Are the polls underestimating Obama's support among African Americans?

[NMMatt]

An interesting analysis:

TNR Blog Entry

I don't necessarily agree with their speculation as to why the numbers are low. I'm curious, though, if this really is the reason why Obama tends to out perform the polls, especially in the states cited.

[TexasObserver]

1. Yes, they're consistently underestimating both turnout and his edge.

When they report something like 64% for Obama and 15% for Clinton, that ought to be a clue.

[Tactical Progressive]

2. The media has been saying 90/10

Are you saying that the media is underplaying Obama's advantage with African Americans?

[NMMatt]

3. No that the polls aren't picking it up enough and that may be why he tends to outperform the polls

[Tactical Progressive]

4. I think I see what you're saying

But that would mean that African Americans aren't admitting that they are supporting Obama.
Why would they do that?

[NMMatt]

5. No idea, but the evidence does show him outperforming the polls -nt

[bunkerbuster1]

6. "Admitting" might be the wrong word.

Democrats in general have, to use Obama's word, a great deal of "fondness" for the Clintons and of course there's still a great deal of support for them among African-Americans.

The idea of a "reverse Bradley effect" might pertain to what happens when black folks actually get to the polls and are looking at two candidates, ideologically similar, and faced with the prospect of voting against their own ethnic group. It's reasonable to imagine that in this unique instance some folks will change their mind on election day vs. what they'd told a pollster a day or three before.

But it'll be interesting to see if exit polling bears this out.

[NMMatt]

7. Actually I think that is what he was saying...

I.e. that exit polling showed 90% support in places like MD, GA and MO while the pre-election polls showed significantly lower margins. If it happens again, WI, OH and TX would represent more examples.

[bunkerbuster1]

10. I don't disagree with the numbers, just the terminology.

I think we might see something of a surge (oh, please strike me dead for using that term in this context!) in Obama's numbers in the primaries ahead, particularly OH and TX.

[johnnydrama]

8. when you poll

When you poll, you try to figure out who's going to vote.

So if you think in a state, 70% of the voters will be white, 20% black, and 10% latino, then you want to poll in those ranges.

But if it turns out that 60% of the voters are white, 35% are black, and 5% are latino, then your poll is going to be off if 90% of the African American's are voting for one person.

So if they aren't polling the proper amount of African Americans, you are going to come in at a low number of votes.

I think the same thing happens with young people. They are hard to poll. They could live in colleges, or only use a cell phone with no main number. So you don't get the proper amount of them in your polls.

[MadBadger]

11. There may be a lot more African Americans voting now because of Obama

They wouldnt usually be considered likely voters.

[TalkAgain]

9. They are

I just read one of the polls from yesterday, I think OH or TX...They said he is getting 73% of blacks, but I know he has won blacks by 90%-10% marging just last week in some places. Anything under 80% is understated.