Pretty much the most interesting blog on the Internet.— Prof. Steven Landsburg

Once you get past the title, and the subtitle, and the equations, and the foreign quotes, and the computer code, and the various hapax legomena, a solid 50% English content!—The Proprietor

Monday, September 27, 2010

Lawyers Live Forever

With the gloom and doom besetting the legal profession, one bright spot is perhaps worth noting. According to a '99 paper,The Impact of Specific Occupation on Mortality in the U.S> National Longitudinal Mortality Study, quoted by Robin Hanson, being a lawyer seems to be just about the healthiest profession you could have.

If I interpret the graph in Hanson's post correctly, being a lawyer is associated with about a 30% decrease in death rate, compared with the population average. This appears to be the case with or without adjusting for age, gender, race, income, and education. Other healthy jobs include MathSciLib (math/science librarians?), RelgSocSv (religious social service providers?), OffManNEC (?), MedEqMchn (medical equipment mechanics?), and farmers, but—perhaps surprisingly—not physicians.

Equally surprising are some of the high death risk jobs. Why do office workers have a 120% (unadjusted)/60% (adjusted) increase in death risk? Why do food service workers suffer a 160%(unadjusted)/55% (adjusted) increase? While fire fighters and police only have their risk increased by 40%(unadjusted)/10% (adjusted)?

Friday, September 24, 2010

Dahlia Lithwick Can't Do Math

The demonstrated intellectual limitations of Slate's legal correspondent Dahlia Lithwick are a subject of frequent amusement in the blogosphere these days, so—before anybody else jumps in—let me offer an example from her most recent column "Lady Killer." In it she laments the cultural and institutionalized sexism which so often condemns innocent women to death for offenses as trivial as murdering one or more family members:

While women are sentenced to death far less frequently than men, often the offenses for which they are sentenced are also rooted in antiquated gender stereotypes. When women are sentenced to die, say experts, it tends to be for the most sexist reasons. Often, their crimes involve the murder of a spouse or a child, which comes with the assumption that they are bad mothers or unnatural wives. Experts say that men on death row, by comparison, have more often than not killed a stranger and done so in the commission of another crime. Why the double standard? Maybe it's because, from the days of the Puritans, Americans have thrilled to stories of fiendish and beastly women who killed their loved ones. The culture expected white women to be "kindly, passive, virtuous caretakers," writes Phyllis Goldfarb, a professor of law at George Washington University. When they committed murder instead, she writes, "execution seemed utterly appropriate."

Spot the basic statistics error? The high-profile "legal correspondent" didn't!

To obtain probabilistic evidence of gender bias in death sentences for murders of family members, look at the fraction of those charged with (or convicted of) family murders who receive the death sentences. If courts and juries really have sexist biases, then the fraction of women convicted of family murders that receive a death sentence should be higher than the fraction of men similarly convicted that receive a death sentence.

But that is not the measure that Lithwick looks at. She looks at the fraction of men on death row for family murders compared to the fraction of women on death row for family murders. The women's fraction being higher she cries "SEXISM!" But that is no evidence at all. In fact it is entirely expected if men—as is universally conceded—commit more stranger (or more generally, outside-the-family) murders.

To see that, consider a simple stylized example. 100 men commit family murders. Every single one of them receives a death sentence. 100 women commit family murders. Half of them receive a death sentence. In addition, 900 men receive death sentences for other murders. Hence, only 10% of men on death row are there for family murders, while 100% of women on death row are there for family murders. Do we conclude that the "system" judges women particularly harshly for family murder? Not at all. To the contrary, in this example, women are much more likely to receive leniency for family murder than men are.

Bonus Lithwick Math blunder:

Lithwick originally wrote:

Hard to imagine even the staunchest feminist insisting that if women commit 10 percent of the murders, they should die 10 percent of the time for it.

I can take no credit for spotting this one. It was apparently reported by a sentient reader or editor after publication of the article and corrected to:

Hard to imagine even the staunchest feminist insisting that if women commit 10 percent of the murders, they should compose 10 percent of those executed for it.

Postscript: Ms. Lithwick reportedly received a law degree from Stanford University in 1996. Wasn't a reasonable LSAT score required to attend a highly selective law school back then or did it not have logical and quantitative sections? If so, it is difficult to conceive how she was admitted.