http://www.lowellsun.com/todaysheadlines/ci_10803787
I was relieved this morning to see that 59% of Bay Staters are opposed to the ballot measure to repeal the state income tax. Good stuff. And with only 26% actually in favor, it looks like this thing has a snowball's chance in hell of passing.
Polls are often wrong, but those numbers are pretty statistically significant.
Much talk has been made this year of whether the polls for the Presidential race may or may not be off due to the so-called "Wilder Effect" or "Bradley Effect" (for Douglas Wilder of VA and for Tom Bradley of CA, respectively). The idea behind it is that there's some hidden pocket of racism in people that leads them to tell a pollster they will (or have, in the case of exit polls) vote for a black candidate even when the reality is otherwise.
Well, I don't live under a rock, so I know there are some Americans out there who won't vote for Sen. Obama solely because of the color of his skin. I'm also aware enough of my surroundings to recognize there are also many Americans of all colors and backgrounds for whose votes Sen. Obama's identity will be a positive factor. Most won't state that outright -- they'll find other ways to justify their votes, but I really don't believe they'll feel the need to somehow distort that when they talk to a pollster.
But as long as we're going to talk about the "Wilder Effect" I think it's only fair we talk about the Dewey Effect (remember the famous 'Dewey Defeats Truman' headline?), the Gore Effect, and the Kerry Effect.
In fact, why stop there?
What about the Dukakis Effect, the Hillary Clinton Effect (in many of this year's primaries), or even the Howard Dean Effect (Iowa, 2004)?
The point is, polls are flawed.
They contradict each other, they're based on small samples, they can change greatly based on slight wording tweaks, and they don't always capture the portion of the population that actually votes.
I have no formal statistics background, but at least I'll start by admitting that I know what I don't know. On the whole polling issue, I think that's more than most of the cable news media is willing to admit.
But somewhere in America on the morning of November 5, some Sociology graduate student will get his or her hands on some poll that somehow did not accurately reflect the way some slice of America voted on the 4th. And that will somehow be fodder for some Ph.D. thesis about the Underlying Biases of our Epistemiological Dystopian Freudian Self-Conceived Images of the Idealized Forms of Reflected Idolatry.
My tax dollars will help fund it.
But because I make considerably less than $250k (got my W-2 if you ever want to see), President Obama and his Party in Congress will take less of it, and that will be the bright side.
I believe that people are dishonest with pollsters, but
ReplyDeleteThe Bradley Effect is pretty much bullshit. People forget that Bradley actually WON the number of votes cast on Election Day -- what did him in was losing in the absentee ballots. Of course, absentee voters are (usually) never polled, so The Bradley Effect says more about methodology than race.
Matt,
ReplyDeleteGood additional points about Bradley. And L. Douglas Wilder won in Virginia, period.
I agree that people lie to pollsters on hypotheticals like, "Would it bother you if your daughter brought home a person of such-and-such ethnicity?" but it seems strange to me that on an up-or-down type of poll like "Who will you vote for?" people would be afraid to admit they were not voting for a candidate of color.
Either way, 24/7 news outlets will need stories to drum up, so they'll find them where they can.
-gp