A new Zogby telephone survey of 1,000 likely voters (conducted 1/24 - 1/26) finds 32% approve of Bush's job performance; 68% disapprove.
By Eric Dienstfrey on January 26, 2007 3:50 PM | Permalink | Comments (0) | TrackBacks (0)
Additional analyses from recent Gallup national surveys find:
By Eric Dienstfrey on January 26, 2007 1:15 PM | Permalink | Comments (1) | TrackBacks (0)
In addition to their new national survey that we posted on yesterday, Time is debuting a new "Election Index" feature. In a related piece on the surprisingly early start to the election season, Time's Karen Tumulty describes it as follows:
To track voter sentiment--and candidates' odds of winning--TIME is launching the Election Index, a regular feature that will pinpoint the intersection of how much Americans know about each candidate and how much they like what they see. The surprising news is that this week's Election Index puts former New York City mayor Rudy Giuliani ahead of Arizona Senator John McCain, despite the latter's formidable organization and resources, for the top spot in the G.O.P. Hillary Clinton leads the Democrats, but the Election Index (see page 34) shows she has slightly less potential general-election support than Giuliani.
The index (shown graphically for Democrats and Republicans) appears to combine survey measures data other than the traditional horse race question (the percentage who have never heard of a candidate and, among those who know the candidate, the percentage that are favorable, will definitely support that candidate or think they will win) with non-survey measures (betting odds and number of blog posts linking to a candidates web site). I say "appears" because Time does does not tell us exactly how they constructed the index nor what the precise score is for each candidate. The graphic rank orders the candidates -- and Tumulty's piece tells us that
Hopefully more details will follow. Meanwhile, the brief introduction of the feature makes this entirely valid point about the shortcomings of early trial heat results:
Pundits and bloggers would do well to tape a copy of these poll results, from ABC-Washington Post in January of 2003, on their monitors/mirrors/most viewed flat surface in their office:
Joseph Lieberman 27 Richard Gephardt 14 John Edwards 11 John Kerry 10 Al Sharpton 7 Howard Dean 3 Don't know 24
Looking back at the last open GOP race, over a year and half out from election day, the results are less shocking than the prospect of a Lieberman-Gephardt ticket, but they do tend to support the hypothesis that early polls aren't so much even popularity contests as trivia quizzes (The question being, "Have you ever heard of this person before?")
By Mark Blumenthal on January 26, 2007 11:00 AM | Permalink | Comments (0) | TrackBacks (0)
A new Public Policy Polling automated survey of 501 likely voters in North Carolina (conducted 1/22) finds:
By Eric Dienstfrey on January 26, 2007 10:09 AM | Permalink | Comments (0) | TrackBacks (0)
Hillary Clinton and Rudy Giuliani are doing better in national polling than in the critical early primary and caucus states. Clinton leads all national Democratic polls, while Giuliani leads McCain in 33 of 43 national polls. But neither candidate does as well when we turn to the polling from the early voting states. Conversely, John Edwards does considerably better in all but two state polls than he does nationally.
Clinton's national support is significantly higher than it is in Iowa and New Hampshire. Nevada and South Carolina (with only 3 polls together) are just about equal to her national standing.
Giuliani has seen some convergence in state and national polls, though his current national support is still above support in three of the states, IA, NH and SC.
The sharpest contrast is with Edwards who does considerably better in IA, NH and SC than he does nationally. Nevada is again the state closest to his national standing.
McCain, Obama and Gingrich show the states spread around their national numbers. McCain in particular has seen some apparent decline in NH over the past 12 months, bringing initially favorable NH number down to his national levels. South Carolina with only 4 polls appears to have slipped a bit as well, but remains well above the national. In Iowa, McCain does no better but also not much worse than he does nationally. The trend in NH must, of course, be of concern for his campaign. Likewise, McCain's national support has been essentially flat over the past year.
In Obama's case, there are relatively few polls so trends are not yet clear. There is not a wide gap between state and nation, though the national is a bit above most of the state polls.
Gingrich is still well behind Giuliani and McCain and his state polling if pretty evenly distributed above and below the national trend. Given the limited nature of his campaign so far there have been few events to affect these trends.
As we all know, the early primary states are not a representative sample of the country, so divergence of state from national polls is to be expected. But the political consequences of doing well nationally but less well in these four states can be profound. We'll keep an eye on these trends.
Cross-posted at Political Arithmetik.
By Charles Franklin on January 26, 2007 9:19 AM | Permalink | Comments (1) | TrackBacks (0)
A new Rasmussen Reports automated survey of 800 likely voters (conducted 1/22 and 1/23) finds:
By Eric Dienstfrey on January 25, 2007 1:53 PM | Permalink | Comments (0) | TrackBacks (0)
Back to basics for a moment, to a point we cannot make often enough. Yes, polls report a "margin of error," which derives from the random variation that comes from drawing a sample rather than interviewing the full population. But that is all it is. The "margin of error" statistic is not an overall measure of quality and can tell you nothing about the potential for other sources of error, including as a bias from missing certain voters (such as those who lack a landline telephone or who hang up when called) or from asking questions in a way that produces misleading results.
A friend sent this take on sampling error from Good Math, Bad Math, a blog by a computer scientist:
People frequently believe that the margin of error is a measure of the quality of a statistic - that is, that a well-designed poll will have a smaller margin of error than a poorly-designed poll. It doesn't - the MoE only represents sampling errors! A great poll with a sample size of 100 will virtually always have a considerably larger MoE than a terrible poll with a sample size of 1000. If you want to know the quality of a poll, you need to know more information it than just the margin of error; If you want to gauge the relative quality of two different polls, you need to know more than just the margin of error. In either case, you really need to know the sample size, how the sample was collected, and most importantly exactly what they measure.
It's worth a click.
By Mark Blumenthal on January 25, 2007 1:43 PM | Permalink | Comments (3) | TrackBacks (0)
A new Cronkite-Eight statewide telephone survey of 600 registered voters in Arizona finds:
By Eric Dienstfrey on January 25, 2007 1:03 PM | Permalink | Comments (0) | TrackBacks (0)
A new Time national survey (story, analysis, results) of 1,064 registered voters (conducted 1/22 and 1/23) finds:
By Eric Dienstfrey on January 25, 2007 12:00 PM | Permalink