home

Ras Tracker: Obama By 1

The GOP Convention bounce begins its fade. McCain has gone from 1 up to 1 down in the Ras poll in the past few days. In addition, Ras pooh poohs the move of white women to McCain/Palin:

A Washington Post poll generated comment yesterday by noting that White Women favor McCain. Rasmussen Reports polling has consistently shown a similar result, but it is important to note that there is nothing unusual about this finding. Four years ago, President Bush managed to defeat Senator John Kerry 55% to 44% among white women. Today’s tracking poll data shows McCain leading 51% to 44% among this group. Among all women, Obama leads by eight. Among men, McCain leads by eight.

By Big Tent Democrat

< Can A Family Court Judge Restrict The Right To Petition The Government? | McCain/Palin Ape Bush/Rove Political Tactics >
  • The Online Magazine with Liberal coverage of crime-related political and injustice news

  • Contribute To TalkLeft


  • Display: Sort:
    A shoo-in! (5.00 / 1) (#1)
    by Steve M on Wed Sep 10, 2008 at 09:15:48 AM EST
    I heard it from a reliable source.

    And I said. . . (none / 0) (#13)
    by LarryInNYC on Wed Sep 10, 2008 at 09:35:34 AM EST
    Obama by 10 to 15.  We'll see. . .

    Parent
    Gosh (none / 0) (#41)
    by Steve M on Wed Sep 10, 2008 at 09:57:34 AM EST
    I hope it's not too late to claim you were speaking in terms of electoral votes.

    Parent
    No, actual votes. (5.00 / 2) (#83)
    by LarryInNYC on Wed Sep 10, 2008 at 10:48:57 AM EST
    Obama over McCain by ten to fifteen voters.

    Parent
    Even better (none / 0) (#85)
    by Steve M on Wed Sep 10, 2008 at 10:52:39 AM EST
    No wonder Obama picked a Catholic for VP, heh heh.

    Parent
    Still tied up. (5.00 / 2) (#4)
    by Ga6thDem on Wed Sep 10, 2008 at 09:28:12 AM EST
    It's probably going to stay tied up forever. The McCain campaign has said that if they can win white women by 10% then they can win the election. I guess we'll have to wait and see if they can pull it off.

    Wrong (5.00 / 1) (#7)
    by bluegal on Wed Sep 10, 2008 at 09:32:11 AM EST
    The McCain campaign said that if they win Hillary Clinton voters they win the election.

    First, according to Ras. McCain is still doing a little worse than Bush in the final outcome and this is almost two months BEFORE the vote.  What's to say that Obama won't shrink the lead or even regain it among white women?  

    The only way McCain can win is if he not only improves his share of white women but also get the HRC white women who hate Bush and our in economic trouble.  

    You seem confident that this bounce will last, I don't.

    Parent

    Problem: the poll doesn't yet show (5.00 / 2) (#11)
    by andgarden on Wed Sep 10, 2008 at 09:34:42 AM EST
    Obama doing any better than Kerry with these voters.

    PPP is noticing a trend, and it's a predictable but concerning one, that most undecided whites are shifting to McCain. Obama has to stop that, or he loses.

    Parent

    Yes (5.00 / 2) (#14)
    by bluegal on Wed Sep 10, 2008 at 09:38:13 AM EST
    These are the republican women coming home. I never expected Obama to win North Carolina but if you look at a state like VA which has been red since 1964 and the numbers haven't budged at all the idea that women are going to McCain in droves just doesn't hold up.  These are Republican women for the most part coming home to McCain.  I expect his share of the white women vote will decrease in the coming weeks as those  independent white women switch once again.

    Parent
    It's all about Colorado now (5.00 / 0) (#18)
    by andgarden on Wed Sep 10, 2008 at 09:39:55 AM EST
    The vaunted financial advantage is almost meaningless.

    Parent
    Obama has to DO stuff to get Colorado (5.00 / 0) (#23)
    by Militarytracy on Wed Sep 10, 2008 at 09:42:51 AM EST
    He has to advocate and sell his ideas like Bill Clinton did.

    Parent
    You are right (5.00 / 2) (#24)
    by bluegal on Wed Sep 10, 2008 at 09:43:15 AM EST
    My money is on Obama in Colorado.  There was an article in the Boston Globe that had Charlie Black McCain's top adviser discussing how they are not capitalizing on the huge crowds because it would be time consuming. Um, isn't that how you target voters?  It is more and more obvious that the McCain campaign is not going to have a good ground game.

    I remember Kerry having big rallies and didn't bother to use those rallies as a way to target voters and he lost.  McCain is doing the same thing.

    Parent

    I tend to think that big rallies (5.00 / 2) (#28)
    by andgarden on Wed Sep 10, 2008 at 09:47:52 AM EST
    are of limited utility. They are also off putting to many undecided voters.

    Parent
    There is no ground game in Colorado (none / 0) (#32)
    by Militarytracy on Wed Sep 10, 2008 at 09:52:32 AM EST
    after the caucuses.  It isn't machine politics in the wild west. You are talking about independent minded people with a very large Libertarian streak  (and most of them don't even know that about themselves) and sadly the Republicans are usually found more appealing to them because they do actually do stuff and make decisions (poor decisions).  The repubs stand for something though, just walking into Colorado and saying that you are the good guy will not get you your votes, you better bring some goods and then tell the people why your goods are better than McCain's.  Where oh where is Bill Clinton when we need him?

    Parent
    Don't tell me that! (5.00 / 1) (#20)
    by Militarytracy on Wed Sep 10, 2008 at 09:41:10 AM EST
    I don't want to hear that and expecially not from you, who knows how to read and add and subtract and such stuff!

    Parent
    Bleh. I expect McCain to win. (5.00 / 2) (#77)
    by cosbo on Wed Sep 10, 2008 at 10:38:04 AM EST
    If Obama then I'll be happily surprised. I will not invest myself in the democratic ticket like I did in 2000 & 2004. I'm tired of the heartbreak.

    And frankly if the democrats really wanted to win they would not have selected an untested, inexperienced candidate for the top of the ticket.

    Parent

    Andgarden....many times we have not (2.66 / 3) (#17)
    by PssttCmere08 on Wed Sep 10, 2008 at 09:39:19 AM EST
    agreed, but you do seem to be one of the few grounded in reality regarding this poll...seems it is way to early to tell anything and I see alot of straw grasping going on.  obama is going to have to do a whole helluva lot better before he is out of the woods, imo.

    Parent
    on the bright side (none / 0) (#36)
    by coigue on Wed Sep 10, 2008 at 09:53:37 AM EST
    McCain isn't doing as well as Bush, either

    Parent
    Well (5.00 / 1) (#21)
    by Ga6thDem on Wed Sep 10, 2008 at 09:41:16 AM EST
    he's not doing what he needs to do to regain those votes. He has never worked for those votes that I've seen, he's always just expected them. You can't expect people to vote for you, you have to ask for and earn their votes.

    Parent
    It's a bit funny that a back and forth (none / 0) (#68)
    by fercryinoutloud on Wed Sep 10, 2008 at 10:22:39 AM EST
    swing of a statically insignificant 1 pointis a "fade".

    As for Ras's women's vote comment what wasn't referenced was what I posted a few days ago:

    Ras:

    Obama leads by three among women. On Tuesday, when Obama's lead peaked, he had a fourteen point advantage among women.

    So does Ras really pooh pooh the move of white women to McCain/Palin?

    Well the facts are that Obama actually has a 14 point lead among women. McCain now leads that category by 7 points compared to Obama's previous 14 pint lead. That is a 21 point swing of ALL WOMEN.

    So BTD, what women have moved to McCain/Palin in BIG numbers to give him the overall lead in women? Blacks? Latinos? Whites?

    I'd say Whites but I'm just guessing.

    Parent

    But isn't it of note... (5.00 / 0) (#6)
    by Exeter on Wed Sep 10, 2008 at 09:29:52 AM EST
    ...that McCain has gotten back his white women margins to where they need to win?  

    Not necessarily (5.00 / 1) (#10)
    by bluegal on Wed Sep 10, 2008 at 09:34:25 AM EST
    I have a feeling that McCain right now has seen a surge in Republican and Independent white women. I'm sure that he will keep the Republican women but the Independent women I doubt will hold once Palin's record is exposed.  I cannot tell you how many chain letters I have gotten in the last few days about Palin from former HRC supporters and it is not pretty.

    Parent
    i suppose it all depends (5.00 / 1) (#40)
    by TimNCGuy on Wed Sep 10, 2008 at 09:57:27 AM EST
    on what states they live in, right?

    If McCain is up among ALL men and Obama is up equally among ALL women....  and there are MORE women voters than men voters, Obama should be in the lead.  

    What really matters electorally is which voters live in which states.

    There was a report on CNN last night that explained that whoever wins southern Ohio (not the whole state, but just southern Ohio) has won all the elections since the 70's.  The only times southern ohio went for dems was for Carter and Clinton.

    Southern Ohio is in Appalachia.

    Parent

    Then (none / 0) (#47)
    by Ga6thDem on Wed Sep 10, 2008 at 10:06:11 AM EST
    taking all this to the logical conclusion it would mean Obama losing the election in Nov.

    Parent
    not necessarily (none / 0) (#51)
    by TimNCGuy on Wed Sep 10, 2008 at 10:09:46 AM EST
    the report on CNN last night did point out that many younger voters in southern ohio are backing Obama.  More so than who backed dems in the past.

    Parent
    thanks for saying that (none / 0) (#59)
    by coigue on Wed Sep 10, 2008 at 10:15:48 AM EST
    it's a bit more nuanced than just racism. But racism is definitely a factor.

    Parent
    it also comes down to (none / 0) (#64)
    by TimNCGuy on Wed Sep 10, 2008 at 10:20:21 AM EST
    how big the repugs win in southern ohio.  in past elections the reougs have won HUGE there and it was enough to counter the wins by dems in places like Cleveland.

    So, repugs could still win southern ohio but not by a large enough margin to counter dem wins in other parts of the state.

    Parent

    The Republicans are also (none / 0) (#71)
    by ruffian on Wed Sep 10, 2008 at 10:25:58 AM EST
    challenging new voter registrations in Cleveland. Von Spokovsky (sp?) is leading the fight.

    I think Ohio is the key, just like last time.

    Parent

    I don't know. (none / 0) (#74)
    by coigue on Wed Sep 10, 2008 at 10:30:41 AM EST
    There are lots of keys this time. And a few ways we could win w/o Ohio.

    FLA may be in play because Palin's views may be too extreme for the retired NYers there.

    Interesting times.....

    Parent

    Cincinatti will probably (none / 0) (#75)
    by coigue on Wed Sep 10, 2008 at 10:31:54 AM EST
    never be blue, but Athens and other places with Universities may be the key. (Ohio has many small University towns)

    Parent
    The 1+ for Obama is ... (5.00 / 0) (#8)
    by Robot Porter on Wed Sep 10, 2008 at 09:33:07 AM EST
    with leaners.  Without it's even.

    While the 1+ from the eighth, the biggest bounce Ras gave McCain, was both with AND without leaners.

    I'm not buying that the bounce is fading based on the Ras polls, since they never really saw a bounce.

    Um (5.00 / 2) (#12)
    by bluegal on Wed Sep 10, 2008 at 09:35:24 AM EST
    What do you call having Obama go from 51% (his highest ever) down to a tie of 47-48%

    That would be what we call a bounce.

    Parent

    Of course you're right ... (5.00 / 0) (#27)
    by Robot Porter on Wed Sep 10, 2008 at 09:47:01 AM EST
    my brain isn't working too well this morning.

    But either way Ras isn't showing much movement this week.

    Nothing to rejoice about.

    Parent

    Democrats (5.00 / 1) (#9)
    by KeysDan on Wed Sep 10, 2008 at 09:33:26 AM EST
    often seem a little slow in the response department to Republicans.   Hopefully, enough time has now passed to reclaim discussion of the germane issues facing the country.  In addition, with each day it may become clear that knowing more about the McCain/Palin ticket is not loving them more.  

    Well, this is something I feel a little more (5.00 / 1) (#15)
    by Militarytracy on Wed Sep 10, 2008 at 09:38:55 AM EST
    up to date on.  Listen up Obama!  I know you have to put up the slugging back ads but you have to put some ads up there making some promises dude to fix the swack.  Right now it only seems like you are saying, "I'm not as bad these guys"  Not many people are.  In Birmingham I saw a bumper sticker.....AlObama.....live up to it my man.

    I don't think it matters much (5.00 / 2) (#19)
    by Jlvngstn on Wed Sep 10, 2008 at 09:40:54 AM EST
    the polls are going to bounce up and down for the next 60 days as there are a ton of undecideds out there.  As the economy sinks over the next 60 days it will be interesting to see if my theory holds up, that the stimulus package approved by the democratic congress which has clearly helped the republicans as the layoffs that would have happened in June/July start happening now.  Foxnews, McCain et al are still saying that the economy is blipping and not a recession, most Americans feel differently.  

    Sarah Palin who is not the focus of the election but is the cause of the bounce is not wowing the crowds with the same stump speech.  Her interview with Gibson is over 2 days which means that it will be a puff piece and I expect another bounce for McCain after the interview.  Obama needs to avoid discussing her but it is fair game for Biden.  

    Two-three weeks post Gibson when Palin is campaigning on her own and fielding questions from reporters, her star will diminish a bit.  The economy will become a bigger concern. Independent voters who watch fox will become tired of their incessant claims that the economy is due for a rebound out of its "funk"  and will identify or translate change as democratic and not maverickal.  

    McCains' numbers will start to decline then and the registration efforts and behind the scenes efforts of reaching actual voters as opposed to sound byting will gain traction.  Ohio is most likely lost although I think O as a result of HRC's efforts will win Florida.

    4 years ago the republicans stayed on message but more importantly got the vote out, which is what I believe will be the changer in this election.  O and his team will get the vote out not because of sound bytes but because of the behind the scenes work.  Dems have registered 2 million new voters and may be counting on the younger generation to get out and vote a bit too much, but my guess is that we will see a far more younger voters in this election comparatively speaking to 04 and 00.

    4 years ago Kerry put me to sleep and the entire election was a big fat bore, at least we have a fun race this year and I am enjoying it for a change.  

    I am glad you're having fun (5.00 / 1) (#57)
    by coigue on Wed Sep 10, 2008 at 10:14:49 AM EST
    my anxiety is through the roof.

    Parent
    the bump is what it is (none / 0) (#70)
    by Jlvngstn on Wed Sep 10, 2008 at 10:25:50 AM EST
    and we have yet to hear Palin speak regarding policy.  Palin is an exciting candidate and there will be another bounce after Gibson wraps her in a flag and does a Larry King - Barbara Walters piece on her.  When she begins to campaign on her own she will inevitably make a similar mistake like she did with fannie/freddie.  

    Too many people are drawing comparisons to Quayle in this regard without taking into account that Quayle did not add a bounce, did not add life to the campaign and did not inject adrenaline into Bush.  Palin did all of the above and then some.

    I do not expect her to be as miserable as Quayle as I think she is pretty sharp and not prone to drone on like he did.  I do expect her to make enough mistakes that it will dim the star, fade the enthusiasm and turn the polls enough to question her ability to step in.

    More importantly, economics will play a much bigger role when the layoffs start and the economy stalls.  I was awed and that is not an understatement, at the ground game in 04 by the republicans and I think axelrod and crew paid close attention to that and are playing that game.  

    The question is when will Palin have reached her apex and I would argue that she already has the bump coming from Gibson notwithstanding.  Momentum is a funny thing and carrying that new and exciting label is awfully hard, just ask Obama who has been completely eclipsed for the time being by a VP candidate.  Problem with that is, people are not excited about McCain.  If there is not excitement about the top of the ticket it simply cannot hold.

    Parent

    how Palin does in her first interview (5.00 / 2) (#89)
    by TimNCGuy on Wed Sep 10, 2008 at 11:33:34 AM EST
    will mean a lot.  If she comes up with good answers for the questions about the "Bridge to NoWhere" and also strongly addresses the lies about her positions that have been put out there like lowering funding for special needs children when she actually increased the funding, I think it will help her and McCain.  And, if she is able once again to get in some shots at Obama, it will help again.  But, if it turns into a fluff interview by Gibson, I think it will hurt her bacause everyone has been calling for an interview with serious questions.

    Parent
    I appreciate your insight (none / 0) (#76)
    by coigue on Wed Sep 10, 2008 at 10:37:30 AM EST
    I have thought alot this year about how Dems seem to peak too early in elections. Maybe this year the GOP will peak too early???

    I agree with you about FLA, but I also think that Palin hurts McCain there...so it's all good.

    Hil is a champ for campaigning there.  

    Parent

    HRC (5.00 / 1) (#78)
    by Jlvngstn on Wed Sep 10, 2008 at 10:41:48 AM EST
    has clearly demonstrated that she is the strongest party advocate, I could not do what she is doing.  She is a better person than I could ever wish to be....

    Parent
    No kidding. (none / 0) (#82)
    by coigue on Wed Sep 10, 2008 at 10:48:35 AM EST
    She really does believe in the party and the platform.

    And it shows.

    Parent

    and I am an idiot (none / 0) (#81)
    by Jlvngstn on Wed Sep 10, 2008 at 10:47:58 AM EST
    with a keyboard, internet access and an opinion.  I hope I am right but conjecture is the lifeblood of blogs and it is full of analysis that is above my pay grade.  The first test to my theory will be after Gibsons interview, 4 weeks after that I will know whether or not I am just a flake like most of the online posters or whether or not I was right like I was about wmd, foreign relations consequences as a result of the Iraq war, O winning the primary and the economy.  All of which I accurately predicted on this very site.  

    That said I also predicted O would win by 8 points or better.  I would like to see more predictions by the bloviators here so that we can actually see whose analysis is based on heart vs mind....

    Parent

    I predicted the WMD and war stuff too (none / 0) (#84)
    by coigue on Wed Sep 10, 2008 at 10:52:12 AM EST
    but because I listened to Scott Ritter and the other weapons inspectors.

    Too bad more legislators did not.

    sigh.

    Anyhow. My prediction is currenlt that Obama will lose. I predict that because the GOP almost always wins the presidency, they are good at it, and they appeal to the culture of most of America, apparently.

    I hope to God you are right and I am wrong.

    That's why I am so anxious.

    Parent

    as i predicted (none / 0) (#94)
    by Jlvngstn on Wed Sep 17, 2008 at 09:56:14 AM EST
    The anxiety and bed-wetting that many Democrats have been experiencing about the state of the presidential race over the past few weeks should be subsiding as the Palin effect is wearing off and we're seeing clear signs that the McCain bounce is over and the pendulum is swinging back to Obama. Effectively, the race has reverted to where it stood before the political conventions.

    Today's four daily tracking polls show a clear trend towards Obama over the past four days. Since state polls tend to be lagging indications -- and they're much more dated -- we should see those numbers settling back in the next week.

    Democrats may now exhale. Here's the data for the past four days of tracks:

    Parent

    The DNC is targeting female voters (5.00 / 0) (#25)
    by Fabian on Wed Sep 10, 2008 at 09:43:24 AM EST
    here in Ohio.

    Unfortunately, they are using the "McCain s*cks on women's issues." tropes.  In fact, everything we've gotten so far is "McCain S*CKS - vote Democrat!".

    Sigh...

    Strong, Positive Brand.  Okay DNC?  Are you listening?  "Not-McCain" is NOT a Strong Positive Brand.

    My second point is Own The Issues.  Let's see if I get any mail about The Issues or if they'll miss the boat on that too.

    And when they send out (5.00 / 1) (#43)
    by tootired on Wed Sep 10, 2008 at 10:01:09 AM EST
    the "McCain s^cks on women's issues" message, many women are yelling back at the TV "and you do, too". Obama needs to get his message out about where he stands on issues that matter to women and their families. And no "just go to my website". And no "it's above my paygrade" answers. He also needs to talk about the role that women will play in an Obama administration. We already know what role they'll play in a McCain White House. Obama has thus far sacrificed a lot of his potential female vote to get the votes of the sexist males. That's not all men, but it's the ones that live on the you know where blogs. It's time to make up some ground if he still can. And Hillary simply cannot do this for him. She's credible on these issues, but he's not. Women will flock to hear Hillary speak because we love her, but it won't change how we vote.

    Parent
    The Dem platform is strong (5.00 / 1) (#55)
    by coigue on Wed Sep 10, 2008 at 10:13:55 AM EST
    and supports people. We should lift that up for people to mull over.

    Obama supports that platform, and you are right, he needs to sell it.

    Parent

    Not McCain (none / 0) (#31)
    by Ga6thDem on Wed Sep 10, 2008 at 09:49:52 AM EST
    has run it's course. I think that picking Palin undid the Bush=McCain narrative.

    They are continually telling the voters that there's no reason to vote for Obama.

    Parent

    Ironic (none / 0) (#38)
    by coigue on Wed Sep 10, 2008 at 09:56:10 AM EST
    since Palin reminds me more of Bush (2000) than McCain does.

    Parent
    Me too (none / 0) (#48)
    by Ga6thDem on Wed Sep 10, 2008 at 10:07:13 AM EST
    However, the fact that shes a woman probably alters that perception somewhat.

    Parent
    that's sexist (none / 0) (#52)
    by coigue on Wed Sep 10, 2008 at 10:12:16 AM EST
    (snark)!

    But as a woman, I do have to bristle at the idea that another woman can't be as much of an idiot as a man (not that that is what you were saying, but others like to put women on a pedestal of middlin': not too good, but not too evil either)

    Parent

    Many Republican voices (none / 0) (#66)
    by tootired on Wed Sep 10, 2008 at 10:22:10 AM EST
    are suggesting that she is the new Ronald Reagan. Richard Viguerie and Michael Reagan among others. I think that Democrats are under-estimating her appeal to Republicans in general. "Mother, governor, moose shooter" probably isn't going to be enough to knock her out of play. Obama needs a new game plan.

    Parent
    She speaks better than Bush (none / 0) (#67)
    by nycstray on Wed Sep 10, 2008 at 10:22:23 AM EST
    and doesn't seem afraid of hard work. They may have some views the same, but I suspect she governs much different then him. And you can bet yer behind you won't hear her whine "it's hard work . . ."

    Parent
    I suspect she governs the same. (none / 0) (#80)
    by coigue on Wed Sep 10, 2008 at 10:47:08 AM EST
    According to Molly Ivins, Bush's MO as governor was to cut funding for services while speaking about his "support" for those services.

    Palin asked for earmarks and spoke against them at the same time.

    Bush governed (as Pres) often from his Crawford ranch, Palin governed from her home in Wasilla.

    Bush has strong support of the evangelicals and thinks God advises him. Pailn (apparently) thinks the same.

    Bush uses folksy language, so does Palin. People are saying they want to have a beer with her...sound familiar?

    She surely has a better work ethic than Bush, but she still needs to be sequestered and taught about the issues. Remember Condolissa Rice's foreign policy "briefings" in 2000? Same thing.

    She appeals to the same people that Bush did, and she offers them the same thing, with the added factor that she worked for everything she has and he was handed everything. That is to be respected.

    Her views, atrocious.

    Parent

    You and I may see it that way (5.00 / 1) (#87)
    by nycstray on Wed Sep 10, 2008 at 11:21:48 AM EST
    but will enough voters? I gave you a shorthand impression above of why Palin=Bush may not sink in with people and it isn't just gender :) I think her views are atrocious, but I wouldn't mind having a meal with her. Preferably in Alaska  ;)

    Parent
    I should clarify my governs comment (5.00 / 1) (#88)
    by nycstray on Wed Sep 10, 2008 at 11:32:43 AM EST
    sorry, jumping back and forth with work :)

    Governs in terms of actually working. I think she gives the impression she would put more into it than Bush. They just don't "look" the same in "presentation". She comes off much different. That is the image people are getting of her. Her first impressions were very good. How many voters are actually sitting back thinking about what MI said about Bush governing TX right now?

    I think that's why Obama needs to clarify policy differences loudly and often. It's how he can de-shine her along with McCain without even mentioning her.

    Parent

    I gotcha. (none / 0) (#90)
    by coigue on Wed Sep 10, 2008 at 11:46:04 AM EST
    it's perception v reality.

    i agree.

    but in reality, I think she is even more extreme Bush - with a work ethic.

    and she scares the sh*t out of me.

    Parent

    National tracking polls make little sense (5.00 / 2) (#61)
    by TimNCGuy on Wed Sep 10, 2008 at 10:17:38 AM EST
    to me at all.  It makes no difference at all what the demographics say in solidly blue or red states about either McCain or Obama.

    Show me state polls for the swing states and do it a little more frequently than every two weeks.

    Don't worry, (none / 0) (#62)
    by andgarden on Wed Sep 10, 2008 at 10:19:35 AM EST
    We're about the get plenty of state results. Finally. . .

    Parent
    When does Gallup Daily post ... (none / 0) (#2)
    by Robot Porter on Wed Sep 10, 2008 at 09:22:50 AM EST
    their number?

    They've had McCain ahead by 5% for the last two days.

    They posted a little while ago (none / 0) (#91)
    by Marvin42 on Wed Sep 10, 2008 at 12:26:06 PM EST
    No change from yesterday, still 5% gap, both candidates down 1%.

    Parent
    Ras weights for party (none / 0) (#5)
    by andgarden on Wed Sep 10, 2008 at 09:29:29 AM EST
    Which means that they never saw a huge bump in the first place.

    I've decided that Ras's methodology is junk for these purposes.

    Ras methodology (5.00 / 1) (#26)
    by frankly0 on Wed Sep 10, 2008 at 09:45:31 AM EST
    certainly presents a real issue because it weights for party id, when it is quite possible party id numbers are themselves shifting in a real way.

    Of course, if you don't correct for party id, you may also be overrepresenting certain groups.

    It's not obvious which is "fairer", and mostly it depends on unknowns. I'd guess that Ras. general methodology would be to shift the party id numbers over time if they stay in the same neighborhood -- which is not an unreasonable compromise. Even if it doesn't give the best possible snapshot, it probably gives a more reliable long term measure of the state of the campaign.

    Parent

    Problem is that party ID is an attitude, (5.00 / 1) (#29)
    by andgarden on Wed Sep 10, 2008 at 09:49:39 AM EST
    not a demographic category.

    When you weight for party, you're substituting your own opinion for the data you've collected.

    Weighting beyond the census is unsound methodology.

    Parent

    Not necessarily ... (none / 0) (#42)
    by Robot Porter on Wed Sep 10, 2008 at 10:00:56 AM EST
    weighing only for census might miss a clear sign of movement if you're getting high non-answers from that group.

    One of the reasons for weighing results is to take into account problems in the randomization method.

    Parent

    yes necessarily (none / 0) (#45)
    by andgarden on Wed Sep 10, 2008 at 10:05:06 AM EST
    What is the basis for weighting to party? Past election results? Well, which election? Every Presidential election in recent memory has had a different D/R/I breakdown. Ras polls separately for party ID, and weights monthly. But that's just as stupid. Who's to say that your poll for party breakdown is any more reliable than your presidential tracking poll? And why would you introduce the noise of last month's party breakdown into today's tracking poll?

    Parent
    I'm not saying Ras's choice ... (none / 0) (#53)
    by Robot Porter on Wed Sep 10, 2008 at 10:12:37 AM EST
    is the best one.  But any weighing of results is a recognition of the fact that you think you have some bad data.

    Weighing ONLY for census will often just amplify that bad data.

    If you have another controlling factor you may be able to decrease that problem.

    Parent

    I think the problem is not so much (none / 0) (#54)
    by frankly0 on Wed Sep 10, 2008 at 10:12:48 AM EST
    inconsistency with election results.

    It's inconsistency between polls, or inconsistency with well known facts about party id.

    Only too often, for whatever reason, one party or another is overrepresented. The numbers themselves make that glaringly obvious. People who look at the polls immediately see the problem, and, correctly, discount the results of the poll on that basis.

    Ras adopts one particular strategy to deal with these anomalies. It seems not unreasonable -- as would the opposite strategy.

    Parent

    This continues not to make sense to me: (none / 0) (#60)
    by andgarden on Wed Sep 10, 2008 at 10:16:52 AM EST
    Only too often, for whatever reason, one party or another is overrepresented.
    My question: sez who? Couldn't I just as reasonably argue that support for a particular candidate is "over represented" in a poll because I "know" that some other candidate should be ahead? Why even bother polling at all if you're just going to massage the results to come up with an answer that you think looks right? Miss Cleo would be a little cheaper and probably more entertaining.

    Parent
    All weighing of results ... (none / 0) (#69)
    by Robot Porter on Wed Sep 10, 2008 at 10:24:27 AM EST
    is massaging data.

    Opinion polling, especially daily tracking, isn't an exact science.

    There's some art in there too.

    In actual scientific statistical analysis you will usually have a control group to solve for the kinds of anomalies you see in opinion polling.

    The introduction of various types of weighing is an attempt to insert something like this in an opinion poll.

    Parent

    Don't you think ... (none / 0) (#63)
    by eustiscg on Wed Sep 10, 2008 at 10:19:43 AM EST
    the margin of error from their party ID poll is smaller than the sum of potential errors in response and randomization?

    Parent
    Why should it be? (none / 0) (#65)
    by andgarden on Wed Sep 10, 2008 at 10:21:55 AM EST
    Are they asking more people about their party ID than their Presidential preference?

    Parent
    Do you know ... (none / 0) (#73)
    by Robot Porter on Wed Sep 10, 2008 at 10:27:25 AM EST
    what a randomization error is?

    Parent
    My understanding is (none / 0) (#79)
    by andgarden on Wed Sep 10, 2008 at 10:42:51 AM EST
    that it's concentrations of important variable/attributes in your sample data that you aren't aware of.

    Why should there be a difference between party ID and Presidential preference?

    Parent

    It isn't the difference ... (none / 0) (#86)
    by Robot Porter on Wed Sep 10, 2008 at 11:20:12 AM EST
    between Party ID and Presidential preference, but between Party ID and sampling errors.

    Parent
    It generally should be ... (none / 0) (#72)
    by Robot Porter on Wed Sep 10, 2008 at 10:26:24 AM EST
    much smaller.

    Parent
    Well, the problem would seem (none / 0) (#46)
    by frankly0 on Wed Sep 10, 2008 at 10:05:16 AM EST
    to be that, when data is in fact collected, there's a lot more variability along party id lines than seems to be warranted, even after correcting for all demographic features. It seems often to happen in polls which don't correct for party id that there is an obviously anomalously high or low representation of a given party. Why should that happen? I certainly don't know. But it does.

    So how does a pollster respond? Try to correct for those obviously anomalous results? Refuse to do so? Whatever one does, it seems one is pretty much obliged to hold to a consistent policy.

    As I said, the probable consequence is that snapshots may be skewed by the weighting, especially under circumstances in which one would expect party id to vary (such as after the conventions). But if the party ids hold up over time, and the numbers for party id are therefore shifted in any further weighting, then the more important trends are preserved.

    How important is it, after all, to know the exact size of the bounce (esp when its measure might be affected by significant and random -- not real -- variability in party id), if the "party id" numbers are themselves short lived?

    How to handle that seems to me to be a policy decision that could go either way.

    Having said all that, I do think it very likely Ras is in this instance significantly underestimating McCain's actual bounce. I think that Gallup's methodology gives a fairer picture, because I think party id has very likely, in fact, significantly changed after the conventions.

    Parent

    Problem is, using your argument for weighting, (none / 0) (#49)
    by andgarden on Wed Sep 10, 2008 at 10:08:00 AM EST
    tracking polls themselves are a sham. They often show "obviously anomalous results," do they not?

    Parent
    Yes, they do (none / 0) (#56)
    by frankly0 on Wed Sep 10, 2008 at 10:14:10 AM EST
    but the approach of Ras at least has the virtue that it tends to tamp them down, rather than exaggerate them.

    Parent
    Now, with the bounce fading somewhat (none / 0) (#16)
    by tigercourse on Wed Sep 10, 2008 at 09:38:56 AM EST
    the convention hoopla over and 50 odd days to the election, would be a good time for Obama to go on the offensive against MCCAIN. Talk about how he's too unstable to be trusted with the red button, point out each and every one of his bad policies and his hypocrisies/flip flops. Go after MCCAIN.

    Now, with the bounce fading somewhat (none / 0) (#34)
    by AvianoTeamB1 on Wed Sep 10, 2008 at 09:53:17 AM EST
    You hit the nail right on the head Obama needs to take off the gloves.

    Parent
    I officially declare (none / 0) (#50)
    by Faust on Wed Sep 10, 2008 at 10:09:23 AM EST
    by powers not invested in me, that we may now view polls as having significance from here on out.

    I'm seeing chins down (none / 0) (#58)
    by Lahdee on Wed Sep 10, 2008 at 10:15:12 AM EST
    from some friends and colleagues who were early and avid Obama supporters. The devil in me wants to kick them in the head, but the practical, pragmatic me seeks to buck them up. It's a difficult fight when facing republicans I advise. They smile softly.
    I can almost see the tears.
    Gosh, I hope they don't give up.

    Best news I've heard in awhile. (none / 0) (#92)
    by Southsider on Wed Sep 10, 2008 at 12:27:34 PM EST


    Polls. (none / 0) (#93)
    by AvianoTeamB1 on Wed Sep 10, 2008 at 01:29:05 PM EST
    Frankly speaking I think this tracking polls thing should be outlawed altogether.
    This kind of information can some times manipulate voters into voting for the wrong candidate.
    Just imagine what a bias news network like Fox News can do with such information?