Clinton and Obama as tied like Missouri and Guam? The delegate chart shows them both getting 9. If some of Edwards delegates for that state have declared for him, than that needs to be changed in the delegate chart. If not, then New Hampshire should reflect being tied, delegate-wise. Also, it would probably be better than showing that mixed purple-and-gold color if one simply stripes the state with yellow and gold diagonally. --Criticalthinker (talk) 20:54, 27 May 2008 (UTC)

Florida & Michigan

New maps with Florida/Michigan

I made SVG versions, with Florida and Michigan. Both are derivatives of this map. Inkscape is acting funny so I wasn't able to add a box for Texas' caucus, nor was I able to add the text from the PNG popular vote map to the SVG one, so I left the PNG popular vote map in the article. However, I replaced the PNG delegate map with the SVG one, since I didn't need to add a Texas caucus box and didn't need to add much new text.

Popular vote map: http://commons.wikimedia.org/wiki/Image:Democratic_presidential_primary,_2008.svg

Delegate map: http://commons.wikimedia.org/wiki/Image:Democratic_presidential_primary_delegates,_2008.svg


SteveSims (talk) 03:21, 1 June 2008 (UTC)

  • Fix this map or go back to the old one. There is no legend to discern which color represents which candidate, and Clinton did not win Oregon or Washington. Krazy19Karl (talk) 03:56, 1 June 2008 (UTC)
I fixed Michigan and Oregon. Thank you for pointing them out.
The delegate map reflects the results of the primaries and caucuses, not the results of the state conventions, other post-primary events, and uncommitted delegate endorsements. This now is explained in the map's description.
Also, most maps, such as the Senate and congressional election maps, put the legend in the thumbnail, not in the image itself.
However, if it's still important to you, feel free to put legends in the images themselves and place the lengthy text from the popular vote PNG into the popular vote SVG. As you know, I am unable to "fix" these stylistic differences since Inkscape isn't working properly on my computer.
You can download Inkscape for free here. SteveSims (talk) 04:27, 1 June 2008 (UTC)

South Dakota must be colored for Clinton

A new poll places her ahead by 26 points, and the one before that was already obsolete anyways. —Preceding unsigned comment added by 68.243.66.112 (talk) 18:34, 2 June 2008 (UTC)

ARG polls are stupid. They have been consistently wrong. I don't understand why they are even used. —Preceding unsigned comment added by 24.168.35.247 (talk) 18:35, 3 June 2008 (UTC)

What's gonna happen to the article once Montana and South Dakota vote?

Will the article then stay empty without polls?

I suggest that all the polls made for all the states (which are now separated into different articles) be merged together into this article. After all it's "Statewide opinion polling", it must have polls made for all the states. Any opinions? —Preceding unsigned comment added by 70.11.66.185 (talk) 21:51, 2 June 2008 (UTC)

I agree. This article was a great resource and should stay up, and become reorganized, after MT and SD vote, for history's sake. 68.45.9.206 (talk) 00:18, 3 June 2008 (UTC)

  • I think it would be too big to have all together, but the main page could be left with the maps and table and perhaps a recap above with the links to each months seperate page. (69.123.81.12 (talk) 01:35, 3 June 2008 (UTC))

I agree that the article would be too big. This page should have a map and links to each separate page for the different polls organized by state or month, etc. Vikingviolinist (talk) 21:36, 14 June 2008 (UTC)

Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:Statewide_opinion_polling_for_the_2008_Democratic_Party_presidential_primaries/Archive_1&oldid=1136231481"