“Doug Wilder [an African-American and former Virginia Governor] says Patrick shouldn’t put too much stock in the findings [that he has a wide lead – Ed.]. In his Virginia campaign, surveys of voters as they left polling places showed him ahead by a comfortable 10 percentage points.
“When the votes were counted, he won by just half a point.
“Wilder, now the mayor of Richmond, related that cautionary tale to Patrick strategists at a political fundraiser in Washington last month. He says there’s no empirical data to prove that white voters today are any less likely to lie to pollsters about their willingness to vote for black candidates.”
The article concludes, There remains a non-trivial faction of white voters who will not vote for a candidate simply because (the candidate is) black, says Vincent Hutchings, a University of Michigan political scientist who is co-authoring a book called Wedge Politics. We are kidding ourselves if we argue these people have disappeared from the landscape.
I don’t agree with the analysis about race eventually making the lead evaporate because people were lying to pollsters. The reason being is that race has thankfully played a very insignificant role in this race and the polls last time underestimated the lead Deval Patrick had, even though one would expect the same type of biases would appear in the primary polls as in the general election.
Wilder says:
<
p>
<
p>
Fine, but is there any empirical data to prove that white voters are willing to lie to pollsters about their vote only to do otherwise in the voting booth?
<
p>
I’ve heard this canard a million times, in past campaigns and today (Ford’s campaign, and now Patrick’s). I’d love to see some evidence for it other than mere speculation.
The article cites Wilder’s race for governor, where he led by 10 points in the latest polls but won by half a point. I’ve also seen references to Harvey Gantt’s races against Jesse Helms that showed similar discrepancies. You should be able to find the specifics on the web somewhere. So there is at least circumstantial evidence — that it happens in the south.
<
p>
I suspect it will happen far less in Massachusetts. I expect Patrick will outperform the polls because of the Patrick-Murray GOTV organization. But I would not be surprised to see Harold Ford lose by a lot more than the polls say. It will be interesting to compare the Tennessee results with the other close races. Then there are a couple of not-quite southern states with black Republicans, Maryland and Ohio. Those results might show something too.
<
p> – Dan
Is there an Obama data point? How did he do relative to the polls?
He was running against anothwer black candidate, however, so it may not be as relevant data point. The telling ones, insofar as they exist, would be black candidate vs. white candidate.
<
p>
I’m not saying that this phenomenon cannot or does not exist, but I’m not convinced that this is anything other than another piece of “conventional wisdom” that is always bandied about but doesn’t have empirical backing.
It always gets on my nerves when folks claim there’s no evidence.
<
p>
Misleading pollsters, especially around hot button topics like race is fairly common. There’s a pretty good analysis (empirical evidence) of Wilder’s campaign that was put out in 1992.
<
p>
Academic surveys (as opposed to professional polls) tend to work pretty hard to take this kind of misreporting into account. If I was part of Patrick’s team I wouldn’t be able to sleep with less than a reported 10 point lead going into Tuesday.
<
p>
In any event, for the truly geeky, one academic study concluded that “It seems clear to many respondents in our racially sensitized electorate that the “socially desirable” response when asked about a vote for a white and a black candidate facing each other is to indicate support for the black.” Leading to inaccurate results. (Traugott and Price, Public Opinion Quaterly, Summer 1992)
1) It should be noted that this phenomenon did not seem to occur in the Democratic primary. In fact, the opposite happened, and Patrick outperformed the most recent poll at the time (if I recall correctly). Now, that doesn’t mean it won’t happen in the general, but it’s worth pointing out.
<
p>
2) A number of polls done in this race (all SUSA and Rassmussen polls) were done by recorded voice, and not using live pollsters. One could imagine that there is less social pressure to support the black candidate when the voice on the other end of the phone is, in fact, a machine. If that’s the case, then we would see Patrick perform worse in the SUSA and Rassmussen polls than the UNH and Suffolk polls. This has not been the case.
<
p>
That said, the Patrick/Murray team would be foolish to rely on these polls. There’s still a lot of work to be done to make them come true.
in 1989. Times have changed and Massachusetts is no Virginia.
and MA are often considered racial hotspots by northeastern standards. Black athletes still complain about the atmosphere here. This should be a good barometer of progress.