In criminal justice, consistency is an essential element as it reflects fairness and dependability of the justice system. In contrast, lack of uniformity in judicial sentencing decisions undermines the credibility of the judicial system and the legal process. The lack of a higher legal authority to scrutinize the sentencing decisions is largely responsible for some of these inconsistencies. In this research, the statistical model used evaluates the proportionality, consistency, and non-discrimination in the judicial sentencing decisions.
The dependent variable will be the sentencing decisions in the drug offender cases. In particular, elements of the sentencing decisions: form of sentence chosen, favorability of the sentence and the severity of the sentence, will be analyzed. The independent variables for this research will involve the demographic factors such as gender, age, ethnicity, socioeconomic status of the offenders and community factors of the offenders and the sentencing judge, the legal issues in the case and the mandatory sentencing guidelines. This research will use a comparative statistical analysis methodology to analyze the variables.
The comparative analysis methodology brings together aspects of the offender and the offense. This aids in the determination of the severity of punishment and the consistency of judicial decisions. Therefore, a primary rationale for the comparative weighting of the variables is to determine the consistency, discrimination and proportionality issues of judicial sentencing decisions. In this research, the sentencing decisions will be defined in terms of the two distinct elements of consistency. First, “sentences will be consistent if offenders with similar attributes receive similar sentences. Second, sentences will be consistent if different offenders receive dissimilar sentences proportional to their level of dissimilarity” (Bushway et al. 2007, p. 176). Consistency underlies any analytic framework for examining the applicability of the sentencing guidelines. As Greene puts it, “the goal of examining consistency is to ensure that gross crimes receive a more severe punishment” (2000, p. 92). Indeed, offenders with comparable blameworthiness should receive more or less similar penalties. Additionally, this research will examine the seriousness of crimes and compare the previous crime records of offenders to determine dissimilar offenders and the extent to which they are dissimilar.
Inconsistency judicial sentencing is normally viewed as discrimination. It arises when judges fail to agree on the fundamental characteristics of the offense. To evaluate inconsistencies in sentencing, this research will integrate the weighted offender characteristics in the sentencing decision. In summary, this research will test three hypotheses, using a comparative analytic model, with regard to proportionality, consistency and discrimination. First, do offenders that fall in the same situations get the same convictions? Second, does the sentencing reflect the proportional distinctions between serious and less serious offenders? Third, is inconsistency in sentencing an indicator of discrimination?
In order to build a statistical sentencing model, this research will review two separate judicial decisions: prison sentence versus non-custodial prison sentence and the length of prison sentence. For accurate assessment of consistency and discrimination, the dependent variables should be defined clearly in these two decisions (2007, p. 167). In the first decision, the research will use a variable of zero for non-custodial sentence and one for a prison sentence. In the second decision, this research will use an algorithm derived from the sentence to assess the magnitude of the sentence (Breen, 1996, p. 84). This will involve the number of months of the offender’s prison sentence. Breen argues that, this measure has a distinct advantage; it reflects the actual length of the sentence served ignoring parole time or pretrial duration (1996, p. 89). The severity of the sentence-a dependent variable-is expected to correlate with the severity of the crime. Normally, in judicial decisions, the length of sentences increases with an increase in severity of the crime. In this view, the dependent variables regarding the sentencing decisions will follow this concept in order to draw reliable conclusions regarding consistency and discrimination in judicial sentences.
The joint estimation of the sentence magnitude and sentence type raises concerns about sample selection bias. This research will employ the maximum likelihood Heckman estimation as a sample selection model to minimize bias (Bushway et al., 2007, p. 170). Using exclusion restrictions as per the Heckman estimation, the departure variables with regard to sentence severity or sentence type can be identified. From the consistency perspective, through the use of an appropriate selection model, discernible patterns with regard to sentencing outcomes can be noted. In the analysis of sentence length, a sentence length equation will be most appropriate. According to Bushway et al., the equation should involve a non-logged independent variable, e.g. gender or race, and a logarithmic dependent variable (2007, p. 155). Accordingly, in this research, coefficients that measure the percent change in the dependent variable alongside one-unit change in the independent variables (age, gender, race, ethnicity among others) will be used.
With regard to offender characteristics, this research, using the sentence length equation and non-custodial sentences, assess the relationship between judicial sentences and characteristics such as age, race or gender. This will help identify potential discriminatory factors in judicial sentencing decisions.
Reference List
Breen, R. (1996). Regression Models: Censored, Sample Selected or Truncated Data. Thousand Oaks, CA: Sage Publications, Inc.
Bushway, S., Johnson, D., & Lee, A. (2007). Is the Magic still there? The use of the Heckman two step correction for selection bias in criminology. Journal of Quantitative Criminology, 23(2), 151-178.
Greene, W. (2000). Econometric Analysis. Upper Saddle River, NJ: Prentice Hall.