Mark Kleiman notes that the Arrestee Drug Abuse Monitoring (ADAM) program has been killed. This was a useful dataset on patterns of drug-use amongst criminals. In his post, Mark quotes John Coleman, a former bigwig at the DEA, who says

The importance of ADAM always has been its stark statistics showing the large percentage of criminals high on drugs and alcohol at the time of their crimes. ADAM surveyed arrested felons and then drug-tested them to confirm their statements about drug use. It was all voluntary but showed, nonetheless, extraordinary levels in some cases of drug use by criminals.

This confirms my non-expert belief that there’s a great deal of evidence telling us that a big chunk of violent crime happens when the perpetrators have been using alcohol or some other drug. People under the influence of drugs tend to have a diminished capacity for rational decision-making. This makes me skeptical about, e.g., fiendishly clever analyses of the rational deterrent effect of prison sentences on crime rates. It’s not that there’s anything wrong with the detail of such analyses per se, it’s that they throw away reliable knowledge before they begin. Ignoring information of the sort that ADAM provides may make an elegant theory of crime more tractable, but it makes a true theory of crime less likely.