jackman.stanford.edu/blog
bannerImage

Republicans looking for Obama coattails?

Thursday June 26, 2008

Filed under: politics — jackman @ 11:21 pm

A sign of how bad things are for Republicans this cycle comes from Gordon Smith’s re-election campaign. Smith is a Republican Senator from Oregon. But you’d never know it from a quick look at the front page of his web site:

Picture 2-7

Down below the luscious images of Oregon we get a hint:
Picture 1-20

The kicker is something that has been attracting considerable media attention over the last day or so: an ad that sees Smith reaching for Obama coattails, replete with a green (green!) logo stressing bipartisanship.

Picture 3-5

Oregon, of course, is the state where Obama drew 75,000 people to a pre-primary rally (although the crowd may or may not have been there for a free concert ahead of Obama’s appearence, by an indie band who may or may not be very popular, but whatever…).

Smith won 56-40 in 2002; Kerry beat Bush 51.3-47.2 in Oregon in 2004. which gives you some sense of the underlying partisanship of the state. 2002 of course was an “off-year”; 2008 isn’t, and it is shaping up as a tough slog for Smith, and quite a few other Republican incumbents this cycle.

Comments Off

passport renewal

Filed under: Australian Politics,general,politics — jackman @ 9:37 am

The Australian consulate in San Francisco just called to say that renewing your Australian passport means turning in your old one, boo hoo. So I scanned mine for posterity. The stamps etc will come in handy when I get to that part of the US citizenship application where you have to list all overseas travel in the last 5 years (“use extra pages if needed”).

And is there anything as sobering as looking at your old passport photo?

And Queen Elizabeth would like it if you let Australians cross your borders, a delightful anachronism:

Passport02

Comments (6)

Latinos and the poor, computer/Internet use

Filed under: politics,statistics — jackman @ 7:46 am

Driving to work this morning I heard that a Public Policy Institute of California study has found that just 58% of Latinos in California use computers, and only 50% of Latinos are Internet users. Apparently that 58% number is down from 64% in a previous study.

Another finding that caught my attention: among households with incomes under $40,000, half have home computers, but only 40% have home Internet access and 33% have broadband.

Food for thought for those of us using the Internet as a survey research platform. Far from fatal, but a major hassle.

Comments Off

the spatial voting model (EITM at St Louis)

Friday June 20, 2008

Filed under: computing,flight nerdery,politics,statistics — jackman @ 11:01 am

I taught at EITM@WUSTL this past week. I did 3 days on operationalizing the spatial voting model, a little theory and then showcasing R packages like my own pscl and wnominate. I also showed them the so-called two-cutpoint/party-presure model (a version of differential item functioning, in the roll call context), using JAGS/rjags.

Here are some photos from the flights and a trip to the baseball with Andrew Martin.

Teaching this stuff gives me a chance to see how the R packages either do or don’t do what you want them to do; I unearthed some bugs in pscl (and co-incidentally, Jeff Lewis found one this week as well). So pscl has been given a version bump as a result, and might get another one later in the summer.

Teaching is a learning experience for the instructor, too. So here are some lessons learned, or re-learned, or finally committed to writing:

  1. The vanilla spatial voting model is just that (vanilla): voting is conditionally independent across roll calls and legislators, given proposals, status quos and legislator ideal points, with “non-spatial” sources of utility considered zero-mean iid shocks. The start of the art is to write out richer models and see what statistical models they give rise to, and ideally, to estimate those models. This requires getting outside the “black box” of the canned routines (factor analysis, wnominate, ideal, etc). Its tough to get to the more advanced stuff in the allotted time: I suspect some of the students have never seen the spatial voting model before, let alone things like eigen-decompositions, SVD, factor analysis, IRT, MCMC, hierarchical Bayesian modeling, and so on. Even R is relatively new for some of them, to say nothing of BUGS/JAGS etc.
  2. Recent U.S. Congresses aren’t that much fun to analyze. The proper nouns are vivid (legislators who are also presidential candidates, such as Obama, Clinton, McCain, Edwards, Kerry, Biden, etc…). But relatively high levels of party line voting mean one dimension fits very very well and there isn’t a lot more to say (other than snooping for the few roll calls that aren’t fit well by the recovered dimension). The Supreme Court is fun.
  3. So I think more “exotic” examples might help: e.g, Voeten’s analyses of the General Assembly of the UN, Hix et al looking at the European parliament, decisions of the Federal Reserve, etc.
  4. On the other hand, the Congressional case has been the context in which we’ve seen the most elaboration on the standard model (e.g., the Snyder/Groseclose two-cutpoint model, the Clinton-Meirowitz agenda-evolution model). But then again, Quinn and Martin developed their dynamic ideal point model with Supreme Court data and Londregan ‘s proposer-power/valence model was developed with data from Chilean Senate committees and the Constitutional Convention (and see this follow up by two of my former students, Shawn Treier and Jeremy Pope). The point is that there is a lot more to do other than run recent Congresses through the canned routines, but not a lot of time for that in 3 days of EITM (approx 10 hours of class time).
Comments (2)

United news

Wednesday June 4, 2008

Filed under: flight nerdery — jackman @ 12:55 pm

Goodbye Ted (see here).

Further down in the article comes the news that

United, a unit of the UAL Corporation, said it would retire 100 aircraft, including all 94 of its Boeing 737 medium-range jets, assuming the airline can reach agreements with aircraft lease companies.

The number includes 30 whose retirement had already been announced. In addition, United said it was retiring six Boeing 747-400 series jets that are used on long flights, like those overseas and to Hawaii.

One wonders what the implications are for United’s LAX/SFO-SYD/MEL routes, currently exclusively serviced by 747-400s. Might we see 777 ERs on those routes? Or with 30 747-400s currently in the United fleet, maybe they’ll still have plenty of 747s available for Pacific long hauls.

United’s 747s are certainly showing some signs of age. I think I’ve seen ceiling panels fall down on passengers on takeoff on about 5 or 6 occasions on UAL 747s out of SFO or SYD (mid-plane, either just as we rotate, or shortly thereafter). And United’s 747s do not have in-seat entertainment systems in Economy class, which can be a real horror on a long haul.

Comments Off

superdelegates for Obama

Tuesday June 3, 2008

Filed under: politics,statistics — jackman @ 10:48 pm

The New York Times continues its cute series of election graphs. Here is the super-delegate story, rendered as an annotated time series.

Maxine Waters flipped from Clinton to Obama on Tuesday. Here she was spinning for Clinton at the LA Democratic Debate (Jan 31, 2008; my photos, click on the thumbnails).

Img 0685 Img 0687

And today’s news reminded me of this shot I got outside the LA Democratic Debate:
Img 0585-1

Comments Off

Running the Numbers

Filed under: statistics — jackman @ 9:19 pm

Cool renderings of large numbers by Chris Jordon, brought to my attention by Jean-Pierre Khoury.

Comments (1)

Powered by WordPress

Bad Behavior has blocked 4861 access attempts in the last 7 days.