Skip to main content

Mitt Romney listens to questions in Jacksonville, FL
Smile, Mitt! CNN says you are tied in Florida and Virginia!

A couple of nights ago, during the nightly Daily Kos Elections Polling Wrap, the topic of discussion was whether or not the press was deliberately suppressing the breadth of the Obama post-DNC "bounce" in order to continue the appearance of a tossup race that could go either way. At the time I wrote that piece, I actively wondered if I was perhaps donning a tinfoil hat.

CNN's coverage of yesterday's trio of NBC/Marist polls seems only to underscore that point.

For those who missed it, the NBC/Marist polls found the president staked to 5-point leads among likely voters in Florida and Virginia, and a larger 7-point advantage in Ohio. With registered voters, the margins were spread out to a slightly larger spread of 7-9 points.

CNN's Gregory Wallace, writing up the polls on the media outlet's website, categorized Florida and Virginia as a tie. A tie.

This is a classic case of being technically correct and yet committing journalistic malpractice, nevertheless. Check out the characterization of those polls:

HEADLINE: Ohio trending for Obama; Florida, Virginia tied

President Barack Obama has a narrow advantage over Republican presidential nominee Mitt Romney in Ohio according to a new poll released Thursday which also found the race for the White House locked up in Florida and Virginia.


In Thursday's polls, Ohio likely voters split 50% for Obama, 43% Romney, according NBC/Wall Street Journal poll, with 83% strongly supporting their candidate of choice.

Voters in both Florida and Virginia split 49% for Obama and 44% for Romney, a five point difference within the poll's sampling error of 3.1.

This is one of those classic examples of a journo knowing precisely jackshit about the concept of margin of error. Wallace, in this case, dubs the race a tie because a Romney lead could, in theory, fall within the margin of error. Which it could, if the actual result was at the absolute low end of Obama's MoE, and at the absolute upper end of the Romney MoE.

However, what Wallace conveniently ignores, presumably because he either did not know or did not care, is that the converse is equally plausible. Indeed, if you want to make the parsing argument that Wallace was not technically wrong to declare the race a "tie," you could therefore follow that it was equally accurate to write the header "Obama up double digits in Florida and Virginia". After all, with an MoE of 3.1 percent, an 11-point Obama lead in both states is just as plausible as an infinitesimal Romney advantage.

The Associated Press actually addresses how to deal with polling margins of error, and they are quite explicit about how to characterize races such as these (emphasis mine):

If the difference between the candidates is more than twice the sampling error margin, then the poll says one candidate is leading.

If the difference is less than the sampling error margin, the poll says that the race is close, that the candidates are "about even." (Do not use the term "statistical dead heat," which is inaccurate if there is any difference between the candidates; if the poll finds the candidates are tied, say they're tied.)

If the difference is at least equal to the sampling error but no more than twice the sampling error, then one candidate can be said to be "apparently leading" or "slightly ahead" in the race.

Double the sampling error, in this case, is 6.2. By the AP standards, Obama is "slightly ahead," at worst. But he is closer to having an unambiguous lead than he is to being what the AP would characterize as a "close race" or "about even."

The bottom line is while there is a statistical chance that sampling error could account for the entirety of Obama's lead in the polls in Florida and Virginia, it is a very small chance, one that falls in just a small fraction of the range of possible outcomes. And to characterize the race as "tied," either in the headline or the text, falls damned close to willful misrepresentation of the state of the race.


Your Email has been sent.
You must add at least one tag to this diary before publishing it.

Add keywords that describe this diary. Separate multiple keywords with commas.
Tagging tips - Search For Tags - Browse For Tags


More Tagging tips:

A tag is a way to search for this diary. If someone is searching for "Barack Obama," is this a diary they'd be trying to find?

Use a person's full name, without any title. Senator Obama may become President Obama, and Michelle Obama might run for office.

If your diary covers an election or elected official, use election tags, which are generally the state abbreviation followed by the office. CA-01 is the first district House seat. CA-Sen covers both senate races. NY-GOV covers the New York governor's race.

Tags do not compound: that is, "education reform" is a completely different tag from "education". A tag like "reform" alone is probably not meaningful.

Consider if one or more of these tags fits your diary: Civil Rights, Community, Congress, Culture, Economy, Education, Elections, Energy, Environment, Health Care, International, Labor, Law, Media, Meta, National Security, Science, Transportation, or White House. If your diary is specific to a state, consider adding the state (California, Texas, etc). Keep in mind, though, that there are many wonderful and important diaries that don't fit in any of these tags. Don't worry if yours doesn't.

You can add a private note to this diary when hotlisting it:
Are you sure you want to remove this diary from your hotlist?
Are you sure you want to remove your recommendation? You can only recommend a diary once, so you will not be able to re-recommend it afterwards.
Rescue this diary, and add a note:
Are you sure you want to remove this diary from Rescue?
Choose where to republish this diary. The diary will be added to the queue for that group. Publish it from the queue to make it appear.

You must be a member of a group to use this feature.

Add a quick update to your diary without changing the diary itself:
Are you sure you want to remove this diary?
(The diary will be removed from the site and returned to your drafts for further editing.)
(The diary will be removed.)
Are you sure you want to save these changes to the published diary?

Comment Preferences

  •  CNN: Gotta keep people watching the "horse race." (15+ / 0-)

    "A young man who wishes to remain a sound Atheist cannot be too careful of his reading. There are traps everywhere ". C. S. Lewis

    by TofG on Fri Sep 14, 2012 at 10:19:18 AM PDT

    •  Is it possible? (0+ / 0-)

      That CNN and other media outlets are commissioning polls and surveys with small pools of subjects specifically to increase the margin of error? I would to see if someone has checked MoE between battleground states and states that are solidly red or blue.

    •  Sheesh. CNot the NEWS is right! They had to (0+ / 0-)

      do this from incompetence or desire to have it be close or "tied."

      Great explanation here! The real results:

      Obama is "apparently leading" or "slightly ahead" in the race.

      "extreme concentration of income is incompatible with real democracy.... the truth is that the whole nature of our society is at stake." Paul Krugman

      by Gorette on Fri Sep 14, 2012 at 10:33:33 AM PDT

      [ Parent ]

    •  CNN reports on what the horses leave on the track (1+ / 0-)
      Recommended by:

      The choice of our lifetime: Mitt Romney, It Takes A Pillage or President Barack Obama, Forward Together.

      by FiredUpInCA on Fri Sep 14, 2012 at 11:33:14 AM PDT

      [ Parent ]

  •  Run like we are 10 points down. (8+ / 0-)

    Keep calling.

    Keep sharing videos.

    Keep knocking on doors.

    Keep helping people to vote early.

    Increase our margin until it is beyond the margin of error.

    Run.  Don't coast.

    "When I use a word," Humpty Dumpty (Mittens, Paul Ryan, Scotty Walker, Limbaugh, pick your favorite) said in rather a scornful tone, "it means just what I choose it to mean -- neither more nor less."

    by Eman on Fri Sep 14, 2012 at 10:19:38 AM PDT

    •  I hate this meme... (11+ / 0-)

      Play like you are going to crush your opponents is much better.

      Why is Mitt Rmoney so happy that an American embassy was attacked? Why does he hate America?

      by RichM on Fri Sep 14, 2012 at 10:34:42 AM PDT

      [ Parent ]

      •  THIS (3+ / 0-)
        Recommended by:
        FiredUpInCA, Eman, MarcKyle64

        Can't stand it either. I'll GOTV even if Obama is LEADING BY 10 POINTS. Sheesh; dumping a bucket of ice on dancing people doesn't inspire me, it pisses me off.

        Romney = Nixon without the sweating problem

        by YoungArizonaLiberal on Fri Sep 14, 2012 at 10:58:02 AM PDT

        [ Parent ]

      •  I can live with it (0+ / 0-)

        if it means big ticket donors keep throwing money at Romney instead of switching downticket. As much as I dislike the idea of one-party control of gov't we can't afford not to have it at this point. We don't need validation from the MSM.

      •  The problem with propaganda that expresses... (3+ / 0-)

        ... tremendous confidence is that it might work.  

        When the Soviet Union invaded Finland in 1939, many of the Red Army soldiers didn't even bring proper winter clothing.  After all, according to their leadership, they were going to be back home in two weeks!  Instead, they brought brass bands, as well as flowers and candies for all the Finnish girls who were going to welcome them as "liberators" and lay down for them...   You see, the Red Army was many times more powerful than the Finnish Army.  It was going to be a rout.  It could last six days, six weeks. I doubt six months....

        Instead, of marching bands and girls fighting over who would get to take one of the "liberating" red army soldiers home, they ran into the Mannerheim Line.  Some estimates place Soviet casualties at 1,000,000 over a three-month period.  

        The Finns' major problem was that there were more Russians than they had bullets to shoot them with.  The Soviets' major problem was that they believed their own propaganda.  The Soviet Army was so powerful and mighty, and the outcome of the war was such a foregone conclusion, that only the Finns decided it was worth it to show ready to fight.

        There have been plenty of instances where people leading in the polls 60/40 have lost on election day due to voter complacency.  Every. Vote. Matters. No matter how far ahead our guy is.    

        I once had the unfortunate opportunity to live in a very conservative mid-sized town for a couple years (it was a suburb of the city my College was in, and I got a great deal on a really nice apartment).  The place had been solid Republican territory since the end of WW2.  Repubs historically won elections by the 60-70% range.  It was so deep-red, that for many years, quite a few candidates ran unopposed by the Dems.  The local Republican party essentially appointed whomever they wanted...    

        Eventually though, a string of overwhelming victories made their base complacent.  "Our guy always wins" people started to think, and they stopped going to the polls.  It just didn't matter.  

        One particular Wednesday in November of 1992, this deep red, conservative suburban enclave woke up to discover they suddenly had a democratic mayor, city council, and school board! And worse, it was made up of elitist, liberal college professors who ran never thinking they'd actually win.  Apparently, The Republican party was so powerful and mighty, and the outcome of the election was such a foregone conclusion, that only the Democrats decided it was worth it to show ready to vote.

    •  Right, we can't afford to relax. (0+ / 0-)

      We have to get everyone we can to vote in order to counteract the shenanigans the republicans are up to in trying to supress the vote.  Also, if we want to get anything done in the next four years, we've gotta give Pres. Obama people to work with.

      Gotta tune out the noise and keep our nose to the grindstone.

    •  So you (5+ / 0-)

      worked just as hard in 2010 as you are this year? Because we were actually running 10 points down. And as far as I saw, no one felt like busting their asses for the losing side.

      Seriously, there's nothing worse than pretending we're losing. I hope you never played sports with that attitude, because the best way to take an opponent out of the game is to wipe them the fuck out.

      •  Sorry my point wasn't clear. (0+ / 0-)

        Not pretending we are losing.  Never.  But playing as hard as if we had to catch up rather than coast.

        In sports I was fortunate enough to learn from a basketball great named Dolph Stanley.  Even when you have a 20 point lead in the 4th quarter, you never take your starters out until the other guy does first.

        In football, when you are ahead 28 to 0 you do put in the 2nd and 3rd strings in the 2nd half... but if the other team gets into the red zone with their starters you put the starting defense back in to protect their shutout.

        In my view of politics, you don't just try to retain votes when you have the lead.  You work every hour of every day to GAIN votes and expand the lead.

        As for 2010, had Democrats run on the President's record rather than away from it and confronted the Teabaggers head-on, perhaps it would have been different.  But that is hindsight.

        "When I use a word," Humpty Dumpty (Mittens, Paul Ryan, Scotty Walker, Limbaugh, pick your favorite) said in rather a scornful tone, "it means just what I choose it to mean -- neither more nor less."

        by Eman on Fri Sep 14, 2012 at 01:20:43 PM PDT

        [ Parent ]

  •  It is in their financial interest (3+ / 0-)
    Recommended by:
    Remembering Jello, MJB, bear83

    News organizations profit from ratings/views/clicks.

    These will be higher if the race is a tie.

    It is the news executives' ethical obligation (to their shareholders) to maximize profits.

    Therefore it is their ethical obligation to lie and spin in order to keep the appearance that the race is close.  

    Just another example of how the free market being king of everything doesn't always work out....

  •  What is their incentive? (1+ / 0-)
    Recommended by:

    Why would they do this?

    Are they inept?  

    I know we always say that they need this race to look "neck and neck" and "down to the wire".....

    ...but why?

    Please, somebody tell me what they have to gain by doing this.

    Aldus Shrugged : The Antidote to Ayn Rand. Buy ALDUS SHRUGGED on amazon, and ALL royalties will be donated directly to HELP ME TO HELP THE BIG O!!! And follow the fun: @floydbluealdus1

    by Floyd Blue on Fri Sep 14, 2012 at 10:20:30 AM PDT

  •  Oh, and don't forget a 7% difference being a (6+ / 0-)

    "narrow lead."

    "A young man who wishes to remain a sound Atheist cannot be too careful of his reading. There are traps everywhere ". C. S. Lewis

    by TofG on Fri Sep 14, 2012 at 10:21:19 AM PDT

  •  Post it (5+ / 0-)

    The AP guidelines you quote here should be sent directly to Greg Wallace and posted in the commentary on their site.

    •  and the "twice the MOE" (1+ / 0-)
      Recommended by:
      Blue in VA

      is simply a style recommendation, there is not any statistical significance to that whatsoever.

      And I'm not even convinced, unless I saw the calcs and raw data, that any particular MOE is accurate to a tenth of a percentage point either (significant fiqures and all that). Poor questions, data not normally distributed, too small sample size all influence the MOE's interpretation. It seems all distribution curves are thought of as bell-shaped, but that is often not the case.

      Without geometry, life is pointless. And blues harmonica players suck.

      by blindcynic on Fri Sep 14, 2012 at 10:34:54 AM PDT

      [ Parent ]

  •  I NEVER watch that shi...y network. NEVER!! (1+ / 0-)
    Recommended by:
  •  The press does not get paid to reflect reality. (4+ / 0-)
    Recommended by:
    A Runner, RichM, Snow Camp, bewareofme

    The press get paid to sell advertising views.

    Calling it a tie draws a lot more eyeballs, for a lot more days in succession, than reflecting the numbers honestly.

    I honestly don't understand how anybody can find this surprising.

  •  Only 5 possible explanations (6+ / 0-)

    1 - He/they want Romney to win
    2 - He/they want to artificially play up the "horse race" narrative
    3 - He/they can't stand Obama or really like Romney
    4 - He/they are idiots
    5 - All of the above

    I pick 5.

    "Liberty without virtue would be no blessing to us" - Benjamin Rush, 1777

    by kovie on Fri Sep 14, 2012 at 10:22:42 AM PDT

  •  Further proof that the news media is liberal (0+ / 0-)

    After all, everyone around here with the notable exception of the guy who founded the website (you know, the site that gives us the basis for saying "around here" in the first place) thinks we should act like we're in a dead heat.  CNN is just trying to motivate the base.

    Romney '12: Berlusconi without the sex and alcohol!

    by Rich in PA on Fri Sep 14, 2012 at 10:22:42 AM PDT

  •  Ohio had a D+10 sample! (0+ / 0-)

    and Virginia had a D+5 sample.

    In reality, Obama has a one or two point advantage in Ohio, if not dead even.  This race is still neck and neck.

    The dKos PPP poll only had Obama up by 6 with a D+7 sample.

    •  In 2008, the turnout was D+8 in Ohio. (8+ / 0-)

      D+5 in VA aligns with voter registration in the state I believe.  The FL polls from Marist actually underweighted Dems.  They had a D+2 sample, when Dems have a +4 advantage in registration.  Romney is behind in the battlegrounds.  He has fallen further behind since June 1 in these key states. It is not insurmountable, but when you look at the few number of undecideds out there, it starts to look like a stable margin for Obama.

      The Obama campaign tends to bear down as it gets closer to the finish line and uses the momentum as fuel. Last weekend, over 200,000 voters were canvassed in VA.  More to come.  We're not slowing down.  Obama is going to turn these poll numbers into votes and Romney's erratic campaign is going to cause more undecideds or late deciders to stick with the incumbent.  In 2008, undecideds broke almost evenly between the candidates.

      Alternative rock with something to say:

      by khyber900 on Fri Sep 14, 2012 at 10:30:31 AM PDT

      [ Parent ]

    •  trying to adjust party ID to (2+ / 0-)
      Recommended by:
      Supavash, Free Jazz at High Noon

      a mythical standard is a complete waste of time.

      rinse and repeat.

      Only people who are behind waste their time doing it.

      "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

      by Greg Dworkin on Fri Sep 14, 2012 at 10:59:42 AM PDT

      [ Parent ]

      •  So, you predict higher Dem turnout in Ohio (0+ / 0-)

        than in 2008???

        That's 2 points higher than 2008.  And, that was probably the peak of Dem turnout.

        These polls are ridiculous.

        •  no, it's during a post convention bounce (2+ / 0-)
          Recommended by:
          Supavash, EcosseNJ

          and people in Ohio especially hate Romney. Polls are snapshots, not predictions. It's "right" for this week.

          If/when the national polls get closer, OH will get closer. But Obama will still lead.

          did you see the buzzfeed story that Romney internal polls have Obama +4? So you have a range, 4-7 point Obama lead.  Probably 5.  But it's been a lead all along, this is where the bain ads hurt, and it's not a lead Romney can easily make up.

          "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

          by Greg Dworkin on Fri Sep 14, 2012 at 11:24:53 AM PDT

          [ Parent ]

        •  further, if Obama is ahead by 5 nationally (1+ / 0-)
          Recommended by:

          you have to expect VA and FL to be ahead by 5. Rule of thumb:

          If Ohio and Florida figure to be dead even if the national race is dead even, then they figure to show a five point lead for Obama if he's up five nationally.

          "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

          by Greg Dworkin on Fri Sep 14, 2012 at 11:30:38 AM PDT

          [ Parent ]

        •  heh (0+ / 0-)
          Ignore most sample whining.
          @murphymike via Tweetbot for Mac

          "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

          by Greg Dworkin on Fri Sep 14, 2012 at 05:41:45 PM PDT

          [ Parent ]

          •  until it affects (0+ / 0-)

            mike's team.. heh.  Then he'll be squealing..

            Funny coming from a guy who throws out polls when they don't fit the meme.. heh.

            I didn't bother responding above cuz I am busy.. but here's the thing..

            Convention bounce shows in increases in responses.. not sample bumping.  real bounce shows in bumps in responses that are presented in consistent samples.

            These samples bouncing all over the place make these  polls meaningless.

            I've told you this before.. there will be a lot of people with their jaws dropping (and squealing "fraud!" like Mike) come November when they find the polls have been showing them a very distorted reality.

            That said.. I have no idea where the polls will go after this week's events .  But my impression is this is still a 2-3 point race in Obama's favor.

  •  And when is the last time... (3+ / 0-)
    Recommended by:
    EcosseNJ, bear83, IM

    .....that Romney had any kind of polling lead in any of those states?

    I can't even remember.

    If you have poll after poll showing the same person in the lead, even if every one is within the "technical" margin of error, wouldn't that tend to lead you to the conclusion that the polls are accurate?

    I guess statistically each one is viewed as a stand alone, but clearly a pattern has emerged for Obama, and it seems to me that fact should be highlighted.

    "Pay no attention to that man behind the curtain!"

    by jkay on Fri Sep 14, 2012 at 10:23:13 AM PDT

  •  We're under 8 weeks out. The Horse Race MUST (2+ / 0-)
    Recommended by:
    Thomas Twinnings, keetz4

    go on.

    At least for media outlets to enjoy advertising dollars, that is.

  •  rasmussen has Obama up 1 in Va (0+ / 0-)

    Pretty funny he has Obama up 1 in Ohio and  VA and 2 in Fl. just enough to flip if things get better for Romney.

  •  Must. Maintain. Audience. Interest. (1+ / 0-)
    Recommended by:

    At. All. Costs.

    Ack. It's so hard spinning without getting dizzy.

    "Nach dem Spiel ist vor dem Spiel." ~Sepp Herberger

    by surfbird007 on Fri Sep 14, 2012 at 10:24:28 AM PDT

  •  And Polls are not independant ... (1+ / 0-)
    Recommended by:
    Thomas Twinnings

    This is where Nate Silver is so valuable.  If VA, OH, and FL were truly tied then 1 would be very close, 1 would lean OHB and 1 would lean Mittens ...

    More technically, yes there is some uncertainity, but the maximum likely estimate is that the poll is exactly correct.  It is the expected result, and the further away from the estimate you get the less likely the result.

    I am a statistician, not a magician although we are easily confused. I guess that explains why people keep trying to tie me in chains and place me under water.

    by Edge PA on Fri Sep 14, 2012 at 10:24:39 AM PDT

    •  they are not independent of national polling (0+ / 0-)
      Remember, the first assumption should be uniform swing. If Ohio and Florida figure to be dead even if the national race is dead even, then they figure to show a five point lead for Obama if he's up five nationally. Indeed: if that's the situation, then a new state poll showing uniform swing isn't new information about that state.

      "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

      by Greg Dworkin on Fri Sep 14, 2012 at 11:01:05 AM PDT

      [ Parent ]

  •  Math is hard (2+ / 0-)
    Recommended by:
    bear83, keetz4

    Especially for journalists from the land of holograms, virtual conventions, and cube studios.

  •  Is it the horse race thing or (1+ / 0-)
    Recommended by:

    that they don't want to be accused of being part of Rush's great media consipiracy to "suppress" the GOP vote?

    Ds see human suffering and wonder what they can do to relieve it. Rs see human suffering and wonder how they can profit from it.

    by JTinDC on Fri Sep 14, 2012 at 10:26:21 AM PDT

  •  Will they say we have 2 Pres when Obama wins? (4+ / 0-)
    Recommended by:
    live1, Stroszek, FiredUpInCA, keetz4
  •  That's just so desperate (0+ / 0-)

    Mitt Romney treats people like things. And he treats things - corporations - like people.

    by richardak on Fri Sep 14, 2012 at 10:26:40 AM PDT

  •  So... the Red Sox are "tied" for first place? (1+ / 0-)
    Recommended by:

    Romney/Lyin' - 2012

    by kitebro on Fri Sep 14, 2012 at 10:27:48 AM PDT

  •  No wonder CNN is losing viewers. lol (4+ / 0-)
  •  Bigger issue here IMO (1+ / 0-)
    Recommended by:
    Greg Dworkin

    is that a single poll here or there continues to be news when everybody has access to aggregators like RCP or 538. In a Presidential election few polls tell us much if anything about the race by themselves, but in aggregate, where observations pile up and the error decreases, they tell you almost all you need to know about the race, particularly when accompanied by trend lines.

    Run for office. It's fun!

    by Alfonso Nevarez on Fri Sep 14, 2012 at 10:28:16 AM PDT

  •  One word: arithmetic. (2+ / 0-)
    Recommended by:
    bear83, keetz4

    Republicans and their apologists are all about fuzzy math.

    The real enemy of the good is not the perfect, but the mediocre.

    by Orange County Liberal on Fri Sep 14, 2012 at 10:28:28 AM PDT

  •  So Obama has to win by a touchdown?? (0+ / 0-)

    This would have been EXCELLENT NEWS for John McCain!

    •  It is going to work on a point spread (0+ / 0-)

      this election based on a complex algorithm of political and golf handicaps. Right now, Obama has to win the electoral college by 35.

      "You can die for Freedom, you just can't exercise it"

      by shmuelman on Fri Sep 14, 2012 at 10:33:47 AM PDT

      [ Parent ]

  •  i guarantee you neither the Obama camp (5+ / 0-)

    nor the robme camp are thinking that is a tie.
    one of those camps is panicking

    "It's never too late to be who you might have been." -George Eliot

    by live1 on Fri Sep 14, 2012 at 10:30:09 AM PDT

  •  MOE declines when you add other polling (2+ / 0-)
    Recommended by:
    Sue B, bear83

    I'm glad to see that the AP hates the term "statistical dead heat," for exactly the right reasons. And what it says about how to describe the results of an individual poll is fair enough.

    But, if you look at the entirety of recent national polls or recent state polls, the combined sample sizes reduce the margin of error and we can be much more certain than 95% that a candidate with a 5-point lead is ahead.

    In general, I follow what Nate Silver says as the most likely accurate, ahead of Poll Tracker (which kos relies on a lot), Pollster, or RCP. Heck, RCP just put Michigan in Obama's column 2 days ago, and still calls Nevada a tossup despite the fact that Romney has never led there in more than a year's worth of polling, with one tie last October.

  •  The margin of error is CNN. (5+ / 0-)
    Recommended by:
    OnlyWords, RichM, Jaleh, rg611, bewareofme

    Once in a while you get shown the light, in the strangest of places if you look at it right. I'm riding for MS in September. Please donate here if you can.

    by darthstar on Fri Sep 14, 2012 at 10:30:45 AM PDT

  •  Who still watches CNN? (1+ / 0-)
    Recommended by:

    "Rick Perry talks a lot and he's not very bright. And that's a combination I like in Republicans." --- James Carville

    by LaurenMonica on Fri Sep 14, 2012 at 10:31:14 AM PDT

  •  Thank you for this. (2+ / 0-)
    Recommended by:
    live1, FiredUpInCA

    Being new to paying any close attention to polling has proved demanding of me.

    Anything that makes it all more accessible/understandable  is much appreciated.

    "The most potent weapon in the hands of the oppressor is the mind of the oppressed." ~ Steven Biko

    by Marjmar on Fri Sep 14, 2012 at 10:31:42 AM PDT

  •  I'm fine with that. (1+ / 0-)
    Recommended by:

    I hope everyone on our side and every independent goes into the election thinking it's close.

    Hyperbole will be the death of us all!

    by MrHinkyDink on Fri Sep 14, 2012 at 10:31:50 AM PDT

  •  Middle income is now between $200K/$250K (4+ / 0-)
    Recommended by:
    RichM, bear83, FiredUpInCA, Supavash

    Wow, all this time i thought i was middle class, i guess i am in poverty! According to Rmoney:

    A man's character is his destiny.

    by Jaleh on Fri Sep 14, 2012 at 10:31:52 AM PDT

  •  So when my daughter's... (4+ / 0-)
    Recommended by:
    live1, bear83, FiredUpInCA, keetz4 team lost 2-1, I could point to it as a clear victory? It's within the margin!

    Twitter: @michaelhag

    by MichaelPH on Fri Sep 14, 2012 at 10:32:03 AM PDT

  •  it's OK (0+ / 0-)

    The D side should not get complacent, and voter suppression might be worth a couple of points for Romney. If CNN wants to pretend the race is closer than it is, this doesn't really hurt Obama any.

  •  CNN is pulling the same shit they did 4 years ago. (5+ / 0-)
    Recommended by:
    Stroszek, The Wizard, keetz4, askew, EcosseNJ

    The GOP campaign pleads with them for fairness...says their internals show a closer race...they say wait for the debates before you state that Obama is ahead....

    It doesn't matter.  Democrats know there is a lot of work to be done between now and November 6th, so it won't slow us down.

    Alternative rock with something to say:

    by khyber900 on Fri Sep 14, 2012 at 10:32:13 AM PDT

  •  So when my daughter's... (1+ / 0-)
    Recommended by:
    live1 team lost 2-1, I could point to it as a clear victory? It's within the margin!

    Twitter: @michaelhag

    by MichaelPH on Fri Sep 14, 2012 at 10:32:33 AM PDT

  •  Thanks for hammering this (1+ / 0-)
    Recommended by:
    Greg Dworkin

    point. In fairness, commenters here are every bit as likely to use the nonsensical descriptor "statistical tie" as the average national journalist. But, perhaps with clear explanations like this one, that will subside.

    Grew a mustache and a mullet / Got a job at Chick-Fil-A

    by cardinal on Fri Sep 14, 2012 at 10:32:45 AM PDT

  •  NYT has RV at 8% but LV at only 3% for Obama? (0+ / 0-)

    The NYT/CBS poll results today showed Obama with a 3% lead over Romney after a likely voter screen, but with a commanding 8% lead (51 - 43) using registered voters.

    Now I can accept that the LV screen will usually benefit the Republicans, but is a 5% effect large compared to other polling?  In this year, there can't be any particular enthusiasm for Romney but more an antipathy to Obama which is pushing the LV numbers towards Romney.

    If nothing else, the RV vs LV disadvantage makes the GOTV and anti-suppression efforts all that more important in the swing states.

    •  LV screens before october are iffy (0+ / 0-)

      what's a likely voter, anyway?  marist says:  

      A likely voter is determined based on interest in the upcoming election, the chance of voting, and prior participation in past elections.
      so younger and occasional voters are sometimes excluded. Should be a 2 point difference more or less, not 7 or 5.

      "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

      by Greg Dworkin on Fri Sep 14, 2012 at 11:05:13 AM PDT

      [ Parent ]

  •  Landslides are bad for business. (1+ / 0-)
    Recommended by:

    Prepare for a new "Weekend at Bernies" where the media drags Rmoney/Lyin' Ryan along and pretends they're still viable.

    Prefers Obamacare over Vouchercare

    by Palafox on Fri Sep 14, 2012 at 10:33:43 AM PDT

  •  If... (5+ / 0-)

    The numbers were reversed, the headline would be: "Obama in big trouble in swing states as Romney takes decisive lead!"

    Why is Mitt Rmoney so happy that an American embassy was attacked? Why does he hate America?

    by RichM on Fri Sep 14, 2012 at 10:34:01 AM PDT

  •  I know it's not worth much, but (1+ / 0-)
    Recommended by:

    Obama keeps rising on Intrade. He's risen steadily for the last 5 days and is now up to 66%.

    Of course, the one to really watch is Nate Silver's 538. I'll only begin to worry about this race if Nate shows it as 50/50. Currently at 78.6% for Obama. Anything below 60% and I'll start to get concerned. As it is now, it's clear Obama is winning. And I don't see the debates changing that.

    Romney's team keeps thinking they can win over the undecideds (because the odds of flipping Obama voters at this point are negligible). But NBC said something interesting today about the undecided voters in their swing state polls:

    * Who are the undecideds? By the way, we looked at the undecideds in all three states -- collectively -- and these are voters who simply aren’t paying attention. And frankly, they don’t look like voters. Consider: A third of them are also undecided on the president’s job rating!?!! They like the president personally (more so than Romney), but they also view the direction of the country a lot WORSE than the rest of the voters. Bottom line: These folks look like DISAFFECTED voters, they seem disengaged from the campaign, and they don’t call themselves enthusiastic about the election. They are probably NOT voters.

    I'm a dyslexic agnostic insomniac. I lie awake at night wondering if there's a dog.

    by rennert on Fri Sep 14, 2012 at 10:34:30 AM PDT

  •  According to NYT (0+ / 0-)

    With their back-to-back political conventions behind them and the general election season fully engaged, the poll found Mr. Obama and Mr. Romney running essentially even among those seen as the most probable to vote. The president has 49 percent and Mr. Romney has 46 percent, a difference within the margin of sampling error of plus or minus three percentage points on each candidate.

    A man's character is his destiny.

    by Jaleh on Fri Sep 14, 2012 at 10:36:28 AM PDT

  •  Margin of error. (0+ / 0-)

    Look, guys, if an editor wanted fo know what's happening to farms in the drought, he wouldn't look at his yard and decide.

    He would F****ing well ask an agriculture professor.

    On statistics, though, editors figure that their ignorance should be imposed on everyone.

    First of all, no, you can't say that the margin of error in the difference is O's margin of error plus R's margin of error.

    Second of all, the possible sample answers from a population follows the "Bell Curve." It isn't a simple number. (And, too, the problems of the sample being non-representative are much more worrisome than the statistical variance.)

    Somewhere on this Blog, there is a staistician. My specialty, back when I studied Math, was abstact Algebra. Staistics isn't my field.

    Please, would a statistician writte on the subject?

    •  standard practice is considering 2 x MoE as a lead (0+ / 0-)

      as Ap does. Their style sheet is useful even if not gospel.

      "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

      by Greg Dworkin on Fri Sep 14, 2012 at 11:06:53 AM PDT

      [ Parent ]

      •  Yeah, they compare 2 doubliings. (0+ / 0-)

        I thought about this more. If nobody who knows more about statistics posts something relevant, I'll bring my 1/2 of an education on the subject to bear.

        The lead is really a doubling; it's Smith's number minus Brown's number. Then they use twice the margin of error to compare.

        The problem is that too many of the innumerate hear "margin of error" as something absolute. It ain't. The difference caused by statistical error is unlikely to be a little inside the margin of error; it is slightly less likely to be a little outside the margin of error.

        And since the poll results are reported to the nearest whole percentage p[oint, how far outside the poll results are is a matter of guess work. I wish that they would say "This great a difference would happen X% of the time if the underlying population was evenly divided." But they never will. That would only be helpful to those who know ariithmetic, and that is not the audience which advertisers want.

        •  what they re trying to get at is this: (0+ / 0-)

          given this lead of X points, does that mean candidate A is likely to win if the election were held today? That's all.

          What does likely mean? How much is enough to say that? Imperfect though the AP guidelines may be, if someone is leading by twice the margin of error, the public would be shocked if the guy lost, and it would be considered a true upset. In that sense, the do the job.

          "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

          by Greg Dworkin on Sat Sep 15, 2012 at 09:49:37 AM PDT

          [ Parent ]

    •  at the same time (0+ / 0-)

      only a tie is a tie. I like their guidelines.

      "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

      by Greg Dworkin on Fri Sep 14, 2012 at 11:08:06 AM PDT

      [ Parent ]

  •  Rep. Rangel Says What Everyone Is Thinking: (5+ / 0-)
    Recommended by:
    Jaleh, Sue B, bear83, Supavash, keetz4
    "Mitt Romney is Threat to Our National Security"
  •  CNN now calls Obama's 2008 win a toss-up (2+ / 0-)
    Recommended by:
    rennert, Supavash

    Turns out it wasn't within their margin of narrative error.

    "Liberty without virtue would be no blessing to us" - Benjamin Rush, 1777

    by kovie on Fri Sep 14, 2012 at 10:37:57 AM PDT

  •  Front Page New York Times right now (7+ / 0-)
    Recommended by:
    live1, Jaleh, bear83, IM, The Wizard, Supavash, askew

    Poll Shows Obama With Narrow Edge Over Romney

    A New York Times/CBS news poll found that Mitt Romney and President Obama remain essentially tied among likely voters

    [my underline]

    The 'liberal' media is so careful not to look liberal. I've caught the NYT a couple of times with very misleading, ie slanted right, headline/title links on the front page

  •  On "liberal" MSNBC right now.... (3+ / 0-)
    Recommended by:
    FiredUpInCA, Supavash, bewareofme

    .... The anchor, Mrs. Andrea Greenspan, is hosting a panel consisting of Susan Page, Major Garrett and Ruth Marcus.

  •  Is CNN smarter than a 5th grader in math? (0+ / 0-)

    Psst!!!......Willard let us see your income taxes.

    by wbishop3 on Fri Sep 14, 2012 at 10:40:06 AM PDT

  •  It's simple greed (1+ / 0-)
    Recommended by:

    There is no ideological component here.  When all is said and done, there will have been upwards of $1 Trillion spent on this campaign.  Most of that money goes to media buys.  The money blitz is fueled by the perception that it is possible for Romney to win.  This is because the big money suckers are dumping huge sums into his campaign and then of course Obama has to follow suit or else get swamped.  So, the moment it looks like the race is over, the gravy train starts to slow down.  The longer they can make the thing look competitive the more dollars they can vacuum up from Romney's wealthy suckers and those trying to shield the President from them.  

  •  while you might be technically correct (0+ / 0-)

    my first instinct these days is to worry about complacency and a low turnout, so who really cares if someone doesn't point out Obama's lead right now? Let everybody think it's close, for the benefit of turnout, until the 6th.

  •  94% chance of winning is called a tie (0+ / 0-)

    Using the numbers from their poll you can calculate the odds of Obama being ahead in Florida using this:  Ballot Lead Calculator

    With a sample size of 980, a 49/44 race gives a 94% chance of winning to the leader.

    Since that is less than the 95% level of certainty, they call it a tie.


  •  A bit of tomato-tomahto, but you're right (1+ / 0-)
    Recommended by:

    that it's misleading to call it a tie.

    Yes, it could be tied.
    For that matter, Romney could be ahead.
    Most likely, however, is that Obama leads by some not-very-large amount.
    Less likely, but not impossible, is that Obama leads by a large margin.

    LG: You know what? You got spunk. MR: Well, Yes... LG: I hate spunk!

    by dinotrac on Fri Sep 14, 2012 at 10:43:25 AM PDT

  •  We're 10 points down buster and don't you start (0+ / 0-)

    hooping and hollerin' otherwise.


  •  We'll wait for Election Day of course... (0+ / 0-)

    ... but it seems like all the $$ and the media BS is not making much a dent in this race. Seems like enough people can filter out the noise and figure it our for themselves.

    The reactions to both conventions in the polling was very encouraging for my faith in the electorate.

    No snowflake in an avalanche ever feels responsible.

    by Magster on Fri Sep 14, 2012 at 10:46:44 AM PDT

  •  No no no, it's not technically right at all (6+ / 0-)

    You are too gentle on them.

    As a statistician, let me say, there is no such thing as a "statistical tie". Within the margin of error does not mean tied. Not technically, not literally, not metaphorically, not any way.

    The only way for a poll to show a tie is if exactly the same number of people say (e.g.) "Romney" and "Obama".  That's a tie.

    If 201 people say "Obama" and 199 say "Romney" then that is a lead for Obama. It is not a statistically significant lead. But statistical significance just isn't very significant. It is the most misused term in statistics.

    In analyzing a poll, what we want to know is "Given the poll results, what are the chances that Obama is winning?"
    That is not what statistical significance answers. It answers: "If the race were really tied, what are the chances of getting results like these or more extreme than these?"

    These two questions are NOT the same and they do NOT always get the same answer.  E.g

    Given that someone is a white male American, what are the chances he is s United States Senator?

    does NOT get the same answer as

    Given that someone is a United States Senator, what are the chances that he is a White male American?"

    Obama is winning. Period.

    •  oh, good, a statistician.... so next (1+ / 0-)
      Recommended by:

      can you explain Bayesean probability and credibility interval? That's what on-line panels use. There is no confidence interval.

      You have 3 minutes. You may use texts and the internet.

      Go. ;-P

      "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

      by Greg Dworkin on Fri Sep 14, 2012 at 11:11:35 AM PDT

      [ Parent ]

      •  Sure! (2+ / 0-)
        Recommended by:
        Greg Dworkin, Twilight Jack

        In Bayesian statistics we take our "prior beliefs" into account and modify them based on new data, combining the two into a new probability called the posterior.

        We can limit the role of prior beliefs by assuming a "noninformative prior".

        Properly done, Bayesian statistics gives an answer to the right question. Improperly done, it gives nonsense. The proper question was stated in my earlier comment, and Bayesian probability and credibility intervals are about that question: Given these results, how likely is XXX?

        Of course, improperly done frequentist statistics also gives garbage, but (in my view) it is easier for two trained and capable people to get very different results when using Bayesian methods than when using frequentist ones.

        •  hence Zogby panels vs Ipsos panels (1+ / 0-)
          Recommended by:

          makes perfect sense, actually.

          Do you want to know the kind of jar given the cookies, or the kind of cookies, given the jar? Different ways of describing the same physical thing, but different info produced..

          "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

          by Greg Dworkin on Fri Sep 14, 2012 at 11:46:24 AM PDT

          [ Parent ]

  •  a tie (1+ / 0-)
    Recommended by:

    much more easily followed by "so stay tuned to CNN" than "its over for Romney"

  •  total face palm.. this is beyond ignorant.. (2+ / 0-)
    Recommended by:
    IM, RichM

    a 5 point lead is "within the margin of error of 3.1"????

    Holy smoke where to begin..

    simply.. the  margin of error is based on a number of standard deviations from the norm when obtained data is normalized into a bell shape of probabilities..

    being within a certain number of standard deviations from the average creates certain levels of confidence that a poll is accurately portraying reality.

    an margin of eorror of 3.1 .. is + or - 3.1 from tne average which is in this case Obama +5. ..

    So.. what is being said in this poll is the odds at around 90% that Obama's lead is within the range of 8.1 - 1.9

    the odds that this is a tied race ar out of that range and therefore less than 10%

    therefore it is over 90% likely CNN's comment that this is a tied race based that poll is wrong.

    Obama has a statistically significant lead.. PERIOD.

  •  See my post from earlier today (0+ / 0-)

    I'm no philosopher, I am no poet, I'm just trying to help you out - Gomez (from the song Hamoa Beach)

    by jhecht on Fri Sep 14, 2012 at 10:55:53 AM PDT

  •  50%-43% is "trending to Obama" ???? (0+ / 0-)


    "A squid eating dough in a polyethylene bag is fast and me?" - Don Van Vliet

    by AlyoshaKaramazov on Fri Sep 14, 2012 at 10:56:50 AM PDT

  •  if a candidate is ahead by 5 points in ALL (0+ / 0-)

    of the polls, then why are we even TALKING about MOE?

    I'd call that a CONFIRMED lead.

    "A squid eating dough in a polyethylene bag is fast and me?" - Don Van Vliet

    by AlyoshaKaramazov on Fri Sep 14, 2012 at 10:59:42 AM PDT

  •  The Pollsters; (1+ / 0-)
    Recommended by:

    Romney's state leads are only coming from the same two pollsters. Rasmussen & Purple Strategies and/or newbies like McLaughlin & Gravis a few local staties and Civitas in NC.

    Obama's come from all over the spectrum including those above.

    Quin, NBC, CBS, NYT or PPP have not given Romney any real, regular state leads/momentum in months, not since he was primary jousting with the Ricks & Noot, one Colorado poll aside and PPP giving him the WI bounce last month. Nothing much. Even EPIC-MRA said to hell with this with their recent MI poll and I know, sometimes even Scotty can't twist the numbers to suit his narrative in his state polling; OH, VA and FL the last 24hrs... Outside the.. ahem, "battlegrounds" and into the swing states sure Romney gets the odd 1-3pt lead from Ras or PS or from Gravis and the bloody 'send us a letter' Columbus Dispatch etc but never much of anything from any other renowned polling firm, not during his primary wrap, not during Lyan's announcement, the RNC, nothing... it was pretty much all from the two firms above...

    Cheers, CNN like your polls, shame about your failed nooz spin...

    Moving stations "The Clinton bounce?" Chucky Cheese Todd is an utterly daft wanker...

    "Never trust a man who, when left alone with a tea cosy, doesn't try it on!!"

    by EcosseNJ on Fri Sep 14, 2012 at 11:05:15 AM PDT

  •  I'm OK with this (0+ / 0-)

    Let's not get complacent. It's the biggest threat to an Obama win right now.

  •  Nice job (1+ / 0-)
    Recommended by:
    Greg Dworkin

    This is a fairly minor problem (in that the statistics would not support a conclusion that either candidate is ahead, statistically ahead.  However AP has technically good way to discuss the issue.  In these cases polls show a lead but not a significant one

    However it is also important to recognize that each poll"s MOE is only with reference to that poll only. When we have a bunch of polls showing the same thing the aggregate power of the polls goes up such that three similar boy non significant polls may produce a significant difference

    I'm going to cut  reporters a little slack since few Americans are that numerate

    Courtesy Kos. Trying to call on the better angels of our nature.

    by Mindful Nature on Fri Sep 14, 2012 at 11:07:19 AM PDT

    •  true but (0+ / 0-)

      just don't call it a !@#%%^ tie.  ;-P

      "Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong remedies." - Groucho Marx

      by Greg Dworkin on Fri Sep 14, 2012 at 11:27:37 AM PDT

      [ Parent ]

      •  It's a bit of a philosophical problem (1+ / 0-)
        Recommended by:
        Greg Dworkin

        As Steve notes, on a technical level, we certainly can't say for sure it's not a tie, which is the null hypothesis. (I'd quibble with his formulation somewhat, since in a poll we are trying to discern whether someone is ahead, which means the starting assumption is that there's a tie, unless we can show otherwise).  In that sense, it's a (marginally) defensible statement statistically.

        However, journalistically it does not give the accurate portrayal that the AP guidelines I think capture well.

        Courtesy Kos. Trying to call on the better angels of our nature.

        by Mindful Nature on Fri Sep 14, 2012 at 12:06:52 PM PDT

        [ Parent ]

  •  So what CNN is really trying to say is... (2+ / 0-)
    Recommended by:
    pollbuster, bewareofme

    This is great news for John McCain!

  •  But if Romney leads by 1 in a poll, (1+ / 0-)
    Recommended by:

    it's romney surges into the lead.

    The definition of INSANITY: Voting Republican over and over and over and expecting the economy to get better.

    by pollbuster on Fri Sep 14, 2012 at 11:14:00 AM PDT

  •  ZNN all the way. Bottom-of-the-barrel tv news. (0+ / 0-)

    I ♥ President Barack Obama.

    by Blue Waters Run Deep on Fri Sep 14, 2012 at 11:15:42 AM PDT

  •  I totally agree with your assessment but CNN and (0+ / 0-)

    the other outlets who engage in this journalistic malpractice are doing us a huge (and entirely unintentional) favor. Anxious grassroots activists and GOTV workers will work harder. The harder and more unrelenting our efforts, the greater our down-ticket return. We need to ensure the advantages the GOP have handed us translate into a secure-from-Blue-Dog-apostasy House and a Senate which Harry Reid's caucus feels comfortable running without the filibuster. Only with those two things can we hold a lame duck's feet to the progressive fire for the next four years.

    I'm going to risk ruining the effect of the above by quoting the truism from those ridiculous commercials: "Stay hungry, my friends!" And leave everything on the playing field. This race still needs us to turn it into a blowout. If you find yourself a little tired, think about people like me, working on and perhaps looking at our last elections. Think about your grandchildren in 20 years who will ask how hard you worked in 2012. Ignore the polls and leave nothing to chance.

    Enough fossil fuel remains on Earth to warm it 6 degrees C by 2100 AD if it is all used. A +6 C planet will only sustain half a billion humans. Human population will rise to 9 billion by 2050. Any questions?

    by davidincleveland on Fri Sep 14, 2012 at 11:21:36 AM PDT

  •  As a prof pollster, AP's criteria generous (0+ / 0-)

    Double the margin of error is usually reserved for between sample polls, and is 95% accurate.  So if CNN conducted two polls and got an average result that was greater than 2 times the MoE, the average result would be 95% accurate.  For within-sample analyses of a poll, results that exceed the samples MoE are 95% accurate on their won, so calling a result a little bit greater than the MoE slightly ahead is being safe, and gives greater weight to the 5% chance that they're wrong.  

    CNN's assertion is basically journalistic malpractice.  To call a result that is 90% accurate when comparing between-sample polls, or one and a half times the MoE, a tie, is an organization trying to sell coverage that overwhelmingly doesn't exist.

    Unfortunately I've seen people incorrectly characterize one point margins regularly on this site.  A 46 to 45% Obama advantage a not a lead, it's really a toss-up, just as a 46 to 45% Romney lead would also be a tossup.  

    That said, there's a big difference between a person, who represents themselves only, commenting on a finding, and a journalist, who represents a major media outlet, making such claims.

    The only thing we have to fear is fear itself - FDR. Obama Nation. -6.13 -6.15

    by ecostar on Fri Sep 14, 2012 at 11:32:48 AM PDT

  •   The pressure on CNN from the corporate suits (1+ / 0-)
    Recommended by:

    to make it look like a close race is probably enormous.

  •  I commented earlier on a similar post... (0+ / 0-)

    The craziest thing is that I've been following CNN doing this before.  Even crazier is that when Romney is up 3 or 4 in a poll they clearly state that he is leading, or has an edge.  

  •  So if CNN (1+ / 0-)
    Recommended by:

    gets 5% more viewers than their nearest competitor, I wonder if they tell advertisers they're tied with them or if they proudly state 'we're winning'?

    The choice of our lifetime: Mitt Romney, It Takes A Pillage or President Barack Obama, Forward Together.

    by FiredUpInCA on Fri Sep 14, 2012 at 11:40:59 AM PDT

  •  Tidbits like this are why CNN (0+ / 0-)

    is falling to third place in the ratings.

    ... the watchword of true patriotism: "Our country - when right to be kept right; when wrong to be put right." - Carl Schurz; Oct. 17, 1899

    by NevDem on Fri Sep 14, 2012 at 11:42:31 AM PDT

  •  Margin of error usually determined by 95% (0+ / 0-)

    probability.  If there's a 5% chance that Romney is in fact ahead in one of these "tied" states, the chance that he is in fact ahead in two of them is 1/400, or 1/4%.  So the chance that Romney is ahead in two is beyond the bounds of statistical probability, as it is normally determined.

    The influence of the [executive] has increased, is increasing, and ought to be diminished.

    by lysias on Fri Sep 14, 2012 at 11:43:06 AM PDT

  •  error is correlated though isn't it (0+ / 0-)

    So Obama is up 49-44-7, with the 7 being an other category to get the numbers to add up to 100.  Because at the end of the day it would seem sans rounding artifacts they would need to add up to 100.  Say in the unlikely case that Obama is at the top of his margin of error and the 'real' number of Obama support is 46.  Now we still have to allocate the 3% somewhere, and it seems better to allocate ~86% of it to Romney[the same percentage of the existing non-Obama vote].  This would put Romney at 46.6.

    What exactly is the flaw in this reasoning?  While an unlikely state of the race it is at least possible.  Put another way, if you assume that one party is 10% likely to be at the edge of their margin of error it would be wrong to state that the odds of both of them being at the edge is 1%, because once one is at the edge the odds of the other also being close to the opposite edge should be closer to 90%.

  •  What is the confidence interval here ... (0+ / 0-)

    I presume that margin of error refers to precision.  

    Let's take this case:

    President Obama had 49% of the vote, and the poll claimed a confidence interval of 90% with a precision of +/- 3%.  Let's give Romney 43%.

    I would read that as a 90% chance that the support for Obama in the enitire population was between 46 and 52%.  There is a 5% chance that Obama is above 52%, and a 5% chance that he is below 46%.

    There is also a 90% chance that Romney is between 40 and 46%.

    I would say, based on this hypothetical poll, that Obama has a 95% chance of winning the state.  Hardly a tie.

    Am I reading this right?

    No one could make a greater mistake than he who did nothing because he could do only a little. - Edmund Burke

    by AdirondackForeverWild on Fri Sep 14, 2012 at 01:44:07 PM PDT

  •  That is one scary-ass-looking expression (0+ / 0-)

    on his face. Looks like an entirely different man from Smirking Mitt, more malevolent than vapid.

    !! Four more years !!

    by raincrow on Fri Sep 14, 2012 at 03:09:02 PM PDT

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site