Summary:
In Utah, a recommendation of probation, jail, or prison should be based on an offender’s criminal history score and the severity and type of their most recent offense. Additionally, the offenders' mitigating/aggravating factors are taken into consideration. The criminal history score is based on the offender’s past involvement with the justice system. The score then places offenders into a category ranging from 1 to 5, with 5 being the most severe. Based on these factors, the offender is placed in one of three areas of a sentencing matrix. The criminal history score and the sentencing matrices are developed and published by the Utah Sentencing Commission. The row of the sentencing matrix denotes the level of the offender’s criminal history, while the columns denote the severity and type of their current offense. Placement on the sentencing matrix is determined by probation officers and included in a Pre-Sentence Investigation report, which is shared with the defense attorney and the prosecutor, who can review it for errors. The report is also sent to the Judge who will make the final decision regarding the nature and length of the offender’s sentence. While the PSI report itself does not determine an offender’s sentence type or length, it is an important document involved in the Judge’s decision-making process. It should be emphasized that the writer of the PSI report has no control over the offender’s initial charges and their associated convictions.
Data & Method
9,788 unique records were obtained from the Utah Department of Corrections (UDOC) database, O-Track, between October 2015 and September 2017. Here, a unique record is defined as a PSI report written on a unique date. Hence, an offender may have multiple PSI reports within the given time period, each representing a unique date-and-conviction combination. Criteria for inclusion comprise being convicted of a felony offense, a valid criminal history score, and a sentence recommendation of prison, jail only, or probation.
Demographics, including age and gender, are included as important control variables. The relevant criminal justice-associated variables include the type and severity of the current offense, as well as the offender’s criminal history category. The type of the current offense denotes the highest offense associated with the current case and is classified either as drug possession only, drug and/or alcohol, driving, property, person, murder, or a category called “other.” Offense severity indicates whether the current offense was a first-degree, second-degree, or third-degree felony, with first-degree being the most severe. The variable denoting the offender’s race and ethnicity comprises four different binary variables. These include White, Hispanic, Other, and Unknown. The dependent variable denotes whether the offender received a prison recommendation versus a jail-only/probation recommendation, with prison being the most severe. A description of all variables is provided in Table 1.
Severity of New Convictions
The logistic regression’s marginal effects, fixed at each level of offense severity and the criminal history categorical score, and evaluated at the mean values of the other covariates, are shown in Figure 3. As seen, a difference between Hispanics and Whites is evident across the three levels of offense severity and the five criminal history categories. The disparity appears to be particularly elevated with PSI reports involving Felony 2 offenders and reports where the offender had a criminal history categorical score of I and II. In terms of offenses, one-third of PSI reports can be expected to include a prison recommendation for a Hispanic individual, while only 20 percent of PSI’s completed for White individuals can be expected to include a prison recommendation (a 33 percentage change difference). Similarly, for criminal history category II, 9 percent of PSI’s involving a Hispanic offender can be expected to include a prison recommendation, which is reduced to 6 percent for Whites (a percentage change difference of 50%). The figure further shows these marginal effects for the sample’s most common offense types, namely, property (37% of the sample) and alcohol & drug-related offenses (at 17%), while again, holding the other covariates fixed at their mean values. In comparison to PSI reports involving Whites, Hispanics are more likely to receive a prison recommendation for a property-related offense (22% vs. 15%, translating to a 47 percentage change difference). Similarly, Hispanics are more likely to receive a prison recommendation for an alcohol & drug-related offense (25% vs. 17%, translating to a 47 percentage change difference).
Policy Implications
It should be emphasized that this study focuses solely on decisions made at the Pre-Sentence Investigation level and, therefore, does not address disparities that may have occurred at prior criminal justice contact points. Keeping this in mind, the racial and ethnic disparities identified in this study may result from current policies and guidelines governing the PSI writing process. For example, a review of current mitigating and aggravating factors revealed that many are tied to an individual’s financial resources. Specifically, having access to more financial resources is linked to several mitigating and hence, sentence reducing factors, are often not available to non-White individuals.14 These built-in policies may have unintentionally created room for racial and ethnic disparities to occur. Analyzing the extent to which current mitigating and aggravating factors explain differences in outcomes, and how the selection of these factors may vary across PSI writers, is an important first step in addressing these disparities.
Read the full report:
Race-Ethnicity & Pre-Sentence Prison Recommendation