Wednesday August 29, 2012
I created the following graph to look at how current estimates of Obama 2-party vote share in each state compares with how Obama actually did in 2008. The state-by-state 2012 estimates come from the work I’m doing for HuffPost’s Pollster.
The usual way one might make this graph is to simply do a scatterplot. But I also like this form of presentation.
I’ve employed a non-linear scaling of the vote shares (or 2012 voting intentions), devoting more of the graph’s area to interesting states close to the 50-50 point (the purple bar marks 50-50).
Right now, North Carolina and Indiana are the losses for Obama. Iowa, Colorado, Florida, Wisconsin are very close to the edge. Some more/better polls from these places would be terrific, by the way…
Monday August 27, 2012
From a Gerard Henderson column (“Left closes ranks to consign Kerr to wrong side of history”) in the SMH:
In his conversations with me, Kerr made it very clear that he was desperate not to involve Buckingham Palace in an Australian political dispute. There is little doubt that Whitlam would have moved against Kerr if he had been consulted by Kerr along the lines Mason suggests. In the climate of 1975, this would have caused a constitutional crisis and possible political disorder.
Uh-huh. Unlike the constitutional crisis we actually did have and the “political disorder” of turfing out a popularly-elected government…?
Which is of course, wholly consistent with how the Right understands 1975. There was no “constitutional crisis”, because “the system worked”…with the GG exercising the Crown’s reserve powers… There was no “political disorder” because Fraser immediately advised a dissolution…that Kerr granted. Whereas Whitlam sacking Kerr? Why that would have been a terrible “constitutional crisis”.
Sunday August 26, 2012
…and some money comes back to Labor. The 3 bookies I follow all have Labor’s odds improving in the last couple of hours:
Yet more evidence of markets following polls…
Friday August 24, 2012
From one of the R lists I follow:
Today (2012-08-23) on CRAN :
“Currently, the CRAN package repository features 4001 available packages.”
These packages are maintained by approximately 2350 different folks.
2011-05-12: 3,000 packages 
2009-10-04: 2,000 packages 
2007-04-12: 1,000 packages 
2004-10-01: 500 packages 
2003-04-01: 250 packages 
 My private in-house data.
PS. This count includes only packages on CRAN. There are more
Thursday August 23, 2012
I’ve added an Electoral College counter to the code running the poll averaging I’m doing for HuffPost.
This is the usual, posterior/predictive guff: (1) we sample many times from the posterior density for Obama and Romney vote shares; (2) for a given sample, is Obama > Romney in state s?; if so, give the EVs for state s to Obama, otherwise to Romney; (3) sum over states; (4) repeat many times over the many samples from the posterior density. Make a pretty histogram when you’re done. Count the number of times you see Obama at 270 EVs or better. The proportion of times this happens is the posterior probability that Obama wins the Electoral College (at least given your model and its many assumptions, uncertainty over its parameters, etc etc). Note no special treatment of Maine, Nebraska.
Here’s what I get, or got, middle of the day, Thursday August 23; Obama narrowly ahead, with a .684 probability of carrying the Electoral College, but a better shot than that currently on offer at InTrade (.572). Note that the InTrade number is more properly a forecast than what comes out of my “pooling, adjusting, smoothing, x-state leveraging” of the polls (and relying on historical election results, etc); punters get paid on what happens in November, not on what poll respondents say now.
Tuesday August 21, 2012
A reminder that we’re not doing the usual public information meeting (yawn) at APSA this year. Rather, the idea is to do actually present results, findings, etc, since that what is what we’ve been using (as much as time and resources permit) to drive the instrument writing process etc (radical, I know):
You are cordially invited to attend an ANES related roundtable discussion
at the 2012 Annual Meeting of the American Political Science Association
(APSA). The session, titled â€œInstrumentation and Innovation in the ANES:
Results from the EGSS Studies,” will provide information on new
instrumentation that was tested as part of our Evaluation of Government
and Society Study (EGSS).
The session will take place on Friday, August 31, 2012, 10:15 AM-12:00 PM.
The current location for the roundtable is Marriott La Galerie 4, but if
you plan to attend, you should check the program when you arrive, in case
there is a location change.
We hope to see you there!
Thursday August 16, 2012
Perhaps with the Olympics over, perhaps with Labor’s capitulation on offshore processing, punter interest has drifted back to the domestic political betting markets in Australia.
There was a tiny improvement in Labor’s price at both Sportingbet and Centrebet this afternoon (Sydney time); see the spark lines and other graphs etc, above.
Still, it is a very sleepy market and Labor remains gone for all money (well, almost all money, presumably the betting houses moved the prices in response to some money coming on the long odds offered for Labor).