Centerfield

That was funny, Tully! Although, I am a little bit embarrassed that I said some of those very same things on this blog! HA HA HA!!! Geez, is that what we sound like? No wonder Dems didn't win the election. We're nuts! :-)

Thanks for the article, Tully.

Part of me wants to call this article satire. The other part calls it a straw man.

If it's a straw man, why do I see the same thing over and over on the web? Another great piece is from Dean Esmay, Letter to John Perry Barlow From a Pot-Smoking Deadhead Bush Voter. Well worth reading.

As someone who works with counting statistics for a living, I just don't see any reason for all this “advice”.

In my line of work 10% is significant, 20% is definitive, 30% can't be questioned.

Bush won in 2004 by 3%
Bush won in 2000 by -0.5%
Clinton won in 1996 by 8.5 %

None of these results are statistically significant.

The 1/3 of the country that is independent gets to choose the president. Give us a decent candidate he will win by 30%. Give us a couple of sleaze balls like the last three times we get a tie.

As Dick Cheney's favorite web site (factcheck.org) showed a couple of liars ran so we split 50-50.

Bob,

Name the last Presidential candidate who won by 30%.

Good luck.

JonBuck--FWIW, satire relies on caricature.

I reject the premise of this article. It seems clear to me that the republicans won their races not on issues of policy, but rather due to a fundamentally better marketing campaign. Multiple polls have suggested that the public did not have a firm grasp of the policy issues at stake in this election - perhaps besides the gay marriage issue. From my perspective, the democrats were unable to clearly define themselves and their unique values or projects they supported. The bush team was clear that their number 1 goal was to keep the US safe from global terrorism. The reality of their proposals did not matter - what mattered was how clear and consistent the message was. The democrats and their "nuanced" positions on this issue (and others) made it unclear to many voters exactly what they wanted to do. The "swing" voters were particularly affected.

This brings up an excellent article about the moderation (nuance) issue and how it is getting completely ignored by so many people today.
Check it here.

Tully:
The point is that the choices suck and that the current 50-50 split is not a mandate.
But just for the record here are some who could actually say they have a mandate:
1984 REGAN 18%
1972 Nixon 23%
1964 JOHNSON 22%
1956 EISENHOWER 16%
1936 ROOSEVELT 24%

http://presidentelect.org/e2000.html

I reject the premise of this article

Jeffrey, it's satire. It has no premise.

The point is that the choices suck and that the current 50-50 split is not a mandate. But just for the record here are some who could actually say they have a mandate:

Yet not a single one of them was a "decent candidate" by your own criteria, Bob. Of course, by your criteria, we've apparently never had a decent candidate in the entire history of the nation (with the sole exception of George Washington, selected by acclamation). I'd be interested in hearing what your criteria are for "statistically significant." Without context, it's a meaningless term.

By definition, the winner has a mandate. Also by definition, it's exactly as broad as the Congress, the judiciary, and the public allow it to be. All else is back-biting and pre-emptive undermining.

Tully: I previously posted:
"In my line of work 10% is significant"

Statistically significant usually means we can distinguish a value from the normal variation a system shows around its mean value. This election does not qualify as statistically significant.

As for the 30% number, I'm not emotionally wed to it. I would settle for 20%.
After all, every candidate has his warts.

PS: You seem more upset that bush barely won, then the democrats do about barely loosing.

Oh Yea, I forgot to address "mandate".

My definition for mandate is that you won by such a significant margin (minimum 20%) that you could legitimately say the country wants to divert the ship of state to a bold new direction.

Nixon: Peace with honor
Roosevelt: New Deal

Bob, I think Karl Rove would tell you that in his business, 3% is significant.

Yet the assumption in your statement seems to be that electoral returns are a random variable of a quantifiable mechanistic system. How is this assumption justified? What other assumptions are you applying to elections? What confidence intervals are you applying? As I said, without context "statistically significant" is a meaningless term.

How do I seem upset? I said repeatedly for months that Bush was not the Second Coming, and a Kerry presidency would not be a disaster. Frankly, a landslide would have startled me, and in modern terms anything over 5% or 6% is a landslide. If anything, I'm simply annoyed at the misuse of statistical terminology in an inappropriate manner.

As for the 30% number, I'm not emotionally wed to it. I would settle for 20%.

You almost have to settle, as there are no 30%'ers.

After all, every candidate has his warts.

No argument there!

Jason: I'm not disputing who won. I'm disputing that this election signals some kind of massive shift in the political landscape.

Nor am I saying that the democrats should not change. That is their decision to make, not mine. Personally I find both sides kind of sleazy. A little more sincerity and a move to the center for the democrats would suit me just fine.

FWIW, Bob, I don't see any massive shift either. A gradual one, yes. And I'd love to see more sincerity and a move to the center.

He is the first President since Reagan to win with over 50% of the popular vote. He pulled down higher rates of voters in categories people didn't expect (18-20, moderates, hispanics) and the demographic trends are going with the Republicans, not against them.

All of the above means the victory IS statistically significant. Its not a New Deal type mandate, but the crises we are facing are not as serious as the Great Depression either. Regardless, mandate is a meaningless term.

"As I said, without context "statistically significant" is a meaningless term."

Tully: There are two scenarios I can think of to create context:

First, there are sources of error in the actual voting system as it attempts to ascertain the will of the people. Namely votes that were or were not counted due to weather, employment, car trouble, health issues, voting machines, corrupt officials, bad hair day............

What is the margin of error in the system? True we have a huge sample size but we also have a very complex counting system and environmental factors to contend with. I would have no heartburn assigning a +/-2 percent margin of to each of the candidate's numbers. If we held three separate elections that week do you really believe the each tally would still be exactly 50 to 48?

A second completely separate context argument could be made using presidential elections since the great depression (lets call it the modern age). I'm more than willing to accept the premise that without a really big decision to make, the electoral returns will trend to a 50-50 split that generates 10% noise.

Obviously all of this is just spit balling. If you have a link were someone has done a rigorous study of these subject I would be interested in reading it.