The Sentencing Reform Act of 1984 mandated major restructuring of federal sentencing through specific sentencing guidelines. New sentencing guidelines developed by the United States Sentencing Commission and adopted in 1987 explicitly linked sentencing to "relevant conduct"-offense characteristics-and sought to abolish unwarranted sentence disparity. The guidelines substantially reduced judicial discretion and resulted in a criminalization and sentencing process that is largely prosecutor controlled. The author has generated hypotheses that relate defendant characteristics, guilty pleas, and departures from sentencing guidelines to sentence outcomes under the federal sentencing guidelines. She first examined the variables influencing sentence severity for the drug offenders who were sentenced in 1991-92. She then explored the interaction effects by estimating the tobit equation separately for three groups-black, white, and Hispanic defendants-to discover whether defendant's ethnicity conditions the effect of other defendant characteristics, guidelines-defined legally relevant variables, guilty pleas, and departures on sentence severity. Her analysis reveals that disparity in federal sentencing of drug offenders is linked not only to offense-related variables, as structured by the guidelines, but also to defendant characteristics such as ethnicity, gender, educational level, and noncitizenship, which under the guidelines are specified as legally irrelevant.
Under the Sentencing Reform Act of 1984, Congress established the United States Sentencing Commission and charged it with the task of designing a sentencing structure that would avoid "unwarranted sentencing disparity among defendants with similar records who had been found guilty of similar criminal conduct" (28 U.S.C. 991(b) (1) (B) (Supp. 1993)). In November 1987 the Federal Sentencing Guidelines were enacted. Among social scientists, legal scholars, and court officials, the sentencing guidelines ignited a debate over the legal and social consequences of the new structure of sentencing. The focus of my research is to explore empirically three issues that are central to the goals of federal sentencing reform and the policy debate that has emerged since state reform efforts began in the 1970s.
My first concern is to estimate empirically the direct effect on the length of imprisonment of defendant characteristics (e.g., ethnicity, gender, education, number of dependents, which are explicitly stated in the federal guidelines as legal irrelevant; see U.S. Sentencing Commission 1989: 5H1.1-5Hl.10). Drawing on my earlier work (Albonetti 1991), I specify hypotheses that reflect the merger of uncertainty avoidance/causal attribution theoretical perspectives.
My second concern is to estimate the direct effects on sentence outcomes of guilty pleas and of sentences that depart from the guidelines. Although the federal guidelines substantially reduce the wide latitude of discretion once enjoyed by sentencing judges, the guidelines do not restrict prosecutorial discretion. Numerous legal scholars and social scientists argue that the federal sentencing guidelines shift discretion away from sentencing judges to prosecuting attorneys (Tonry 1996; Standen 1993; Nagel & Schulhofer 1992, to name a few). Under the federal guidelines, a prosecuting attorney can circumvent the guidelinedefined sentence through charging, guilty plea negotiations, and motions for a sentence that is a departure from the guideline sentence. As noted by Nagel and Schulhofer (1992), the guidelines' emphasis on directly linking sentence outcomes to relevant conduct was intended as a structural constraint that would eliminate unwarranted disparity resulting from judicial control over sentencing. However, in the absence of constraints on prosecutorial discretion over charging decisions, guilty plea negotiations, and motions for "substantial assistance"'l departures, these …