Bradley Effect Myth Persists

[SPC]

My dad and I have an ongoing gentleman's bet over who will win the election. Whenever I point to polling showing that Obama is currently favored in most of the swing states, he replies that Obama's position is overstated due to the Dinkins effect. I would assume that this belief is prevalent among conservative optimists. I don't understand how people continue to believe that the Bradley effect is going to seal Obama's fate when it was nonexistent in 2008?

(Disclosure: In 2008, I did think the Bradley effect would make the election much closer than it actually was. However, seeing my hypothesis fail caused me to discard this belief.)

[後援会]

The original Bradley effect was not even real either. It was a result of Bradley having bad pollsters. Demun(however you spell his name)'s pollsters were spot-on regarding his margin of victory.

[Talleyrand]

The original Bradley effect was not even real either. It was a result of Bradley having bad pollsters. Demun(however you spell his name)'s pollsters were spot-on regarding his margin of victory.

George Deukmejian. He is the son of Armenian immigrants.

[Landslide Lyndon]

Just wait till JJ discovers this thread.

[Grad Students are the Worst]

I agree -- I don't get this line of thinking.

It's really hard for me to believe that the social desirability effect could possibly be higher this year in 2008, when there wasn't any indication people lied to pollsters to avoid being perceived as racist and then voted Republican.  Obama now is an incumbent with a record.  Surely it's a lot easier now than in 2008 to oppose Obama for a reason perceived as race-neutral.

[○∙◄☻¥tπ[╪AV┼cVê└]

Voter intensity can affect things. Remember how Kerry was creaming Bush in the early exit polls?

[Free Palestine]

My father claimed Obama would loose due to the Bradley effect, in 2008.

[Free Palestine]

Also, part of the problem with the Bradley effect theory is that Bradley lost to an Armenian.  And the jury's still pretty much out as to whether or not Armenians are "white."

[Talleyrand]

Also, part of the problem with the Bradley effect theory is that Bradley lost to an Armenian.  And the jury's still pretty much out as to whether or not Armenians are "white."

I think the only way that results could be skewed because of his ancestry is through his name, and I don't think that made a difference.

[TJ in Wisco]

The reason people think this is that they heard it suggested on TV once or twice and thought it sounded smart. So they latch on to it as political truth without ever having looked at any data themselves. It's one of these "Word is .... " sort of deals.

[Smid]

I don't think it's going to happen this year, but it may have been possible last election with likely voter models, because there were so many enthusiastic people who voted for the first time in that election. I think one of the questions that screens likely voters is "did you vote last election?" - so perhaps that should have resulted in Obama receiving a swag of votes that may not have been picked up in likely voter polling. Then again, perhaps I'm wrong. I'm just putting it out there as a suggestion.

[後援会]

The Bradley effect did not occur in 2008, will not occur in 2012, and has never occured, even in 1982.

http://www.washingtonpost.com/wp-dyn/content/article/2008/10/30/AR2008103002396.html

http://articles.latimes.com/2008/nov/04/opinion/ed-bradley4

[So the Heroes Fall]

The Bradley Effect indeed is a myth that never occurred. In 1982 Bradley did lead by double digits...a few weeks before the election. The polls quickly closed in the last few days, and he was up only one point on the last poll. Also Bradley actually won amongst same day voters, he was done in by absentees, and considering how difficult polling absentees is in 1982 it's possible the polls weren't wrong at all. It's the political science version of psuedoscience.

[J. J.]

I think it did occur with both Wilder and Bradley, but perhaps more clearly with Wilder. 

There was some evidence that it did occur in some of the gubernatorial elections in the 2000's, but never as pronounced as it was with Wilder or Bradley.  It was dwindling over time

In some states in 2008, IA, for example, Obama did overpoll, i.e. run well below his polling numbers; absolutely nobody expected it.  Where I was looking for it, PA, it didn't happen.

Something that did seem to happen in 2008, that hasn't been reported too much, is that Obama underpolled in states with high Hispanic populations, that were of Mexican ancestry. 

[Grad Students are the Worst]

the bradley effect:

1. now responsible for the statistical margin of error and/or arbitrarily state-specific

2. happens backwards for mexicans

[SPC]

My father claimed Obama would loose due to the Bradley effect, in 2008.

So did mine.

[J. J.]

the bradley effect:

1. now responsible for the statistical margin of error and/or arbitrarily state-specific

2. happens backwards for mexicans

Iowa was outside of the MOE.  I think that was the most dramatic.  We also had some large states where there was not reliable polling, CA, NY, and TX.  We'll see if he runs ahead of the polls in those states this time.

Something happened in NM, NV, and even AZ.  Obama ran ahead of polling.  I wouldn't call it a Bradley Effect.  People lying to pollsters is a possible reason. 

Like I said, another factor was that it was shrinking over time.

[CultureKing]

I think the Bradley effect may be driven by awareness of candidates.

For example: In Washington this year we had a supreme court race with very low visibility. Of the two candidates one was MUCH more qualified, and was expected to win by a solid amount. However results showed that he managed to only win counties in which he was either a known quantity or those which sent out voters pamphlets (the secretary of state this year didn't send them out due to budget cuts). In other words when voters have nothing to base their vote on except for the how someone's name looks on paper then the Bradley effect comes into play.

The race by the way was Gonzalez vs Danielson:
http://vote.wa.gov/results/current/Supreme-Court-Justice-Position-8_ByCounty.html

[SPC]

Something happened in NM, NV, and even AZ.  Obama ran ahead of polling.  I wouldn't call it a Bradley Effect.  People lying to pollsters is a possible reason. 

Or perhaps pollsters have more difficulty interviewing ESL Hispanics in those states, which would be largely Democratic?

[Grad Students are the Worst]

the bradley effect:

1. now responsible for the statistical margin of error and/or arbitrarily state-specific

2. happens backwards for mexicans

Iowa was outside of the MOE. 

It's almost like that happens 1-in-20 times or something, or there can be other explanations besides the Bradley Effect for people who aren't trying to force data to conform to their hypotheses.

I'm also glad you wouldn't call an overperformance in Hispanic states the Bradley Effect, because that would be completely and ridiculously incoherent.

[Californian Tony]

I think it did occur with both Wilder and Bradley, but perhaps more clearly with Wilder. 

There was some evidence that it did occur in some of the gubernatorial elections in the 2000's, but never as pronounced as it was with Wilder or Bradley.  It was dwindling over time

In some states in 2008, IA, for example, Obama did overpoll, i.e. run well below his polling numbers; absolutely nobody expected it.  Where I was looking for it, PA, it didn't happen.

Something that did seem to happen in 2008, that hasn't been reported too much, is that Obama underpolled in states with high Hispanic populations, that were of Mexican ancestry. 

So, in other words, there's no Bradley effect.

[Tender Branson]

In some states though there was a Bradley effect in 2008.

Pre-election polls in AL & MS for example showed Obama with 20% of Whites, while in the end he got only 10% in the exit polls.

Of course the results remained the same as in the pre-election polls, because more Blacks compensated the loss of Whites.

[Smid]

Of course the results remained the same as in the pre-election polls, because more Blacks compensated the loss of Whites.

This was the point I was trying to make. That there may have been unexpected voters compensating for other "lost" voters - hiding the presence of a possible Bradley effect.

[So the Heroes Fall]

the bradley effect:

1. now responsible for the statistical margin of error and/or arbitrarily state-specific

2. happens backwards for mexicans

Iowa was outside of the MOE. 

It's almost like that happens 1-in-20 times or something, or there can be other explanations besides the Bradley Effect for people who aren't trying to force data to conform to their hypotheses.

I'm also glad you wouldn't call an overperformance in Hispanic states the Bradley Effect, because that would be completely and ridiculously incoherent.

The most inane thing about this whole theory is that it basically relies on the Bradley Effect being totally random. As in for whatever reason a bunch of voters in Iowa en masse decide to lie to pollsters to not appear racist (racial issues always being rather thought out in Iowa after all), but this doesn't happen in Pennsylvania or Ohio or Indiana. Though of course it supposedly did happen more mildly in Pennsylvania and Ohio just two years earlier with the black Republican candidates but then entirely faded (well not really because it's obvious which way undecideds were breaking in 2006 and even then the cited races weren't really outside the margin of error, but that's beside the point.) So in other words the Bradley Effect just happens, but we can't gauge why or where it happens or consider any other options as to why polls might be wrong (including margin of error.) Does that make sense?

And yeah the Hispanic thing is basically nonsense and probably just a cheap attempt to look "balanced" and not just throwing out the argument as pure GOP hackery.

Something happened in NM, NV, and even AZ.  Obama ran ahead of polling.  I wouldn't call it a Bradley Effect.  People lying to pollsters is a possible reason. 

Or perhaps pollsters have more difficulty interviewing ESL Hispanics in those states, which would be largely Democratic?

Actually nothing all that odd happened. J. J. is as usual cherry-picking results and doing it badly. In New Mexico (which is always notoriously difficult to poll), the polls had an average of him with 55%, and he got a little under 57%. In Arizona Obama got a little under 45%, and the last polling average had him at 46% (and of course why undecideds might break for McCain in Arizona is a little obvious). Nevada's the only one where the polls were off by a significant margin, where Obama averaged 50% and he got over 55%. However the last poll average also showed 6% undecided, and with the swing Nevada took and the economic conditions of the state it's not too hard to simply seeing the undecideds breaking heavily, and with the dynamics of the state using a turnout model based off 2004 in 2008 would be somewhat inaccurate. I'm oversimplifying quite a bit obviously, but these are far more logical explanations than "The Bradley Effect causes black candidates to underpoll in states with lots of Hispanics."

Also look at Texas where Obama polled at 41% with 5% undecided and got 43.6% (aka the undecideds broke almost 50/50) and California where he polled at 59% and got a little under 61% (with 3% undecided.) So this "inverse Bradley Effect" thing with Hispanics basically requires the same total random occurrence to be believed in.

[Sibboleth]

This is a long-running thing on this forum. For those that aren't aware of the main issue at hand, J. J. is not psychologically capable of admitting that he's wrong.