On the exit polls

You can find an interesting discussion on Kos about Richard Morin's column on those damned bloggers who discussed the exit polls on election day.

First, let us note the pattern: Bloggers (the newspapers tell us) did a fine thing when they gave Dan Rather a poke in the eye (even though the only qualified expert to comment on those controversial National Guard documents has disproven the claims of the rightist bloggers). Switch the issue to disputes over this very disputable election, and suddenly we hear that bloggers are either foolish or fiendish.

The "liberal" press is feeding us a simple formula: Right blogs good, left blogs bad.

Morin's piece is (for the most part) the usual crap. Forget the exit polls that called it for Kerry: Exit polls are inaccurate. Ask him why the inaccuracies don't swing both ways, and he'll repeat: Exit polls are inaccurate.

Yes, you say, but shouldn't Republicans register as many false positives as Bush?

Morin fidgets. Umm...did I mention that exit polls are inaccurate?

Yes, but even if they were (and usually they aren't), shouldn't the exits have been inaccurate in (say) Illinois as well as in Florida or Ohio? Why are exit polls accurate in Europe but not here? Isn't it odd that only the Republican vote is undervalued in election after election?

And Morin has no choice but to repeat: Exit polls are inaccurate.

And how do we know they are inaccurate? Because they don't agree with the final tallies.

In other countries, the presumption works in the other direction. Observers presume vote tampering has occurred when the official tallies seriously conflict with the exit polls. After all -- aside from the exits, what other verification do we have?

Nevertheless, exit polls do "damage" according to Morin, because they leak out. For some reason, he does not feel compelled to offer any evidence that this country houses even one (1) Republican voter who stayed home because he or she saw a leaked exit polls on Kos or elsewhere.

"Damage" my ass!

Morin gives us only one new piece of information:

On election night in 1988, we relied on the ABC News exit poll to characterize how demographic subgroups and political constituencies had voted. One problem: The exit poll found the race to be a dead heat, even though Democrat Michael Dukakis lost the popular vote by seven percentage points to Dubya's father.

So...presuming this information to be accurate (a presumption I make only provisionally), the pattern of an undervalued Republican vote goes back to 1988. Does this factoid really help Morin's argument? In my view, the situation becomes even more worrisome.

In election after election, over the course of 16 years, only the Republican vote has been undervalued -- which is to say, the Democrat earned fewer votes than the early exits predicted. (The earlier exits provide more independent data, since later exits are polluted by incoming "official" numbers.)

This pattern cannot occur by accident.

So why do the exit polls consistently favor Dems? The Morins of this world offer no answer beyond "shit happens."

At least one diarist on Kos attempts to offer a more substantive answer. The diarist makes reference to Mark Blumenthal:

My point is that there are two competing theories for the discrepancy: The first is that the exit polls were slightly biased to Kerry due to a consistent pattern of what methodologists call "differential non-response" that has been evident in exit polls to a lesser degree for a dozen years (Republicans were more likely to refuse to fill out the exit poll than Democrats). The second theory is that systematic and consistent vote fraud occurred in almost every state and using every type of voting equipment. The first hypothesis seems plausible to me; the second wildly improbable.

Reverse the terms "plausible" and "wildly improbable," and you get closer to the truth.

"Differential non-response" is simply an impressively polysyllabic term for what I have called the "chatty Democrat" theory. The idea that conservatives are taciturn by nature is, of course, pure horseshit. The modern Republican party is a cult, and most cultists are verbal bullies. Liberals, by contast, have been cowed; they speak up only in non-mixed company.

Once we toss the "chatty Dem" theory onto the trash heap where it belongs, what do we have left?

Vote fraud.

Is the notion "wildly improbable"?

How many pundits in 1972 would have used those very words to describe the proposition that Nixon's men would break into Democratic headquarters? How many pundits would still use such a term to describe the (pretty much proven) thesis that Reagan's team cut a deal with Iran in 1980? How many pundits in 2003 would have described as "wildly improbable" the thesis that Bush lied about weapons of mass destruction in order to grab another nation's oil reserves?

Blumenthal places a lot of stock in Warren Mitofsky's interview in Mayflower Hill. (Mitovsky is one of the heads of the polling consortium.) This precis (by a Kos reader) sums up the situation well:

Mitofsky in Mayflower says he has already checked to see whether there's any divergence between precincts with electronic voting and others, and he didn't find any. Case closed, said Blumenthal.

But wait a minute. Which precincts did Mitofsky check? Many of us are not concerned that all precincts have voting irregularities. Only a select few. Why is Blumenthal so quick then to close the case? Why is he so ready to call us delusional?

I appreciate his web site as an aggregator of info, but i haven't appreciated his analysis at all.

With regards to Mitofsky, the only thing the revered exit pollster offers to rebut questions of possible fraud is speculation. Mitofsky refuses to consider possible machine irregularities.

Once again, the debunker's argument presumes the validity of the very data being questioned. Why should we consider the final tallies accurate? They're accurate because they're accurate. That's why.