The Hibbs model (which takes into account the strength of the economy, combined with the intertial factors of incumbancy and the effects of no longer having the President's coattails in the midterm elections) >predicted that Democrats would lose about 45 seats. It's currently looking like Republican gains are about 65 seats. Assuming the Hibbs model is valid, only about 2/3 of Republican gains were due to shark attacks, and Democrats lost an additional 20 seats due to factors not taken into account by the Hibbs model.

Actually, thinking about things a bit more, the Hibbs model predicts based on the combined effects of shark attacks and regression to the mean from previous elections. In 2006, Republicans lost a ton of seats to the Democrats due to shark attacks (predicted on the nose by the Hibbs model). 2008 isn't modelled since it's a presidential election year and Hibbs is specific to midterms, but it's pretty safe to conclude that shark attacks played a major role in Democratic gains then as well. Now it's 2010, and the sharks are still attacking, dammit, so the voters have turned on the Democrats for failing to get rid of the sharks they way the voters expected them to do.

My point in my previous comment still stands, though. Hibbs takes all that into account, and still predicts significantly (eyeballing the chart, 20 seats seems well above the standard error) fewer than the actual Democratic losses in the House.

Thank you for teaching me about the Hibbs Model. That's interesting, and somewhat worrying. It makes actually going through the voting ritual seem kind of like a waste of resources.

I only just learned about it myself on Monday. Even with the Hibbs model, the voting ritual still accomplishes a few things:

1. Making broad exceptions. Occassionally, there's something big voters care about that's not taken into account by the model (such as household gremlin infestations or plutonian frosting), but it gets taken into account by voters.

2. Making individual exceptions. If an individual congressman gets caught with bribe money in his freezer or underage congressional pages in his bed, voters have a role in dealing with that.

3. Even if the general election is pointless, the primary is not. Hibbs predicts how Republicans and Democrats will be elected, but it says nothing about which Republicans and which Democrats will be elected.

4. If we allocated seats based on the Hibbs Model rather than holding elections, there would be an enourmous incentive to game the measurement of personal income growth, to improve the number of seats held without benefiting the actual health of the economy. For instance, one could raise taxes in Q2 then declare a tax holiday in Q3 in order to increase income growth between Q2 and Q3. Or one could just instruct the Bureau of Labor Statistics to fudge the measurments as much as they can get away with.

Everything affects everything else. It's all interdependent. There is no such thing as "a cause". There are always pretty much infinite factors involved!

Of course, it's useful to look for some of the more direct causes, especially if they are things that are optional and under our control, so that we might be able to sway chaos in a particular direction a bit in our favor in the future. But there's always that damned butterfly flapping it's wings in Brazil, or China, or whatever...

There's no single cause, but there are things that explain certain precise amounts of variance.

Not really. At best we can get a general idea. Anyone telling you that they have precise measurements of reality is either trying to sell you something, or doesn't understand physics very well. :-)

Can't you use http://en.wikipedia.org/wiki/Analysis_of_variance to piece apart the causes of variation in a variable?

You can try! Just don't take any results you get as at all accurate. They are more like suggestions pointing to landmarks that might be interesting to look at on your sight seeing tour of life. But the tiny, tiny details of your trip, including toothpaste flavors, and the woman who smiles at you at the coffee shop, are just as important to the overall experience.

This seems like a pretty significant fallacy of grey to me. Causality is sometimes hard to pin down, therefore let's abandon all hope of talking sensibly about causality?

No. I'm saying that causality is, like everything in the universe, always more complex than it looks like on the level you're looking at it so it is indeed impossible to pin down any sort of "cause" definitively, and that some people, when they are looking at some of the statistically probable causes, get too attached to thinking that they've got some "fact", when they've really just got a potentially more likely possibility.

But...

...since possibilities are how we humans function in life, obviously we're going to keep paying attention to them! Just be aware that those who build their entire belief systems and plans for the future on some statistically demonstrated "fact", without at least considering what they'd do if something less or even un- predictable happens, tend to come crashing down hard, eventually. :-)

If the President isn’t responsible for the economy, he can neither be acquitted of causing the current recession, nor be made look pretty good in slowly but surely bringing the country out of it.

He can be acquitted in the same way someone who isn't responsible for a murder is acquitted of a murder, and he can be made to look pretty good, just not proven to actually be pretty good.

I will face my politics, I will permit it to pass over me and through me, etc..

when the politics are gone, only I shall remain. (?)

I recommend "Twelve Days of Terror: A Definitive Investigation of the 1916 New Jersey Shark Attacks" by Richard G. Fernicola if anyone wants a good read about the Jersey Shore shark attacks.

It seems that someone forget to tell the American public that correlation does not imply causation.

Most people are terrible at grasping that concept.

I even see it all the time in Magic The Gathering (the card game). It is often important to realize that just because you lost doesn't mean you played wrong and even more importantly just because you won doesn't mean you played right.

But the American public isn't interested in the causes of their economic misfortune.

The economy hasn't gotten much better in 2 years under the Democrats, ergo they cannot fix the economy. They don't understand that the economy didn't go bad in just a year or two and can't be fixed in just a year or two.

(2007 saw a big fall - on Bush II's watch. And as I recall it was starting to teeter by around 1999 and was generally downward after 9/11 and has never been back to the level the 90s).

For one of the first times in my life, I find myself sympathizing with the man. His electoral loss wasn't proof that Americans hated him

That whole Iran-Contra scandal didn't help his first term popularity either.

Wasn't there a mini-crash in '87? How did he get off the hook for that one?

The people who won Monday's election are thinking

The people won MONDAY's election are imagining the whole thing. ;)

That whole Iran-Contra scandal didn't help his first term popularity either.

Contra came up near the end of Reagan's second term.

Wasn't there a mini-crash in '87? How did he get off the hook for that one?

The crash was actually a pretty big one, the largest one-day US stock market decline in percentage terms since 1929. The reason it had no real impact on Reagan's political fortunes (or Bush the Elder's on Reagan's behalf, since Reagan had already faced his last election) was that it wasn't correlated with any broader economic trends. Rather, it was an artifact of herding behavior and poorly-designed automated stock trading programs. Over the course of the entire year, the stock market went up, as did GDP and personal income.

Rational Ignorance being what it is, restrospective voting isn't that bad a heuristic. It takes a lot of effort to become informed in even a general sense about every candidate on the ballot, much less make a detailed study of each candidate's biography, character, and voting regord. Since the marginal effect of an individual vote is miniscule, people generally only remotely approach the specific knowledge required for a properly informed vote if they follow politics as a hobby. But just about everyone has a pretty good sense of their own personal situation and how it's changed over the past few years, and while the relationship between that and a politician's performance in office is very noisy, on average you're more likely to vote correctly.

Compare alternative heuristics that voters could be using instead:
"He has such an honest face"
"I hear the other guy is a secret Muslim"
"This guy seems like he'd be more fun to have a beer with"
"I'm sick of seeing her commercials"

Even semi-informed votes are problematic, since there are several biases (anti-market bias, anti-foreign bias, pessimistic bias, and make-work bias) that distinctly incline the voting population towards candidates who will enact (or at least promise to enact) bad policies.

I agree with you that they're not *literally* blaming the government for the sharks. But when you say it's not irrational - well, of course it's irrational! If let's say you usually agree with the Democrats, but you vote Republican because the Democrats are in power during a shark attack, you've just frustrated your own interests for no benefit.

I would say the best description of what's going on is that it's an unconscious algorithm (vote for or against the government depending on whether or not you're better off than you were before aka retrospective voting) carried to a ridiculous extreme. This doesn't require the voter to consciously represent the statement "the government causes shark attacks", but that is effectively what their actions would imply.

Most studies of of retrospective voting behavior that I'm familiar with conclude that retrospective voting is used mainly be people who don't have a strong general preference between the two parties.

The red Congress would not be able to hurt much yet. All they can really do is tie Obama's hands for two years. Alright, that would give him time to concentrate on foreign policy and on PR, the latter hopefully to prove Reds crooks and hypocrites by using the "They would not let me, freaking party of NO" viz. their "See, he did not accomplish much" line. Then he will win the 2012 elections, and have Blue Congress and Senate and four years of Fear None at his disposal.

never mind. That was estel