TalkLeft, among others, is pointing to this latest Washington Post-ABC News poll, which, according to the Post, purports to show that President Bush’s approval ratings are at an all-time low (a charge I heard repeated by a reporter this morning during the President’s brief news conference). Here’s how the WaPo frames it:
For the first time in his presidency, a majority of Americans question the integrity of President Bush, and growing doubts about his leadership have left him with record negative ratings on the economy, Iraq and even the war on terrorism, according to a new Washington Post-ABC News poll.
On virtually every key measure of presidential character and performance, the new survey found that Bush has never been less popular with the American people. Currently 39 percent approve of the job he is doing as president, while 60 percent disapprove of his performance in office—the highest level of disapproval ever recorded for Bush in Post-ABC polls.
Other recent polls—including those from CBS News and AP/Ipsos—show similar results, proof-positive that the Bush presidency is in trouble.
Or is it?
The New Editor’s Tom Elia takes a closer look at the demographic and political breakdown of the respondents of all three polls and concludes that all three are, well, cooked. From “More Fun With Polls”:
the CBS News poll, which shows that Bush’s approval rating is at 35%, reports that an unweighted sample shows that 34.8% of its respondents self-identified as Democrats, while 27.6% said they were Republicans. While the unweighted sample yielded a seven-point differential favoring the Dems, a weighted sample had the spread at 11% points in favor of the Dems. This represents at least a 10-11 point swing in the electorate since the 2004 election (and perhaps as much as 14 points), when Bush won by about three points and the Repubs won the aggregate House vote by about four points.
The AP/Ipsos poll, which has Bush’s approval rating at 37%, said its respondents self-identified 49% Dem and 40% Repub, yielding a nine-point differential favoring the Dems. This represents a 12-point swing in the electorate since 2004.
Only 80% of the respondents in this poll were registered voters, while 13% of the respondents reported that they were unemployed (the current unemployment rate is about 5%-6%), and 31% said they were between the ages of 18-34 (this was the only age group that Sen. John Kerry carried in the last presidential election, and it actually voted in a much smaller percentage than that represented in this poll).
The ABC/Washington Post poll, which shows that Bush’s approval rating is at 39%, saw its respondents break down thusly: 52% said they leaned toward the Dems, and 41% leaned toward the Repubs. That represents a 14-point swing since last year’s elections.
Will someone explain how the demographic breakdowns in these polls represent accurate pictures of the electorate, given what we know about the 2004 election (and the 2002 mid-terms)?
Good questions.
And the natural follow-up question is this: if there is no inherent liberal bias in the media, then how on earth can we account for what seems an intentional desire on the part of these news organizations and their polling organs to skew the logistics in order to put the Bush presidency in an unrepresentative light?
In an update to the original post, Tom notes:
The last time a Democratic Party presidential candidate received 52% of the vote, the percentage of Dems responding to Washington Post pollsters’ questions, was in the 1964 election when President Johnson beat Sen. Goldwater.
Well, that explains things, doesn’t it? I mean, who didn’t know that today’s Boomer-run legacy media wasn’t stuck in the heady days of the 1960s…?
What? Low poll numbers? Hope he doesn’t want to run again!
Because of the EXIT POLLS!
GW is in some serious trouble for his next election.
I think what some of the left commentators are missing is that the dissaproval numbers (as accurate as they are) have swung because Republicans are dissatisfied with Bush. There’s this missconception that some great reserve of independant Americans have risen up to voice their displeasure.
Was this poll taken during the Miers nomination?
While the post does a good job of pointing out the odd flaws in these polls…
What would be more convincing to me (about the point that these polls have sampling/weighting flaws) is a comparison to other past poll’s sampling/weighting characteristics where Bush’s fav/unfav ratings were around 50% or higher.
If those old polls had roughly the same characteristics as these polls then I would say that the dismissal of the current polls is akin to sour grapes. If they are different (better balanced) then that would be something. Of course if they are skewed the other way (favoring Rep’s etc) I would be suprised.
Has anyone done this sort of comparison?
This&That
The only poll that counts was taken Nov. 2, 2004.
TW: love. For the love of God, we aren’t in the ‘60s any longer.
There’s lots to chew on here, but Washington Post’s poll follows another pattern that argues against a simple liberal bias claim. There’s another category: Would you say your views on most political matters are liberal, moderate, or conservative? And in that category, as always with these Post polls, the breakdown skews conservative. In this one, 23% liberal, 44% moderate, 32% conservative.
Yeah. But Democrats are always insisting they are the real conservatives—John Kerry said as much—so that bit is hardly surprising.
It’s not reelection, it’s that his polls are approaching the “tipping point” where he doesn’t get to have a Library!
You just said ‘polling organs’. I mean, I was right on track, following it all; very interesting as usual. Then I ran into the phrase ‘polling organs’. Sorry, but I lost my shit there.
Shank, That’s almost as bad as yesterday’s AJC headline, ”Illegal wood probe…” er, something something…
I did the math on the CBS poll data; adjusting the number of respondents to roughly equal (x1.3 for GOP, x.9 for Dems), Bush polls at 39%. Which is probably about right.
936nweighted% app by grp# app by goupnew weightnew approve
r259223771.71289.9223.223
d326326115.86293.432.274
i351388310.28352.7109.337
total936937327.85364.834
358.98%
Don’t know if that will post in a legible format.
Update: It won’t.
n,weighted,%, app by grp,# app by goup,new weight,new approve
r,259,223,77%,171.71,289.9,223.223
d,326,326,11%,35.86,293.4,32.274
i,351,388,31%,120.28,352.7,109.337
total,936,937, 100%, 327.85,936,364.834
old 35&#xne;w 38.98%
Alright…
Who’s calculator just vomited everywhere?
Why doesn’t anyone ever note that the only people home during the workday and willing to talk to these idiots are not the productive, working Americans who voted for W in the first place?
“Why doesn’t anyone ever note that the only people home during the workday….”
1) Who says that the phone calls are placed only during working hours? I can’t find any info. on this one way or the other.
“…and willing to talk….”
2) That is a good question. How many have caller ID, don’t pick up, or hang up on anything that is not a personal call?
This&That
darn it, vw beat me to it.
Call up a bunch of random people, and the people who are home are more likely to be unemployed, so that makes sense. The people with jobs are out working and have too many important things to do than talk to pollsters.
Heck, I’m never going to be polled – I don’t have a home (land line) phone, just two cell phones. The number of people who don’t have phones alone has to skew the results of any poll significantly.
AR, Thar she blows! The ghost in the machine!
Party identification is flexible, unlike more fixed demographic characteristics such as sex and ethnicity. So, some portion of survey respondents will report party leanings based upon their feelings about the president at the time of the survey. IOW, when Bush was riding high, those people would say they were Republicans or lean toward the Republicans; now, with Bush doing poorly, those same people say they are Democrats or lean toward the Democrats. (I put the number of people who flexibly report their party leaning at 5 to 10 percent of survey respondents.)
Actually, with the 3X normal percentage of unemployed in the survey, that would argue for a daytime sample.
But the larger point here, besides the obvious agenda journalism, is that liberals are obsessed with the idea that they are right. It reduces to pure narcissism: they love themselves and their ideas are one with themselves. Therefore, if one rejects their ideas, one rejects them. Thus we have the continual efforts to find the silver lining in every conservative thunderstorm over their heads and rationalize why they should be happy and take another toke from the bong of liberalism.
Bad weed, man.
Wake me up when they repeal the 22nd Amendment. Oh, and let the baby have its pacifier. No one else wants it anyway.
Was there evidence of this type of poll cooking while Clinton was getting low numbers?
Eh, same deal with the Iraq War. The left thinks he’s too harsh, the right thinks he’s to easy. Both are disappointed. Just goes to show there’s not much point in trying to please everyone. Get it done, damn the torpedoes.
figure(s)
Yiz can split hairs all you want, but don’t miss the big picture: growing numbers of Americans think the Prez is a creep. Funny, that.
Bee–
We’re not “splitting hairs,” we’re fundamentally rejecting these polls as seriously flawed and unrepresentative.
P.S. I think you are a creep.
Don’t you miss the big picture, Beetrot. Growing numbers of Americans are looking forward to Bush’s successor, be she Condi or any other Republican, except McCain. Funny, that.
I was just informed by CNN’s Dana Bash, as she introduced her story on Bush, that this is a Presidency In Crisis.
All we need is to give it a color scheme, a font, and some theme music and Bush’s poll numbers could dive to 25%.
Josiah Bartlet has a job approval rating at a stratospheric 75%.
Santos 59
Vinick 29
Zogby
It is readily apparent that the left has been spending too much time smoking poll. Aren’t these the same polling organizations that called dems victorious in the 2000 pe exist poll and for Kerry all the way up until election day?
Regardless, the last seven Presidents have had historic lows in the same areas as what they are reporting for GW as some time or another during their terms, even Clinton.
What does this all mean if you are a conservative? Nothing, nada, zip. Other than the fact that the left can conjure up some fantasy reality where conservatives are all extremist wing nuts it amounts to little more than a rough patch in what will likely be a long 3 years for the dems. If you are a lib and you’ve been touting these polls it means you are a moron (but hey, we already knew that).
To quote a lib if they were conservatives
America will lose everytime any President decides to govern based on the popularity of an idea. Fortunately there is an abundant collection of examples of success won in-the-face of popularity polls e.g. Kennedy’s decision to support Civil Rights, Nixon’s China Policy, and the list goes on… Steve