Chris Hanretty

Magic incantations are good

Chris — Sun, 07 Mar 2010 12:47:07 +0000

Some people say magic incantations are bad.

But then, you find a code snippet — sorry, incantation — on Hadley Wickham’s web site, and find out how to plot that interaction effect graph you need.

Isn’t it pretty? The bare bones:
effectdf <- function(...) { suppressWarnings(as.data.frame(effect(...))) }
my.xlevels=list(x=seq(0,1,0.05), rol=c(TRUE,FALSE) both2 <- effectdf("x*z", lm.mod,xlevels=my.xlevels)
fplot <- ggplot(mapping = aes(y = fit, ymin = lower, ymax = upper) fplot %+% both2 + aes(x=x,group=z,color=z,fill=z) + geom_smooth(stat="identity")

If you want substance, the paper in which the graph features will be uploaded in a week's time, ready for the UK PSA in Edinburgh.

Review of Simona Piattoni’s Il Clientelismo

Chris — Thu, 04 Feb 2010 14:17:27 +0000

Clientelism, like corruption, is a juicy topic; it is made so by the seemingly-inexorable conflict between protestations of public virtue on the one hand and particularistic exchanges on the other. The topic becomes even more interesting when one tries, as Simona Piattoni does in her book Il clientelismo: L’Italia in prospettiva comparata, to rehabilitate clientelism, and rid the concept of its moral overtones by engaging in rigorous comparative and empirical analysis.

Read the rest here. It came out a bit harsher than I meant.

Berlusconi 2020!

Chris — Mon, 01 Feb 2010 13:26:41 +0000

This interview with Carlo Taormina is fantastic. Key points (all according to Taormina):

The “processo breve” is a bargaining tool. Offering to remove it from the agenda can be seen as a concession; a concession can win quick approval of a text of “legitimate impediment”.
The bill on “legitimate impediment” (repeated postponement of trial hearings due to demands of government office) is unconstitutional, and Berlusconi’s circle knows this
Legitimate impediment might be declared unconstitutional, but by the time that happens, a new constitutional amendment (lodo Alfano-bis) will guarantee prime ministerial immunity
Once this is all passed, Berlusconi will call early elections in 2011, and then wait for the presidential elections of 2013. Assuming he gets that, he’ll be with us until 2020.

Nice to see someone from the centre-right putting these things down in black and white. [h/t Mattia Guidi]

I fought the law, and the law won (bleg)

Chris — Wed, 27 Jan 2010 12:48:40 +0000

Two more things:

Why do legal journals insist on this bizarre referencing style of Article Author (Year), “Title”, Vol(No) Journal ppp ? And more to the point, does anyone have a bibtex (.bst) or biblatex (.bbx) file that will replicate it, with proper block punctuation?
Wasn’t the Supreme Court’s decision on the Terrorism Order interesting? Partial dissent by Lord Brown, could have hoped for more there. (I like dissent).

Two things I don’t get

Chris — Mon, 18 Jan 2010 19:13:32 +0000

Why are Italian politicians bending over backwards to say nice things about Craxi when he was, after all, a fugitive from justice?
If your one-dimensional model works fine, but your two-dimensional item response model doesn’t converge after a long long run even with informative priors on items over both dimensions, can you do a Gertrude Stein and conclude that there’s no there there?

Slides for this morning’s thesis defence

Chris — Fri, 18 Dec 2009 08:13:54 +0000

Parochial interest only, but here they are

Why Fini can make himself unpopular

Chris — Wed, 02 Dec 2009 15:29:44 +0000

Gianfranco Fini has been caught saying things about Berlusconi that Berlusconi doesn’t like. (Let’s leave aside for the moment the question of whether Fini’s comments were correct and/or inflammatory).

Berlusconi, according to the usual unnamed sources, wants Fini to be punished, even to the point of requiring him to step down as President of the Chamber of Deputies.

I think this is unlikely. As far as I can see, no President of the Chamber of Deputies has ever resigned, unless to take up another, better post. Three Presidents of the Senate have resigned (Merzagora, Paratore and De Nicola), but Paratore and De Nicola both resigned in protest; only Merzagora’s resignation is surrounded by question-marks (I’m following Wikipedia here).

Nor is there any provision in the Chamber’s rules of procedure to vote no-confidence in the President of the Chamber.

Neither side wins from a protracted fight. I’m betting this blows over.

Leave the country

Chris — Mon, 30 Nov 2009 11:24:45 +0000

Depressing editorial from Pier-Luigi Celli. Celli, a most capable man, is the director of Luiss, the closest thing (with the exception of Bocconi) that Italy has to an ENA; and yet I can’t imagine any Frenchman writing this kind of editorial.

Ideal points for the 16th Camera dei Deputati

Chris — Fri, 27 Nov 2009 00:03:06 +0000

I’ve been scraping voting records for the 16th parliament (lower chamber only). That’s part of my continued (quixotic) interest in trying to get credible ideal point estimates out of situations (western European parliaments) where the assumptions of most estimation techniques (i.e., legislators vote sincerely) are violated.

There’s an article coming out by Luigi Curini and Francesco Zucchini (Italy:Government alternation and legislative agenda setting. In: Bjorn-Erik Rasch and George Tsebelis (eds.) The Role Of Governments in Legislative Agenda Setting) which assumes that ideal points can be estimated by only using final votes on bills.

Now, that works for previous parliaments — and I though I’d test that on the current parliament. The current parliament is interesting largely because of one opposition party — Italia dei Valori — which is centrist in its economic policy, and on many other issues, but is vociferously opposed to Berlusconi. So, it tends to oppose whatever it sees as compromise. As you can see from the graph, that pulls them to roughly the same position as the Partito Democratico.

Ideal points from the 16th Camera

What you can’t see from the graph is that the median Italia dei Valori legislator is slightly to the left of the Partito Democratico. That’s not the result you would expect if legislative rollcalls were determined by left-right positions instead of government-opposition. So, IMHO, we can’t get credible ideal points by restricting the analysis to final votes only.

Karmic Koala

Chris — Tue, 10 Nov 2009 13:32:14 +0000

I’m not too happy about my upgrade to Karmic Koala. That’s due both to my own stupidity (not backing up as fully as I should have done), and to a number of minor niggles.

For example: whilst wireless worked perfectly, my ethernet didn’t autoconnect. I fixed it by following the steps at Craig Mayhew’s blog, but that shouldn’t be necessary: this was a fresh install after all.

Also, there’s a hardware beep when I shut down. No kidding.

On the plus side, everything is very smooth, perhaps also due to the fact that I’m now running 64 bit. Just don’t ask about Flash and/or Java just yet.

Now I just need to wait until TeXLive 2009 hits the mirrors and I’ll be ready to start work again.