Pollster: We Did Not Over-Represent African Americans In Asking About Wright
As Josh noted last night, many of you readers have been writing in to say that the new NBC/WSJ poll oversampled African Americans in measuring voter reaction to the Wright controversy.
Now the pollster who did the poll has produced a new memo discussing the question and clarifying why it is that African Americans are not over-represented in their national sample.
Read it here.
Late Update: I've changed the headline above because the initial one was inaccurate.
Advertisement
Comments (18)
Guess the results can no longer be spun as bad sampling.
Waaaaah!
March 27, 2008 1:37 PM | Reply | Permalink
It was never legitimate spin to begin with.
I read the original story about the poll mentioning the black "oversamples" yersterday and I knew what it meant -- they polled more blacks, but then they weighted the results to make the final numbers fair and representative. The oversampling was simply to improve the MOE among the black sub-sample (which is normally astronomical -- NEVER trust a poll that comments on what "black people think" unless they did an oversample!)
But I also knew ignorant people would go ballistic, not knowing that the results are weighted afterwards to insure the correct outcome.
March 27, 2008 4:14 PM | Reply | Permalink
Last week when Obama was down in Gallup but ahead in Rasmussen we got the Gallup numbers and not the Rasmussen ones. Today, with Obama ahead in Gallup but losing in Rasmussen we get the Rasmussen numbers.
Fair and balanced?
March 27, 2008 1:37 PM | Reply | Permalink
There goes Craig Crawford's, that Hillary shill, primary talking point.
March 27, 2008 1:44 PM | Reply | Permalink
Why?
March 27, 2008 1:47 PM | Reply | Permalink
Why is this necessary?
March 27, 2008 1:47 PM | Reply | Permalink
because the sample size was too small. come on, it's not that complicated
March 27, 2008 2:01 PM | Reply | Permalink
Just another example of how the Media discussion of the day is driven by Rightwing crazies.
A poll comes out that seems favorable to Obama and the next thing we know they are parsing the vocabulary.
I've seen this fake "sampling" meme posted in bold faced all caps on Politco's comment boards.
If this were McCain, there would be no discussion of polling terminology.
MSM = GOP lapdogs ?
March 27, 2008 1:49 PM | Reply | Permalink
The headline should read: "Not over-represented". They certainly did over-sample. As Josh points out, there is a difference.
March 27, 2008 1:51 PM | Reply | Permalink
The headline changed while I was posting. Bravo!
March 27, 2008 1:52 PM | Reply | Permalink
This confusion is a result of MSNBC's Chuck Todd's limited explanation in his First Read article.
I saw Peter Hart correct the record on MSNBC and he also pointedly made a reference to the mistaken interpretation of the sampling of African Americans on their "Morning Joe" show.
March 27, 2008 1:53 PM | Reply | Permalink
Hey Clinton surrogates (TPM), why don't you write an article about Obama now having a statically significant lead in the Gallup poll?
You wrote plenty articles about Clinton's surges in the polls during the Wright controversy.
March 27, 2008 1:58 PM | Reply | Permalink
Speaking of Gallup's tracking poll, Obama has gone from trailing Clinton by 7 points to leading her by 4 in just eight days. I note, however, that Rasmussen has been getting almost opposite results. Is one of these polls more trustworthy than the other - or does this just prove that all these polls are so much noise?
March 27, 2008 2:19 PM | Reply | Permalink
Perhaps it's my partisan blinders, but Rasmussen's finding that Clinton and Obama both trail McCain 51-41 seems a bit extreme. McCain has been polling better recently, but does not lead by more than a few points generally, more in line with Gallup. Not sure if that says anything about the Obama-Clinton numbers though since obviously they use different samples than the GE matchups. Obama lost a lot of ground in both polls after the Wright stuff came out, so either Rasmussen hasn't picked up his recovery yet, or Gallup has picked up a false recovery. Hopefully the former...
March 27, 2008 2:49 PM | Reply | Permalink
With all this gnashing of teeth over polls, one would think the general election is tomorrow.
Relax, people. November is still a long time away. Plenty of time for Obama to re-unite the party against McCain.
March 27, 2008 3:25 PM | Reply | Permalink
The latest Pew poll is out: Obama 49%, Clinton 39%. The result is based on data gathered after Obama's speech on race. Result is virtually unchanged from late February poll.
Anyone have the skinny on Pew's reliability?
March 27, 2008 4:09 PM | Reply | Permalink
I stand corrected; they didn't weight, they simply pulled another sample of blacks entirely.
Either way, it's perfectly normal legitimate pollster practice; which is bound to confuse the hell out of people and let charlatan poll inspectors (and remember, you can't spell charlatan poll inspector without C L I N T O N) whine.
Another reason why America is getting dumber as a result of this protracted nomination battle.
March 27, 2008 4:17 PM | Reply | Permalink
Good question...
CNN, "Cafterty File"March 27, 2008 4:34 PM | Reply | Permalink