28 responses so far ↓

• 1 marktarheel // Jul 28, 2008 at 10:48 am

  lets not forget the bradley effect..obama is most likely getting false poll numbers from people who wont actually vote for an african american.probably 3 to 5 percent..thats all mccain will need

• 2 Jack // Jul 31, 2008 at 9:34 pm

  There’s been a lot of analysis of the Bradley Effect on fivethirtyeight.com and other sites. Most political scientists seem to be in agreement that it has greatly diminished to the point of being virtually nonexistent.

• 3 quincyconscience // Aug 2, 2008 at 2:08 am

  Over at fivethirtyeight.com as Jack mentioned, specifically at http://www.fivethirtyeight.com/2008/07/cellphone-problem-revisited.html
  there is a very interesting analysis of polling technique, particularly cell phone issues. The under 25 market is left in the dark almost completely because most major polling agencies do not use any cell phone numbers. This is a problem as I can name 20 of my friends who are voting and don’t have landlines, including myself. This age group of 18-25 is breaking as much as 30 points to Obama. Also many polls are only conducted in English. Hispanic voters have shown historically and in current polls to be leaning Democrat. These factors will probably either eliminate or mitigate considerably the Bradley Effect.

  270TOWIN: Interesting article in WSJ yesterday (8/1) about the validity of automated polls (Rasmussen, SurveyUSA) vs. in-person/phone live polling. (WSJ subscription may be required to read full articles linked to in this post) Pros and cons to both, but the article indicated that automated poll results were very pretty competitive with their traditional counterparts in terms of accuracy, and are much easier/cheaper to conduct more frequently, allowing for a better understanding of how opinions were changing over time.

  In today’s (8/2) WSJ, there is an article talking about when people lie to pollsters. The introductory paragraphs set the tone:

  One of the toughest questions on any poll is whether people are telling the truth. It is a conundrum that looms front and center as voters look ahead to the first U.S. presidential contest that an African-American candidate has a chance to win. With polls showing overwhelming voter support for the idea of a black president, researchers and pollsters are trying to determine who really means it.

  Peter Hart, a Democrat on a bipartisan team conducting the Wall Street Journal/NBC News poll, estimates that 10% of current Democrats and independents who say they support presumed Democratic Party nominee Barack Obama may not be giving a fully honest answer, at least based on their responses to broader questions about race. “This election is exceptionally tricky,” he says.

• 4 mark // Aug 10, 2008 at 7:58 pm

  sorry jack…………..the facts say otherwise……..obama lost 13 0f the last 15 primaries………….especially…ohio…pennsylvania……and as you know.its a majority blu collar white folks.game over. one more thing.fivethirtyeight is pro obama………nice try

• 5 mark // Aug 10, 2008 at 8:00 pm

  doesnt this site update every day?.of twice a year.jeeez………..most of the poll numbers are weeks old

  270toWin: Most states only get polled once or twice a month… there will likely be increased frequency in the swing states as it gets closer to the election.

• 6 quincyconscience // Aug 11, 2008 at 4:21 am

  mark, the 538 site has the most solid projection model anywhere. It is pro-obama in its blog but math is math. Show me math with McCain ahead and I will believe you, but right now there is no reason to believe that Obama has ground to make up before the election. Don’t use attacks of bias against numbers as 538 includes too many polls to have a bias in its math. Also 270 clearly is just using a limited number of trusted polls whenever possible, hence the slow delay in number updates.

• 7 mark // Aug 11, 2008 at 7:41 am

  the last update was on july 25th…….its the second week in august………and gallup and rassmussen are updated daily……….you mean to tell me that has been no movement in the polls that deserves no blog time?.when obama’s lead has shrunk dramatically……….sorry.but the july 25th entry is way outdated……

  270toWin: We haven’t really made many posts about the polls themselves. If anything, the post on the 25th was trying to say that despite Obama appearing to be well ahead in the state-level polling, that the race was closer than it looked. However, you do indirectly make a good point…that the lack of posts can give the appearance of bias where none was intended. Since it isn’t the goal of this site to offer opinion, we will forego any further blogging on the polls. Polls will appear in the state-by-state pages when we come across them.

  If we have any bias on this site, it is for a close election. We’re running an electoral college informational website. If the race isn’t close, people won’t need to visit the website. Neither of these candidates, if elected, is likely to displace anyone from Mt. Rushmore.

• 8 mark // Aug 11, 2008 at 8:39 am

  .I appreciate your post..and I do like the site…………..however…it seems that there are daily polls that come out.and 270towin is way behind the curve……..would it not be prudent to mention the respected daily polls?…and the mt. rushmore comment.excellent.most excellent…….

• 9 quincyconscience // Aug 11, 2008 at 2:28 pm

  http://www.fivethirtyeight.com/2008/08/persistent-myth-of-bradley-effect.html
  LoL i think they can hear you

• 10 mark // Aug 11, 2008 at 7:34 pm

  quincy…………..rassmussen and gallup are the 2 best…and they have national and state polls almost daily..especially the national…….with obama’s drastic meltdown in the numbers….it would seem that 270 might consider updating the site as to more accuratelt reflect current trends…fair enough??

• 11 quincyconscience // Aug 12, 2008 at 2:57 am

  What is up with your spelling errors, not that it makes any point more valid but you might try to use spell check. Rasmussen and Gallup really don’t update state polls all that often in the grand scheme of things, especially this early. For example Rasmussen yesterday (8/11) released polls in Iowa and Oregon. These will be used by 270 most likely. The problem with using only two pollsters is that neither of these states were touched by Rasmussen in a full month. Its not like they release 20 polls a day. What polls show obama loosing ground on a state level? The national polls show Obama slipping but no organization can point to the states he is loosing in. They could be states like Texas, Louisiana, Tennessee, etc in which case that would not really effect the odds of winning the election for either candidate.
  PS you do not give SurveyUSA the credit it deserves with its prime performance recently

  270toWin: The conversation on this forum needs to remain civil. My sense of this back-and-forth is that Mark is referring primarily to the national polls in his comments about daily updates, with the issue likely being (although I don’t want to speak for either of you) that if McCain is gaining (or if Obama’s lead were to increase), this would be reflected much more quickly in the national polls than the state-level ones. I suspect we’ll definitely see something like that when the candidates each pick their VP nominee. Any kind of bump might show up within a day or two in the national polls, but it could be several weeks before it ripples out into each state poll. On 270, we’re focused on the state-level polls since that is ultimately how the U.S. election works…. and the two will catch up with each other on election day…. but in the interim, the national polling will likely oscillate more than the state-level ones.

• 12 mark // Aug 12, 2008 at 12:01 pm

  spelling errors aside……….I do like survey polls……..Its not a matter of like or dislike…….I like 270towin very much…..The state polls in several have seen obama;s numbers fall ..his lead in states like iowa, minnesota, colorado , even michigan all have fallen by a more than statistical noise……..and my observation was that 270’s didnt reflect those quickly enough……..thats all………..sorry about the spelling errors………sincerely.mark……………………………..p.s. I didnt say I only used 2 polls.I merely pointed out they were two of the best………..

• 13 mark // Aug 12, 2008 at 12:12 pm

  dear staff at 270towin, Thank you for a really good website.I have looked at others but yours is superior.not perfect.but superior…….Im still trying to figure out the simulator and all its nuances……but I have told friends about the site……

  270toWin: Appreciate the feedback. The simulator is an interesting tool, but we’re still wondering about it ourselves. The statistics behind it are sound, but the outputs are only as good as the polls that go into it… and nobody will really know if those are right until Election Day. The two biggest defects we see so far are:

  1. It probably overstates the likelihood of an actual outcome around the ‘average’ because it treats each state as an independent event. There is some possibility that most or all swing states will fall to one candidate or the other on Election Day. For example, North Dakota is polling rather closely in the few polls that have come out, and so it comes up ‘blue’ in some simulations. However, if North Dakota actually were to go for Obama, most if not all of the other swing states will likely have also gone blue. The simulator doesn’t explicitly model for that ‘covariance’ (as our statistical friends call it), as it would require a lot of assumptions and a very complicated model.

  2. As mentioned in a prior comment, the state polls tend to be infrequent (once or twice a month) and so, in the aggregate, they likely lag overall popular opinion.

• 14 mark // Aug 13, 2008 at 11:16 am

  it seems to me that whats not being reported is obama’s falling poll numbers in states like…..pennsylvania..ohio..michigan…missouri…minnesota…….Obama is still leading………but the margin is way down……..an ominous sign for the democrat , who should ,by all accounts be leading by a much bigger number…..what do the numbers say?….obama was up by what?.8-10 percent…..id like your thoughts on the dropping obama numbers and why he isnt pulling away like the generic matchups suggest…..thanks

• 15 quincyconscience // Aug 13, 2008 at 5:20 pm

  I think not counting people with cell phones and not landlines as likely voters, standard practice for all polls mentioned, leads to deceptively good numbers for McCain as many poorer, younger Americans (think college students and recent post grads) are not counted. Also the Maverick factor of McCain helps him vs the generic match up. The main reason for the disadvantage is likely the ugly nomination process that took hold for democrats, severely hurting them vs the generic. A fourth factor in why obama is not pulling away is most likely due to McCain’s history as a war hero and honest advocate of things like environmental policy and campaign finance that are popular among groups of voters the generic republican misses completely with. Most of the close numbers are real, this is closer than democrat versus republican, and it will be fun to watch it all unfold.

• 16 mark // Aug 13, 2008 at 5:57 pm

  the ugly nomination you speak of.it was exactly my point…..the blue-collar white vote of ohio.pennsylvania went 70-30 for hillary clinton{see bitter and cling to guns} comments..another plausible reason is simple: the more voters get to know obama.the less they like him.or trust him.{fisa….campaign finance reform}….as for the cell phone matter…I think there is a good argument for that. However, obama’s numbers with the younger voters has been declining in several polls..as well as his numbers with white males….even mccain picked up 2 points of the african american vote…..this year could very well be the year of the landslide..the mccain landslide…..

• 17 Alex // Aug 13, 2008 at 8:30 pm

  I must say that I am disturbed by this exchange between mark and quincy.

  The reason this website works so well is because the people behind the scenes remain un-biased. They post the numbers when they get them and then feed them into the simulator. Granted the polls themselves may be biased, but the 270towin.com staff have remained resolutely unbiased in their reporting of the information.

  It is quite obvious that Mark is a McCain supporter, a die hard McCain supporter. I am also thinking that Quincy may be in the Obama camp. Both of you in your comments are trying to play up your candidate as the best option.

  The way I have always seen this website since first finding it in 2004, is that this place was for discussing the methods and data behind every national election. What makes this website great in my opinion, is that you do not have people trying to spin the facts about their candidate being the best option.

  I sincerely hope Mark and Quincy will refrain in the future from trying to convince other readers and bloggers on this website to vote for their candidate.

  Remain true to the spirit of this website guys and just go by the facts we have here and remain unbiased in your postings.

• 18 quincyconscience // Aug 13, 2008 at 11:22 pm

  I am first and foremost am an observer of this election. I did mildly support Obama before the whole public finance deal, and now I would be just as pleased if McCain won, but might vote for Obama in the end. I am trying to defend the polls as accurate and that is all. I am not trying to defend Obama, he asked what I thought about the generic numbers versus actual so I explained. Polling has favored Obama, so I believe he is going to win. I have not mentioned an issue positively or negatively except in regards to polling. If the facts on the ground favor a candidate, it is not biased to say that candidate has an advantage. I have tried to remain unbiased as much as possible, but asking someone to cite polls they are requesting to be added is not biased or partisan. I specifically mentioned 3 weaknesses of Obama in my last post without any response to them. I am not even sure you would have ventured the guess at my support were I not on the other side of Mark in this debate. With that said I will try to make sure that my posts are unbiased in the future. I understand that the spelling error comment was biased but it bothered me so I said one line about it.
  If you have any specific complaints of bias feel free to point them out, I won’t be offended and it will be useful to know when I am committing such errors.

  270toWin: Our site strives to be unbiased. I know we’re pretty successful because we get emails from both candidates’ supporters claiming we’re favoring the other guy. However, it would be unreasonable (and maybe kind of boring) to expect everyone who visits and posts on the site not to have a preference for one candidate or another. So, feel free to express your opinions — just try not to bash others for theirs. The polarization that is out there today (when is the Olbermann vs. Hannity boxing match going to take place?) is not really helpful for the country as a whole, and while that genie is never getting back in the bottle, we’d just like to try and stay above the fray, to the extent possible, on 270toWin.

  By the way, coming soon (if we can work out a few programmatic kinks), visitors to 270 will be able to email a map of their projections to others.

• 19 mark // Aug 14, 2008 at 7:36 am

  alex…..I never accused 270 of being biased……….I was only concerned about the outdated polls cited by their website……….I never tried to convince the bloggers to do anything.except update their polls…….as for who im supporting…its irrelevant…..like I stated earlier.I like 270towin and will pass the word on the site.thats as fair as one could possibly be………sincerely.mark

• 20 mark // Aug 14, 2008 at 7:43 am

  quincy……….that was a well-written response to alex……………I appreciate lively debate…thrust and perry.give and take……….just one question……..What do you think the worst poll is/…as far as being biased…one way or the other…….im curious……….sincerely.mark

• 21 quincyconscience // Aug 14, 2008 at 11:33 am

  Obviously PPP and Strategic Vision can have trouble remaining neutral due to their respective party allegiances. Also online polls like Zogby interactive clearly trend obama (shocker: Obama’s supporters are younger!!!!11!!one!). Most polls have a real MoE of about 7-10 historically. Unfortunately for close elections, when a poll shows a one or two point difference it is least accurate, as has consistently been the case in Virginia. I take little stock in the idea that anybody really gets Virginia with its surging population demographics. The Bradley effect seems to be on the ground, alive and well in this particular state (I live in Northern Virginia most of the year). This may be countered by the fact that I literally can’t go to a single food store within a mile of my residence and expect the cashier to know fluent English. Again the young voter thing is also alive with the large amount of universities, but many of them are slightly more conservative than the average institution such as GMU and to a lesser extent Virginia Tech. Nobody knows whats going on in Virginia for the presidential race, so for that reason polls of Virginia, regardless of company are the least likely to be accurate given the large amount of factors counting against accurate polling numbers.

• 22 mark // Aug 14, 2008 at 1:13 pm

  quincy………interesting hypothesis…….in regards too virginia….Im gonna play the historical odds…………virginia votes republican in presidential elections…….and with obama a far left candidate…..will do so again……..and for more anecdotal evidence: the “young” vote doesnt turn out like the older and military..especially in va….84% of under 24 voters that are registered didnt actually vote in 2004……..anyway.nice post…sincerely.mark

• 23 quincyconscience // Aug 15, 2008 at 2:52 am

  I could talk about new registration figures and primary voting figures all day, but what I find most interesting is that this state is polling within 5 points consistently. There are over half a million registered voters in the 18-25 age range in Virginia (http://www.dailypress.com/news/local/virginia/dp-va–youngvoters0721jul21,0,6890764.story)
  This number can’t just be ignored, no matter who they end up breaking towards. Obama obviously is playing the ground game in virgina which can aid turnout. Now this certainly doesn’t mean he has any secret advantage in the state as a whole, as many factors could go against him as well, such as the possibility of a perceived snub of Kaine and his position of offshore drilling which is widely viewed as popular in Virginia just like most of the rest of the country. It is simply a factor that makes polling in Virginia particularly hard to predict this election

• 24 mark // Aug 15, 2008 at 12:35 pm

  I cant argue with your latest post………..darn…….lol……..I would say that kaine isnt obama’s “magic bullet”……..polls ive seen have kaines approval rating under 50%……car tax issues…….along with drilling….seem to doom kaine as his running mate……along with kaines silly comment on “russia listened to obama’s directions” when its clear that mccain has a much better understanding of geo-politics………just my opinion…..

  270toWin: Doesn’t seem like Russia is listening to anyone right now. Foreign policy questions will not be limited to just Iraq in the debates this fall.

• 25 mark // Aug 15, 2008 at 1:55 pm

  dear 270towin staff……….thats a great point……….makes kaine look even sillier……p.s….gallup has it tied today….Im beginning to think that the state polls are coming in line with the nationals…..your earlier analysis was right on point…………thank you

• 26 quincyconscience // Aug 16, 2008 at 4:11 am

  Kaine also adds to the inexperienced claim. Kaine is the most talked about non-clinton who is not going to be VP in my opinion. This VP situation seems like it could become volatile if it is played wrongly. I think with context the Kaine did not mean that Russia was literally listening to Obama, but meant it as a kind of people do what he says because its the best path. Another example of this could be someone saying Clinton starting listening to Obama on Iraq, when she was just changing her mind because of new information. Politicians consistently use this tactic to show how their candidate was right first. McCain is certainly no stranger to this particular technique

• 27 mark // Aug 18, 2008 at 8:19 am

  sorry quincy, that doesnt pass the smell test………..back in march, mccain spoke on the issue of russia and georgia……..he made it clear , his thoughts on putin and the big bear….kaine didnt have a ..”context”…….he said it.he meant it…….sometimes a rose is just a rose.no context needed.just my opinion.

• 28 Richard // Sep 8, 2008 at 12:22 pm

  Bradley effect? Perhaps with southern white dinasours born before baby boom. How about the Hollywood effect? Haven’t the younger generations been exposed to see African American presidents in movies?

Leave a Comment