>and analyzed quasi-experiments. We really don't know whether they
>controlled for the important confounds here until we read the
>research report.
True. However, we need to realize that the expressed criticism of the "three - strike" laws is based on misguided assumption. They seem to assume than any harsh punishment will drive criminals toward more serious crimes - which is a hogwash model of criminal behavior based on the rat-choice approach. Hence, the higher the "cost" we inflict upon them, the more benefits they want to obtain from their criminal activity. If this were true, we should abandon punishing them altogether to avoid pissing them off and committing more serious offences.
The problem with the three strike laws is that they limit judicial discretion in meting out justice and force judges to give long sentences to petty criminals. This may have a positive effect on the seriousness of committed crimes, but only after the petty criminals are released from prison (on the assumption that they become hardened and learn new skills from other inmates). This effect, however, cannot be demonstrated with the data used in the study - you would need a panel design for that (e.g tracking individual criminals before and after their three strikes sentence) and thus a much loner time lag.
wojtek