# The weirdness of polls... What's up? **URL:** \ **Category:** Politics & Elections\ **Created:** [October 17, 2012, 5:39pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033 "2012-10-17T17:39:08Z")\ **Posts on this page:** 20\ **Page:** 4
**Author:** ![Omg\_a\_Black\_Conservative](https://avatars.discourse-cdn.com/v4/letter/o/7ba0ec/32.png) [@Omg\_a\_Black\_Conservative](https://boards.straightdope.com/u/Omg_a_Black_Conservative)\ **Post date:** [October 19, 2012, 2:46am UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/61 "2012-10-19T02:46:10Z")
In most any national poll you click on, you’ll find the following things are almost always true: 1.) Romney is leading among Independents. 2.) The Republican coalition is more united than the Democrat coalition (meaning that Romney is doing a better job at holding down his base than is Obama) 3.) Republicans are more enthusiastic than Democrats. 4.) Obama’s support is down across most every, if not every, demographic. Yet, despite all of this, a lot of those polls still have Obama leading. It really boils down to the D/R/I sample of any given poll, which “we” conservatives feel have unjustifiably high Democratic turnout relative to Republican and Independent turnout. So while Gallup is probably overstating Romney’s performance, there’s little doubt in my mind that the other polls are understating Romney’s performance, which is why whining about the Gallup poll-- and only Gallup-- doesn’t make much sense to me. ETA\> [Case and point](http://www.publicpolicypolling.com/pdf/2011/PPP_Release_TrackingPoll_1018.pdf). PPP has it tied 48 - 48 even with terrible internals for Obama (sample is 40D/35R/24I). ---
**Author:** ![Chronos](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/chronos/32/134_2.png) [@Chronos](https://boards.straightdope.com/u/Chronos)\ **Post date:** [October 19, 2012, 3:06am UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/62 "2012-10-19T03:06:32Z")
The reason that Romney is leading among independents is the same reason that it looks to you like there are too many Democrats vs. Republicans and independents. “Independent” does _not_ mean the same thing as “moderate”, and a large chunk of people nowadays who call themselves “independent” are people who would previously have called themselves “Republican”, but who now think the Republicans aren’t conservative enough. So you get a bunch of extreme righties splintering off from the Republicans, which means that the number of remaining Republicans shrinks, and Romney gets a big advantage among people who call themselves independent. ---
**Author:** ![Omg\_a\_Black\_Conservative](https://avatars.discourse-cdn.com/v4/letter/o/7ba0ec/32.png) [@Omg\_a\_Black\_Conservative](https://boards.straightdope.com/u/Omg_a_Black_Conservative)\ **Post date:** [October 19, 2012, 3:40am UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/63 "2012-10-19T03:40:28Z")
Your rhetoric doesn’t hold up to reality, least because the number of Independents have been growing mostly at the expense of _Democrats_, not Republicans. This is where that state level data that you guys constantly ignores comes into play. [This](http://images.politico.com/global/2012/08/third_way_memo_-_the_i_s_have_it_pre-convention_voter_registration_analysis.html) is a report looking at party affiliation in eight “swing” states. > [@](#): > > Since 2008, the number of Independents in the eight presidential battleground states has increased substantially. By contrast, party registration is down, with Democratic registration declining significantly and Republican registration doing so by a lesser margin. In total, since 2008 in these 8 states: > > • Democratic registration has declined by 800,329, or 5.2%, and now stands at 14,624,665; > • Republican registration has declined by 78,985, or 0.7%, and now stands at 12,114,524; and, > • Independent registration has increased by 486,677, or 6.4%, and now stands at 8,083,829. > > In 6 of the 8 states—Colorado, Florida, Nevada, New Mexico, North Carolina, and Pennsylvania—Independent registration has outpaced both Democratic and Republican registration since 2008. In the remaining two, Republicans have had the edge—likely the result of the competitive GOP primary. In Iowa, Republican registration increased by 6.4% while Democratic and, to a lesser extent, Independent registration fell. In New Hampshire, voter registration has declined across the board; however, Republicans have shed the fewest voters. I’d be willing to bet that the national trend bears out the same. I’d like to know where you guys are getting your informationfrom. As it is, my original point still stands. Obama won Independents by 8 points in 2008. This year, he may very well lose them by 4 - 6 points, which seems very likely. That’s a 12 point swing (if you look at polls of the aforementioned states, Romney is winning Independents in most, if not all, of them). And if it’s true that Independents have grown at mostly the expense of Democrats, then that effective swing is even larger, because it means that Romney would be picking up a great deal of Obama “defectors” who called themselves Democrats in the last election. ---
**Author:** ![Lantern](https://avatars.discourse-cdn.com/v4/letter/l/d26b3c/32.png) [@Lantern](https://boards.straightdope.com/u/Lantern)\ **Post date:** [October 19, 2012, 4:06am UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/64 "2012-10-19T04:06:17Z")
Nate has a long [post](http://fivethirtyeight.blogs.nytimes.com/2012/10/18/gallup-vs-the-world/?)about the Gallup poll which is well worth reading. As he shows Gallup has a track record of being wrong when it is different from the consensus as well as being very volatile in the middle of election campaigns. Read the whole thing but here is his conclusion: > [@](#): > > To be clear, I would not recommend that you literally just disregard the Gallup poll. You should consider it — but consider it in context. > > The context is that its most recent results differ substantially from the dozens of other state and national polls about the campaign. It’s much more likely that Gallup is wrong and everyone else is right than the other way around. ---
**Author:** ![Leaper](https://avatars.discourse-cdn.com/v4/letter/l/4bbf92/32.png) [@Leaper](https://boards.straightdope.com/u/Leaper)\ **Post date:** [October 19, 2012, 4:10am UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/65 "2012-10-19T04:10:44Z")
Psst… 55. 🙂 ETA: A quick Google search seems to indicate that **OMG’s** assertions are common in conservative blogging circles, which may be where he first read of it. I’d be interested in seeing analysis of it, if only for that reason. ---
**Author:** ![Lantern](https://avatars.discourse-cdn.com/v4/letter/l/d26b3c/32.png) [@Lantern](https://boards.straightdope.com/u/Lantern)\ **Post date:** [October 19, 2012, 4:26am UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/66 "2012-10-19T04:26:19Z")
> [@Leaper](#): > > Psst… 55. 🙂 Ah sorry I missed that. To make up I went back and dug up Nate’s [post](http://fivethirtyeight.blogs.nytimes.com/2012/09/29/poll-averages-have-no-history-of-consistent-partisan-bias/?)about “over-sampling” from a few weeks back since the same nonsense is turning up in this thread. Here is the crux of the issue: > [@](#): > > Some have said that the polls are “oversampling” Democrats and producing results that are biased in Mr. Obama’s favor. One Web site, [unskewedpolls.com](http://unskewedpolls.com), contends that even Fox News is part of the racket in what it says is a “trend of skewed polls that oversample Democratic voters to produce results favorable for the president.” > > The criticisms are largely unsound, especially when couched in terms like “oversampling,” which implies that pollsters are deliberately rigging their samples. > > But pollsters, at least if they are following the industry’s standard guidelines, do not choose how many Democrats, Republicans or independent voters to put into their samples — any more than they choose the number of voters for Mr. Obama or Mitt Romney. Instead, this is determined by the responses of the voters that they reach after calling random numbers from telephone directories or registered voter lists. > > Pollsters will re-weight their numbers if the demographics of their sample diverge from Census Bureau data. For instance, it is typically more challenging to get younger voters on the phone, so most pollsters weight their samples by age to remedy this problem. > > But party identification is not a hard-and-fast demographic characteristic like race, age or gender. Instead, it can change in reaction to news and political events from the party conventions to the Sept. 11 attacks. Since changes in public opinion are precisely what polls are trying to measure, it would defeat the purpose of conducting a survey if pollsters insisted that they knew what it was ahead of time. ---
**Author:** ![Leaper](https://avatars.discourse-cdn.com/v4/letter/l/4bbf92/32.png) [@Leaper](https://boards.straightdope.com/u/Leaper)\ **Post date:** [October 19, 2012, 4:31am UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/67 "2012-10-19T04:31:26Z")
Ah, I see. I’d read that before, but didn’t connect it to this issue. So Nate would say that it’s less useful than it seems to say “Obama is down X points among independents than in 2008” because the pool of people who were independents in 2008 is not necessarily/probably not the same as they are now? ---
**Author:** ![Lantern](https://avatars.discourse-cdn.com/v4/letter/l/d26b3c/32.png) [@Lantern](https://boards.straightdope.com/u/Lantern)\ **Post date:** [October 19, 2012, 5:11am UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/68 "2012-10-19T05:11:54Z")
> [@Sam\_Stone](#): > > However, all the trends are moving towards Romney - even after the last debate. Actually Obama has had a great day in the 538 forecast moving up by almost 5 points to 70% based on strong state polls in Wisconsin, Iowa and Colorado which do include polling after the second debate. ---
**Author:** ![Omg\_a\_Black\_Conservative](https://avatars.discourse-cdn.com/v4/letter/o/7ba0ec/32.png) [@Omg\_a\_Black\_Conservative](https://boards.straightdope.com/u/Omg_a_Black_Conservative)\ **Post date:** [October 19, 2012, 5:58am UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/69 "2012-10-19T05:58:05Z")
> [@Leaper](#): > > Ah, I see. I’d read that before, but didn’t connect it to this issue. > > So Nate would say that it’s less useful than it seems to say “Obama is down X points among independents than in 2008” because the pool of people who were independents in 2008 is not necessarily/probably not the same as they are now? Only a fool would make an assertion like that, since that would not only apply to Independents, but to Democrats and Republicans as well, which would make a poll utterly worthless. Anyway, I said this a few weeks back, but the prevalent thought seems to be that “party ID is fluid, therefore whatever partisan split a poll finds must be accurate.” But this leads to two separate questions; if party ID is fluid, and one day a poll found that, say, 80% of the population identified as Republicans and 20% as Democrats, would the results be considered valid? Chances are it would not be, which leads to the secondary question of at what point does a partisan split become _unreasonable_? ---
**Author:** ![Leaper](https://avatars.discourse-cdn.com/v4/letter/l/4bbf92/32.png) [@Leaper](https://boards.straightdope.com/u/Leaper)\ **Post date:** [October 19, 2012, 4:41pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/70 "2012-10-19T16:41:55Z")
I meant in the context of comparing 2008 to 2012 directly. But then, I don’t know what **Lantern’s** point really was; that’s why I was asking, rather than stating. Anyway, you’ve already declared toplines to be utterly worthless. So why do pollsters use them at all? Since the very beginning (from what I can tell), it’s always been the topline that’s been used. So have they been worthless all this time? What’s the mechanism that turns it from useful internals to worthless toplines? ---
**Author:** ![Fotheringay-Phipps](https://avatars.discourse-cdn.com/v4/letter/f/d9b06d/32.png) [@Fotheringay-Phipps](https://boards.straightdope.com/u/Fotheringay-Phipps)\ **Post date:** [October 19, 2012, 5:06pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/71 "2012-10-19T17:06:25Z")
> [@Leaper](#): > > It may or may not help, but Nate Silver [has a post up discussing Gallup’s poll in particular.](http://fivethirtyeight.blogs.nytimes.com/2012/10/18/gallup-vs-the-world/) > > His summation at the end comes down to: don’t ignore Gallup’s result entirely. Take it in the whole universe of current poll results, compare and contrast, and then decide if Gallup is right and everyone else is wrong, or the reverse. It seems evident which side he comes down on. Silver is being misleading in that article, in comparing Gallup to “the average”. This gives the false impression that it’s Gallup against “everyone else”, as you put it. The reality is that “everyone else” consists of varying polls which cover a broad spectrum of results. Most individual polls will not have the same result as the average of all other polls, aka “everyone else”. And you could apply the same logic that Silver applies to Gallup to any other poll. On the whole, the average of other polls is likely to be more accurate than any individual poll, which is why outfits like RCP (& 538 FTM) include a spectrum of polls. There’s no reason to assume that Gallup is right and everyone else wrong, but the same applies to any other poll that you might happen to be discussing. The idea that Gallup is some sort of unique outlier has not been established by anything in that article. ---
**Author:** ![Richard\_Parker](https://avatars.discourse-cdn.com/v4/letter/r/35a633/32.png) [@Richard\_Parker](https://boards.straightdope.com/u/Richard_Parker)\ **Post date:** [October 19, 2012, 5:18pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/72 "2012-10-19T17:18:42Z")
Outlier is not a mathematically precise term. It is a judgment about the distance of a given poll from the mean. Gallup is currently furthest by a substantial margin–twice as far as the next furthest–which is what most people mean when they label a poll an outlier. ---
**Author:** ![Fotheringay-Phipps](https://avatars.discourse-cdn.com/v4/letter/f/d9b06d/32.png) [@Fotheringay-Phipps](https://boards.straightdope.com/u/Fotheringay-Phipps)\ **Post date:** [October 19, 2012, 5:25pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/73 "2012-10-19T17:25:37Z")
I agree they’re an outlier at this time. But he’s making a history of it. ---
**Author:** ![Lantern](https://avatars.discourse-cdn.com/v4/letter/l/d26b3c/32.png) [@Lantern](https://boards.straightdope.com/u/Lantern)\ **Post date:** [October 19, 2012, 5:28pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/74 "2012-10-19T17:28:23Z")
> [@Leaper](#): > > I meant in the context of comparing 2008 to 2012 directly. But then, I don’t know what **Lantern’s** point really was; that’s why I was asking, rather than stating. I thought it was a rhetorical question. Anyway Nate hit the nail on its head with this para which I quoted: > [@](#): > > But party identification is not a hard-and-fast demographic characteristic like race, age or gender. Instead, it can change in reaction to news and political events from the party conventions to the Sept. 11 attacks. Since changes in public opinion are precisely what polls are trying to measure, it would defeat the purpose of conducting a survey if pollsters insisted that they knew what it was ahead of time. So the point is party ID changes over time. It’s the polls which tell you what a reasonable party ID is not some a priori assumption of what it should be. ---
**Author:** ![Leaper](https://avatars.discourse-cdn.com/v4/letter/l/4bbf92/32.png) [@Leaper](https://boards.straightdope.com/u/Leaper)\ **Post date:** [October 19, 2012, 5:29pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/75 "2012-10-19T17:29:56Z")
Well, my question was, how does that quote relate, exactly, to the internals that **OMG** is quoting, and the significance he attaches to them? I can _sort of_ see how, but I’m not 100% sure. ---
**Author:** ![Richard\_Parker](https://avatars.discourse-cdn.com/v4/letter/r/35a633/32.png) [@Richard\_Parker](https://boards.straightdope.com/u/Richard_Parker)\ **Post date:** [October 19, 2012, 5:32pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/76 "2012-10-19T17:32:24Z")
> [@Fotheringay-Phipps](#): > > I agree they’re an outlier at this time. But he’s making a history of it. Huh? His claim is “Gallup Performs Poorly When Out of Consensus,” and that “The context is that its _most recent results_ differ substantially from the dozens of other state and national polls about the campaign. It’s much more likely that Gallup is wrong and everyone else is right than the other way around.” (emphasis added.) You seemed to be disputing that, but now you concede that they’re an outlier. So your complaint is that he’s wrong to have identified previous times when they were also an outlier? ---
**Author:** ![Lantern](https://avatars.discourse-cdn.com/v4/letter/l/d26b3c/32.png) [@Lantern](https://boards.straightdope.com/u/Lantern)\ **Post date:** [October 19, 2012, 6:43pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/77 "2012-10-19T18:43:52Z")
> [@Leaper](#): > > Well, my question was, how does that quote relate, exactly, to the internals that **OMG** is quoting, and the significance he attaches to them? I can _sort of_ see how, but I’m not 100% sure. Are you talking about the internals for party registration. By my calculation it’s about 42/35/23 for D/R/I. Is that so different from what the cross-tabs in the polls are saying? ---
**Author:** ![Leaper](https://avatars.discourse-cdn.com/v4/letter/l/4bbf92/32.png) [@Leaper](https://boards.straightdope.com/u/Leaper)\ **Post date:** [October 19, 2012, 6:49pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/78 "2012-10-19T18:49:49Z")
I’m talking about post 59. ---
**Author:** ![Lantern](https://avatars.discourse-cdn.com/v4/letter/l/d26b3c/32.png) [@Lantern](https://boards.straightdope.com/u/Lantern)\ **Post date:** [October 19, 2012, 7:00pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/79 "2012-10-19T19:00:06Z")
Well that is a national poll. It doesn’t say what the party ID numbers are so it’s hard to say whether they are off the mark. The main point is that you can only know what sensible party ID numbers are from a bunch of polls. If the average of a large number of polls say that Democrats outnumber Republicans than that is a fact. You could criticize a single poll with a very different party ID but you can’t claim on a priori grounds that party ID should be a particular number if a large number of polls say otherwise. ---
**Author:** ![Leaper](https://avatars.discourse-cdn.com/v4/letter/l/4bbf92/32.png) [@Leaper](https://boards.straightdope.com/u/Leaper)\ **Post date:** [October 19, 2012, 7:05pm UTC](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033/80 "2012-10-19T19:05:00Z")
Okay, here’s what I’m not getting (and this is probably pre-lunch famish syndrome, so please bear with my stupid questions): what, specifically, does any of that have to do with **OMG’s** assertions (in post 58 and 61) that Romney is winning independents and not lost any of McCain’s '08 support, that Obama has lost a lot more of his '08 support, and that given that, the topline can’t possibly be right because that means Romney should be way ahead? [Previous page](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033.md?page=3) [Next page](https://boards.straightdope.com/t/the-weirdness-of-polls-whats-up/638033.md?page=5)