• My book: Bayesian Analysis for the Social Sciences
  • Wiley; Amazon; errata 10/4/11
• 112th U.S. Senate ideal point estimates: pdf, csv 9/26/11
  • scatterplot against Obama vote share pdf
  • roll call object: rda, documentation for roll call objects
• 112th U.S. House ideal point estimates: pdf, csv 9/24/11
  • scatterplot vs Obama vote share pdf svg
  • roll call object: rda, documentation for roll call objects
• Next Australian Federal Election, betting market summary, 9pm Oct 3 (Sydney):

  Agency	ALP	Coalition	ALP Win prob	Last 7 days
  Centrebet	4.00	1.23	0.235
  sportsbet	3.60	1.28	0.262
  sportingbet	3.50	1.28	0.268

• time series graphs: entire month week day
• Australian 2010 Election Analysis

This might be an interesting seminar. I wonder if Brad knows about the long history of “poor-person’s” (approximately-asymptotically valid) Bayesian inference in political science via things like Clarify?

Tuesday, October 4, 4:15pm: Statistics Seminar, Sequoia Hall Room 200
Brad Efron, Stanford University Statistics
“Bayesian inference and the parametric bootstrap”

Abstract:
The parametric bootstrap can often be used to compute posterior distributions obtained from complicated Bayesian models. Besides its computational advantages, the bootstrap offers insight on the relationship between Bayesian and frequentist methods. I will discuss some examples relating to exponential families, generalized linear models, and high-dimensional inference.

Cargolux was due to take delivery of a freighter edition of the new 747-800 on Monday. There was much fanfare scheduled up in Everett to celebrate the 1st delivery of the 748s. But today Cargolux has announced it is refusing to take delivery. No details as to why: “unresolved issues” are what the parties are saying in press releases. But are the issues performance-related, or something else?

Horrible news and images from Reno.

And a photographer dropped a Canon lens from a sightseeing flight of some kind over Petaluma, with the lens punching a hole in the roof of a house. Terminal velocity for a Canon lens?

Special elections are, well, special. Sometimes they produce results like we saw last night, where the rest of the Congress stays put, but one seat switches from being represented by a D to an R (or vice-versa).

NY-9 is usually not so interesting. It is an odd shaped district, nestled amid majority-minority districts and is reliably Democratic, although not as much as its maj-min neighbors. Obama beat McCain there 55-45 in 08 on turnout of about 120K votes; Weiner won that election 93-7 against a Conservative candidate. In the 2010 “wave” election, the Republican (and now Representative-Elect) Turner managed to get 39% of the vote; turnout was 110K in that election.

A lot gets read into special elections, but turnout was only 63K yesterday (a figure that will tick up a little as the vote count finishes), still pretty impressive for a special election in September in an off-off-year, but way down from what we can expect in an even-numbered year. Turner has 34K votes thus far, for a 53-47 margin. He got 10K more votes and lost by 10-11 points in 2010. Put differently, while Weiner got 67K votes in 2010, the Dem contesting this special election got less than half that (30K thus far). Turnout — always an important part of the story — played a critical role here, and so any big picture read of NY9 needs to be against the backdrop of what looks like woeful Dem turnout.

My point here is to show something that *is* consequential, and that is the kind of representation that NY-9 will now be getting in the Congress. See the accompanying graph, which plots estimated “ideal points” (a summary, liberal-conservative, measure of each representative’s voting history) against Obama’s 2008 2-party vote in the district.

Weiner chalked up a roll call history that was a little left of what you’d expect for a Democrat in a district that went 55-45 for Obama in 2008. NY-9 now gets a Republican member (Turner), who can be expected to vote a little to the left of the median Republican House member, but about 1.6 units from where Weiner is (was) on the ideal point scale.

Now, the ideal point scale is normalized to have a standard deviation of 1.0, so we’re talking about an abrupt and large change in the legislative representation of NY-9.

Good luck with that, citizens of NY-9, and the same to Representative-Elect Turner.

Of course, it will be interesting to see how Turner actually votes, who the Dems recruit to run against him in 2012 (when Dem turnout will presumably be back up), and whether the district reverts “to type”.

And of course, some post-Census redistricting is in the works too, which will scramble the egg a little further.

In a 10 min span I saw a 747-800 and a 787 fly low in-bound to Boeing Field. 787 in ANA livery. Pure luck, happened to look up while out and about for lunch; grabbed my iPad to verify what I was seeing (PlanePlotter, Flightaware).

An update of my paper with Lynn Vavreck (UCLA) is available over on the research part of my site; it is currently the 1st listed paper there.

Abstract: Would the outcome of the 2008 U.S. presidential election have been different if Barack Obama had not been the Democratic nominee? In this paper we analyze 33 head-to-head match-ups, some real and some hypothetical, to understand more precisely how candidate traits and election-level characteristics affect election outcomes. We find that “old-fashioned” racial stereotyping is uniquely important in decisions about Obama in 2008, relative to its role in past elections or in 2008 choices substituting Clinton or Edwards for Obama. Similarly, “new” or symbolic racism is an exceptionally important predictor of vote choice when Obama in the choice set, and it gains in importance over the year leading up to the election. The Democrats would have won the 2008 election regardless of who they nominated, but the average Democratic party nominee from the last 16 years, and either of Edwards or Clinton, would have done better against McCain than Obama, although in Clinton’s case, not by much.

And teaser (?) thumbnail graph:

United Continental announced plans to upgrade the interiors of some of their aircraft:

United Continental Holdings (NYSE: UAL) today announced it will invest more than a half-billion dollars in its onboard product as it takes another important step in becoming the world’s leading airline. These product improvements include:

• Adding flat-bed seating on 62 additional long-haul aircraft, bringing total aircraft with flat-bed seats to 185, more than any other U.S. carrier
• Adding Economy Plus seating and Channel 9 air traffic control audio to more than 300 aircraft
• Nearly doubling the overhead storage space on more than 150 aircraft
• Installing advanced broadband Wi-Fi on more than 200 aircraft
• Introducing streaming wireless video onboard its 747-400 aircraft
• Completely retrofitting its p.s. fleet with flat-bed seats, Economy Plus, on-demand audio and video and Wi-Fi

These and other planned product changes come in addition to the 25 new aircraft – including the Boeing 787 Dreamliner – that the airline will introduce to its fleet next year. United and Continental have each made significant product and service improvements over the past several years and since the airlines closed their merger Oct. 1, 2010.

I have 1M+ miles on UAL and almost always fly them, but after traveling with SIA and ANZ (and even a IAH-SFO flight on a CO 738), I’m struck by how old and run-down the UAL interiors are. It isn’t just the absence of personalized, on-demand video in coach on a 744 (for instance), or no in-seat power in coach on a 777: lavatories, seats, trim, finishes, everything tends to be older and tired looking on UAL equipment, save for the improvements they’ve been making with the business class product.

2 years ago the WSJ reported the average age of UAL 747s as 13 years, 757s as 17 years (!) and the 777s as 10 years. It shows.
At some point, old planes are simply old planes — so it might make sense to spend a little on making the economy class experience a little less horrible, but thats about it.

I remember being on a brand-spanking new UAL 777 out of ORD many moons ago — the captain invited us up to the cockpit to check it out, on what was like revenue flight #3 for that aircraft.

Bring on the 787s please.

If Australians can shorten a locution they will. The “Ekka” is short for the “Exhibition” as in the Royal National Agricultural and Industrial Association’s Annual Show (or Exhibition). The closest thing for Americans might be a state fair.

I remember the 1974 Ekka, getting off the train back down in Sandgate with my grandmother to see the headlines that Nixon was resigning.

People’s Day is the Wednesday of “Show Week”, a public holiday in Brisbane. It was wicked crowded at the Ekka, but we took it on in good spirit: many photos here, including the Grand Parade.

This sad news struck a little close to home. Janet had worked with Paul Lockyer at A Current Affair in Sydney back in the 1990s, and had also worked with the camera operator John Bean in Canberra.

We’d really enjoyed the way Lockyer had scoped out the regional/environment beat with the ABC, showing city people (and the world) amazing stories and images from the bush. The stories and images of Queensland flood waters working their way south into the Murray-Darling and west to the Channel Country and down into Lake Eyre were wonderful (some samples here).

The news of the crash got us off to a sad start today here in Brisbane. Janet was saying she thinks that Locyker is the 2nd journalist from her A Current Affair days to have died in a chopper crash.

Screen shot from the ABC web site:

We’re having a wonderful time catching up with friends and family in South East Queensland. I’m pushing many photos etc on the gallery.

My sister lives just south of Toowoomba, with lots of farm land and bush nearby. One of my sister’s dogs got a little carried away yesterday, greeting us with a “catch”:

The bird lived…um, we think.

And today my son Tom thought a sheep would look better in his hat (as if there was any doubt to this sheep’s provenance):

From the SMH: It looks like over 2M Australian households opted to take the Census on-line…out of 9.8M households in the nation.

But over 670K on-line Census-takers “forgot to hit the send button”… (which strikes me as a design flaw, and something you’d never have arising as an issue in a conventional on-line opinion survey, with the client browser pushing responses to the server on a per screen basis).

So it would seem that the Australian Bureau of Statistics has accepted the proposition that one should let respondents complete in the [self-complete] mode of their choice (although they do have the force of law compelling responses of some kind).

We did paper forms (and one per person) here in our Brisbane rental apartment. Its a long survey compared to the US Census, but goes pretty quickly for the kids (“if under 15 years of age, skip to Q54…”).