“In an article that has become a classic to sociologists of law, David Sudnow (1965) described how attorneys in a public defender's office acquire -and share with the prosecutors and judges with whom they negotiate pleas - a consensus about the risks in various types of sentence for what they call "normal crimes." Thus, as a normal crime in the court that he studied, "burglary” is seen as involving regular violators, no weapons, low-priced items, little property damage, lower- class establishments, largely Negro defendants, independent operators, and a non-professional orientation to the crime." "Child molesting is seen as typically entailing middle-aged strangers or lower-class middle-aged fathers (few women), no actual physical penetration or severe tissue damage, mild fondling, petting and stimulation, bad marriage circum-stances, multiple offenders with the same offense repeatedly committed, a child complainant, via the mother" (p. 260). The consensus among these lawyers included notions of the risks typically involved with each of the normal types of offense. Negotiation on every separate case usually begins with arguments as to whether it is a normal type of the particular offense or whether it involves a defendant who is a better or worse risk than most others charged with this crime. Usually, the guilt of the accused is assumed.
As Sudnow points out, these risk-classification norms vary between courts and are probably different for private rather than public criminal defense lawyers. Attorneys new to the process have to learn the local norms before they can efficiently resolve cases to the general satisfaction of their coworkers and with a minimum investment of time and effort per case. Comparative studies of the operations of criminal courts in different locations have found that a "courthouse subculture" develops in each, giving each court somewhat different norms as to the risks involved in alternative dispositions for particular types of cases.
Risk-assessment norms probably also vary from one collective workplace to another not only for courts but also for other types of criminal justice agencies. Thus the criteria for evaluating risk are likely to be somewhat diverse among different police stations where decisions to arrest and to book are made, penal institutions where prospects of escape or violence must be judged in allocating housing and work assignments to inmates, parole boards that have release discretion, and probation or parole supervision offices that must decide which releasees should get the most surveillance or other kinds of controls.
These normative variations reflect a basic principle of sociology and anthropology that I call the law of sociocultural relativity: social separation fosters cultural differentiation. Cultures are acquired by communication and tend to be continuously changing. When people in different workplaces do not interact with each other, their subcultures are likely to diverge.
The influence of local subcultural norms in risk assessment is objectionable from at least two standpoints. First, their variation conflicts with the ideal of maintaining uniform justice practices throughout a jurisdiction. Second, local norms may generate quite inaccurate judgments of risks. The inaccuracies of risk-assessment norms, however, probably reflect psychological principles that affect human thought in all types of settings.”
- Daniel Glaser, “Classification for Risk.” Crime and Justice, Vol. 9, special issue on Prediction and Classification: Criminal Justice Decision Making (1987), pp. 252-253.