Comments

  1. dave schutz says

    Grasping at straws. Not that anything means much just now, check again at the end of the month. We are just past Labor Day, for godsake. Look for 'likely voter' polls.

  2. SP says

    When they do tracking poll plots like they they should really make the thickness of the bars equal to the MOE so you can tell when a difference is p<0.05. Otherwise you might as well disconnect the cable from your TV and pretend you can make out an I Love Lucy rerun in the static.

  3. Brett Bellmore says

    Has Gallup started adjusting for likely voters yet? I don't believe so. Remember, according to their own website, those numbers aren't worth squat until they do. And Rasmussen, which IS so adjusting, puts the numbers at 48% to 36% in favor of the Republicans.

    Further, Gallup themselves in this poll say that levels of Republican enthusiasm are at historic highs, (While Democrats are going into the election bummed out.) which suggests that turnout might be even more tilted Republican than usual.

    I suppose it's a straw, so you CAN grasp at it if you like. But I wouldn't bet any real money based on this.

  4. says

    Nate Silver, the star augur for reading poll entrails, pointed out http://fivethirtyeight.blogs.nytimes.com/2010/08/… at the time that the earlier Gallup poll was an outlier – relative to a bad baseline for the Democrats.

    The straw to grasp at is perhaps that the fluctuations suggest a genuine volatility or confusion in voters' minds. They trust Republicans even less than Democrats, but seem prepared to vote them back in anyway. Weird. "Throw the rascals out" irrationally trumps "better the devil you know".

    On the face of it, I think you are in for a very bad two years of Tea Party screeching. At any rate having to use the veto every week may cure Obama of his bipartisan dreams, and force him to bend his considerable talents to the political destruction, not taming, of his irreconcilable enemies.

  5. says

    Both. It's good news insofar as it makes the noisiness of the poll obvious, and gives us an idea of the error bars. Is there a sudden trend? ha.

  6. Bernard Yomtov says

    Gallup says they surveyed 1650 voters. So MOE is about 2.5%. These numbers are very volatile, it seems. Six or seven weeks ago it was D-49 R-41. Before that it was running even.

    I'm not sure how to interpret fact that these results are, in each case, an aggregate of a week's worth of daily polling. It would be interesting to see the moving average of seven daily results.

  7. Brett Bellmore says

    "Gallup says they surveyed 1650 voters. So MOE is about 2.5%."

    That's a lower bound on the error bar, assuming that they didn't have to do any renorming at all, assuming their response rate was 100%, and so on, and so forth. The likelihood of which I'd put somewhere between little and none.

    Realistically, the error bar on a poll like this is probably more like 5-6%.

  8. rachelrachel says

    @Wimberley says: "They trust Republicans even less than Democrats, but seem prepared to vote them back in anyway."

    Or, more likely, many of them when there's a Democratic president, prefer a Republican Congress to keep him in check.

  9. Sargon says

    "Or, more likely, many of them when there’s a Democratic president, prefer a Republican Congress to keep him in check."

    That seems utterly redundant, as Democrats do a very good job of keeping themselves in check. Majorities in both houses and still can't get much done.

  10. Vladimir says

    A prediction: the Dems are going to sweep, because the country is going to wake up in early November and see the Republicans for just what they are. You heard it here first.