When is the death penalty okay?
How much does advertising matter in presidential elections?
Bartenders are Democrats, beer wholesalers are Republicans
The ambiguity of racial categories
No, public opinion is not driven by ‘unreasoning bias and emotion’
Political science: Who is it for?
Modern campaigning has big effects on voter turnout
Vis a vis the piece about public opinion: from my own limited reading and understanding, I would have to agree with the underlying messaging conveyed in the journalism. From ‘What’s the Matter with Kansas’ to ‘Nudge,’ the message seems to be that people are limbic and that only some sort of Vulcan training can cure them of their biases. Things begin to bifurcate there. Some people are willing to trust the training, usually embodied in liberal arts training, ‘critical thinking,’ ‘values clarification’ and the like, with the appropriate test a set of shibboleths about racism and so on. Other people arguing that even the Vulcan training of liberal arts education doesn’t work very well, that it is overridden by bias and needs to go further down its path; or that it reinforces a different set of biases, and is dominated by one political tribe. Very few people seem willing to defend the sensibility of man, a few more defend commonsense/intuition/EQ in women, and so on.
A few people, like J. Haidt (righteous mind, which gets mentioned in the comments under the original article) seem to defend the sensibility of mankind within frames of reference and operating from some very basic assumptions. Unfortunately, the frames are radically different, and there appear to be two mutually incompatible sets of very _very_ basic assumptions. He makes the extremely interesting observation that these assumptions are so fundamental that to be of a different assumption appears as nothing less than intention wickedness from the other; and that
the frames are so pervasive that one does not easily step outside of them regardless of personal advantages or disadvantages with regard to the alternatives.
If the message of individual irrationality and mob madness isn’t the message being conveyed by the research, then the academic communication enterprise is seriously broken.
Re. the death penalty article: Basically, no matter what the penalty, death or not, there will be the occasional innocent wrongly sentenced. Correct?
So the strongest arguments for or against the death penalty would have to be on other grounds. i.e. Even in a world where we could never err about who is guilty, the death penalty might still be abhorrent, but not on grounds of the false positives rate.
I think you’re missing the small point that the death penalty is more difficult to correct than, say, life in prison.
I think he also missed the 47% error rate which is being described as ‘occasional’.
Try correcting ’20 years in prison’ after the fact.
I believe massive wads of cash are standard, no? It’s not really adequate but it beats ‘ah crap, we already killed ’em, sorry about that’ by a wide margin.
Give even more wads of cash to the heirs?
Yes, there will always be type I and type II error.
The first question is whether there should ever be any use of the death penalty. If there should be,
then not using it in those cases is an error. This is what balances against the mistaken
executions.
An alternative perspective would be that anything is justice as long as it is done ‘fairly.’
Thus the opportunity for ‘unfairness’ is an argument against potential mistakes in an action,
particularly a high impact action.
This is open to obvious reduction to absurdity. Not anything fair is just. Fairness is one small
component of justice, and particularly dear to 2 year old children rather than reflective adults.
A return to the root concept of justice would be nice. Against alternative perspectives, there is
the sense that by taking an innocent life, one takes one’s own as well. Whether one is then forced to
live in servitude as compensation, or effectively has one’s blood ‘on your own hands’ in a manner that
cannot be commuted; but must be mechanistically made a concrete reality, is an additional question.
I taught a course on decision theory to honors freshmen/sophomores a number of times in the past 15 years. By the time we get to questions of this sort, the students have had a good deal of experience in assigning loss functions. Invariably they assign loss functions that make the death penalty impossible, unless the evidence of guilt is p(guilt)=1. Any value of that probability <1 has produced a decision tree that favors life without parole over the death penalty.
The basic reason for this is that they rank the loss of (acquit an actually guilty person) to be less than the loss of (convict an actually innocent person and sentence to life), which in turn has the greatest loss of the three (convict an actually innocent person and sentence to death).
Generally they have assigned zero losses to correct decisions. I believe that if they didn't do that, some decisions might be different for some values of p.
In case that second paragraph wasn’t clear, they assign the greatest loss to (convict an actually innocent person and sentence to death).
I’m interested in particularly serial murder in this kind of analysis. Suppose you have a person suspected of being a frequent serial murderer, like maybe once a year for the last 5 or 10 years. Now you’re trying to come up with a loss function, the loss associated with “acquit when guilty” should include the deaths of the future victims, whereas the loss associated with “convict an innocent person” also includes the deaths of future victims (since the serial killer is still out there) but if you include any kind of deterrent at all, even a delay caused by the real killer realizing that if he continues suspicion will be taken off the wrongly convicted… then there’s an imbalance which could potentially favor the death penalty, especially when there’s sufficient lengthy appeals etc such that the deterrent effect might take effect for a while, while there is opportunity to correct the mis-conviction.
In other words, I wonder how rational our current system is, with its “Death penalty highly delayed” structure?
There’s also the cost of concentrating really bad people together in prison, and how mis-conviction might cause death of innocent inmates at the hands of other inmates. When you actually go full decision theory here it can get complicated.
The question of serial murder did not come up in any of the classes I taught. The classes did usually consider the loss involved in releasing an actually guilty person who may go on to murder again, but this did not, at least in any of the classes I taught, outweigh the greater loss, in the students’ evaluation, of executing someone who was actually innocent…remember, the students were determining the loss functions by class consensus, I was just writing things down on the whiteboard.
You raise an interesting question. Still, though, life w/o parole (barring escapes like the one several weeks ago in upstate New York) should be a pretty effective preventive against that.
Don’t know what to say about your last point. Yes, full decision theory can be very complicated, more complicated than these students could probably handle!
Nonetheless, I like these examples (criminal cases tried by a jury) since it represents a kind of decision having nothing to do with money that some of the students are likely to actually have to make at some point in their lives, serving on juries being a responsibility of every citizen. It might not be a murder case, but they all share features of these cases…evaluating probabilities, determining loss functions that take into account what happens if the wrong decision is taken (the action does not agree with the actual state of nature), etc. This course was fun to teach. The students really got into it, they learned some things that probably would be useful in their lives, and the professor really enjoyed the experience of teaching such smart students.
Modern campaigning has big effects on voter turnout
I don’t see how (and the link to the paper is broken in the WaPo column so I can’t confirm this) Enos and Fowler can separate out the increased general interest/advertising that exists in battleground states versus ground campaigns/voter outreach/etc. To separate these out, one should look at states that wer battleground versus non-battleground between two succeeding presidential elections and look at individual voters and try to match them. North Carolina 2004/2008 (even though Edwards was on the ticket in 2004, I don’t believe it was contested) or Indiana 2008/2012 come to mind (whether you can get individual voting records for those states is another matter, but since NC is a Voting Rights Act state (before it got gutted by SCOTUS)), it should have individual voting records.
It’s funny that political scientists are getting around to this. I ran an analysis in the 80’s on voting registration drives for a state party and came to the conclusion that the cost was about $90/vote (not that different than the $87 a vote, though these aren’t adjusted for inflation). I was told to deep-six this as no one would fund such paltry returns and the party still wanted to do registration drives (of course, the Obama campaign had a billion, money any state party can only dream about. At $87/vote that gives 11 million or so votes for a billion (I realize most of the campaign money did not go to GOTV, but I’m making a point).
Hillary should have at least a billion (and the Republicans/Koch will have more). GOTV should help the Democrats more as their voters are, by all the usual political science criteria, more likely to be “marginal” (mobile, non-property owners, lower income, minorities, single parents, etc).
Is it “in,” “on,” or “at” a blog? I always say at. I’m reminded of the story that early users of the WELL policed flaming because they perceived a listserv to be located “in” their own homes, not off in cyberspace.
From sister blog on “When is the death penalty okay?”
Formatting went awry. Only first half of the blockquote is the quote. Starting with “Usually…” are my own musings. Sorry for the inconvenience.
The posts in this series always make me wonder: does your sister read your blog?
Thanks, Erin. I needed a good laugh.
Erin:
Susan doesn’t read the blog every day but sometimes I point out entries to her. Also I think some of her students and postdocs read the blog, so when her work is featured, I think she hears about it.
Andrew is right. And thanks for the plug! :)
There’s additionally the expense of concentrating truly terrible individuals together in jail, and how mis-conviction may bring about death of honest prisoners on account of different detainees. When you really go full choice hypothesis here it can get confused.