It is a collection of striking questions with clever use of data to answer them. The most infamous piece of analysis is that the surge in abortions as a result of Roe v Wade is a major factor in the dramatic drop in the US crime rate 20-odd years later – the argument being that those children most likely to become criminals were those most likely to be aborted by abortion becoming cheaply available to low-income women (pp137ff). When I originally read of the analysis, I was highly sceptical, but an impressive case is put, not least because factors are adduced to explain both the dramatic rise in the US crime rate in the 1960s and additional factors for the drop in the 1990s: more police do deter crime and incarceration does reduce crime. (I like the quote [p.123] from a political scientist that:
apparently, it takes a Ph.D. in criminology to doubt that keeping dangerous criminals incarcerated cuts crime).Capital punishment had almost no effect (pp123-124), because the US simply does not fry/hang/needle enough folk to have a significant effect (even on the optimistic statistical estimate that each execution stops about 7 murders).
Needless to say, this is an analysis with something to offend all sides of politics. (A very critical take on the abortion-cut-crime argument is here. James Q. Wilson declares the case unproven here.)
Another striking result is that election results cannot be explained by how much candidates spend. Analysis of almost a 1000 Congressional races where the same candidates were against each other in successive elections found that:
a winning candidate can cut his spending in half and lose only 1 percent of the vote. Meanwhile, a losing candidate who doubles his spending can expect to shift the vote in his favour by only that same 1 percent (pp11-12).
Then there is the data on how it is not what parents do with their children, it is what they are, which has most effect on their children's student test scores (pp166ff). Keeping the family intact (or not) has no measurable effect, having lots of books around does, reading to them doesn't, nor does the amount of television they watch. High parental socio-economic status is a positive, moving to a better neighbourhood has no measurable effects, a child's mother being 30 or over when she had her first child is a positive, mother staying at home between birth and kindergarten has no measurable effect, child having low birth weight is a negative, speaking English at home is a positive, going to museums regularly has no measurable effect, being regularly spanked has no measurable effect, parents being involved in the PTA is a positive, being adopted has a negative effect compared to what their parent's characteristics would otherwise suggest (this seems to be genetic – high-status parents adopt children low-status mothers give up: on the other hand, adopted children do noticeably better in college-attendance, career and income and avoiding teen marraige than similar children not adopted [p.176]).
Having a backyard pool is about a 100 times more dangerous for one's young children then having a gun in the house (pp149-50). And lots about cheating (and how most people don't).
My favourite section was on the analysis of a crack gang (Black Gangster Disciple Nation, pp94ff). The gang was essentially a franchise arrangement with steeply increasing incomes (an Indian-born researcher, Sudhir Venkatesh spent years studying the gang and managed to get hold of the books put together by the local leader, J.T., a college graduate with a business major). That is, the gang worked basically like McDonalds with the addition of illegality and violence. The "soldiers" lived a home with Mom on very low incomes, but had a chance (if they survived and progressed upwards, about 1-in-4 were killed) of getting very high incomes. When a gang war was on, the soldiers were paid more; such wars being generally started by the soldiers, since violence was their route to status (it disrupted business as far as JT was concerned).
Full of fascinating stuff – and highly readable too!
No comments:
Post a Comment