Underpunishing Offenders: Towards a Theory of Legal Tolerance

Article excerpt

This paper presents the results of two case-sentencing surveys and compares the demand for punishment (as expressed by the public) to its supply (as measured by the sentencing preferences of court actors). Findings show criminal court actors to hold significantly less punitive views than the public. This "underpunishing bias" cannot be explained away in terms of differential assessments of perceived crime seriousness, the socio-demographic background of respondents, or the presumed irrationality of public opinion. The crucial factors are to be found elsewhere: both sets of respondents differ significantly in the amount of responsibility they attribute to offenders; they also differ in how certain they are of achieving their normative goals. Relying on cognitive attribution research as well as on equity theory, the discrepancy in sentencing preferences between criminal court actors and public opinion is shown to be intrinsically linked to their cognitive status as actors and observers of just deserts.


There is some evidence that the demand for punishment expressed by society ("public opinion") is much higher than the actual supply of penalties provided by criminal courts. Is this "underpunishment bias" to be expected or not? The issue of legal "tolerance" has not been of great empirical concern for a number of reasons. First, the sociology of criminal courts has mostly been involved in uncovering latent and manifest discrimination (Black 1988; Hagan and Bumiller 1983). Discrimination or the overpunishment of certain classes of offenders on the basis of legally irrelevant or objectionable criteria may, however, coexist with an overall leniency bias. Second, public opinion is often seen as an irrational body and, hence, an unacceptable standard for assessing to what extent criminal courts over or underpunish offenders: "using public judgements to determine the penalty for burglary may be akin to the use of public opinion to determine the prime rate or the percentage of hydrocarbons that is to be tolerated in the atmosphere" (Durham 1985). The assumption that sentencing per se requires considerable expertise remains, however, controversial (Wilkins 1984). Even in domains where technical expertise is acknowledged (nuclear policy), dismissing "public opinion" has been shown to be problematical (Perrow 1984).

In the United States, public polls of the late 1970's and early 1980's indicate that more than 80% of the population "feel" that criminal courts are being overly "lenient" (Thomson and Ragona 1987; Flanagan, McGarrell, and Brown 1985; Stinchcombe 1980). Canadian or British polls obtain similar results (Hough and Moxon 1988; Doob and Roberts 1988). Such data sets, however, are inappropriate for assessing civil society's actual demand for punishment. First, the conventional poll questions ("would you say that courts hand out sentences that are too severe, just right, too lenient?") "tap into dimensions of generalized malaise" (Thomson and Ragona 1987) that are only loosely related to the actual sentencing preferences of the public. Second, such questions assume, incorrectly, that the public actually knows what the sentencing practices of criminal courts are. In fact, it has been shown that the public overestimates the seriousness of the crime situation (Roberts and White 1986; Roberts and Doob 1990) and the leniency of the criminal courts themselves (Roberts and Doob 1989).

Although opinion polls cannot be considered a relevant data base for assessing differences (if any) in sentencing preferences between criminal courts and civil society, case-sentencing surveys were designed precisely for that purpose. A detailed review of these surveys (e.g., Makela 1967; Gibbons 1969; Boydell and Grindstaff 1972; Grindstaff 1974; Thomas, Cage, and Foster 1976; Blumstein and Cohen 1980; Samuel and Moulds 1986; Thomson and Ragona 1987; Zimmerman, Van Alstyne, and Dunn 1988; Walker, Collins, and Wilson 1988; Diamond and Stalans 1989; Roberts and Doob 1989; Stalans and Diamond 1990) is beyond the scope of this paper. …