>consider those who might commit crimes, but don't. Current studies support
>the idea that most people obey the law because they think it's right rather
>than because of fear of punishment.
I think this implies a rather narrow sense of deterrence, the rat-choice quid pro quod variety so to speak. In a broader sense, deterrence means fear of loss of status and being alienated from one's community by violating the norms of that community which accordind to soem crimonologists is the main factor that keeps people from being delinquent (cf. works of Travis Hirschi). The obverse to it is criminal communities (e.g. gangs) where being delinquent and violent is the norm of respectablity - thus not commiting a delinquent or a violent act is a violation of the nomr that can result in a loss of status (cf. Jack Katz, _Seductions of crime : moral and sensual attractions in doing evil_).
The bottom line is that deterrence has a much broader sense than the dominant paradigm based on individulaism and the rat-choice ideology suggests. It works mainly as informal sanctions and the loss of one's standing in his/her community. No wonder that mainstream pundits and social engineers (including liberal ones, who are devoted to individualism) prefer the narrow (rat-choice) meaning of deterrence.
Another point, the most effective deterrence from crime and delinquencly is strong embeddedness in stable communities - but that goes against the interest of land speculators, profiteers and social engineers. Hence we have the 'war on drugs,' the police state, and the prison-industrial complex - all measure based on the rat-choice model of criminal behavior.
wojtek