Exceptions to the Rule? Exploring the Use of Overrides in Detention Risk Assessment

Abstract

The Juvenile Detention Alternatives Initiative (JDAI) aims to reduce the use of predispositional detention for juveniles. One core strategy of JDAI is the use of risk assessment instruments to ensure that detention decisions are made objectively. These instruments allow for mandatory and discretionary overrides based on statutory guidelines, aggravating factors, or mitigating factors. This study investigates the use of overrides in a jurisdiction that utilizes the JDAI model. Offense seriousness, prior record, history of escape/runaway, and age were consistent predictors of overrides. African Americans were less likely to receive mitigating overrides, and females were less likely to receive mandatory overrides.

Keywords

juvenile justice Juvenile Detention Alternatives Initiative inequality overrides

Introduction

Risk assessment has long been used in criminal justice to predict future behavior, increase objectivity, equity, and consistency in decision making, and facilitate the efficient use of resources (Gottfredson & Moriarty, 2006). Risk assessment has been used in the juvenile justice system to guide dispositional decisions, determine supervision levels, and inform treatment regimens (Hoge, 2002; Schwalbe, 2008). Most jurisdictions also use risk assessment in predispositional detention decision making. Predispositional detention, the incarceration of a juvenile prior to an adjudicatory hearing, has been the subject of much scholarly debate. Research suggests that detention has a negative impact on youth development, is plagued by disproportionate minority contact and race/gender inequities (Bishop, 2005; Leiber & Fox, 2005), and is exorbitantly expensive (Mendel, 2011). Thus, risk assessment helps to ensure that detention is reserved for only the most serious offenders.

In over 100 jurisdictions, risk assessment in the detention decision is implemented as a core component of the Juvenile Detention Alternatives Initiative (Mendel, 2011). JDAI is a widely recognized juvenile justice reform effort developed by the Annie E. Casey Foundation focused on reducing the reliance on secure confinement, reducing racial disparities, improving public safety, and saving taxpayer dollars (Mendel, 2011). Under JDAI, risk assessment instruments are developed and validated at the local level and vary by site (Steinhart, 2006). Instruments used within JDAI sites are point-scale instruments; that is, points are accumulated based on factors known to be associated with an increased risk of failure to appear and threat to public safety (e.g., offense seriousness, prior record), generating a total risk score which indicates whether the youth should be released, monitored, or detained (Steinhart, 2006). Importantly, such instruments allow staff to make exceptions to the decision indicated by the instrument when aggravating or mitigating factors are present; these are known as overrides.

While one of the goals of risk assessment is to increase objectivity and equity in decision making, overrides are avenues through which discretion, subjectivity, and bias may be reincorporated into the detention decision process. A large body of research has found racial disparities in the detention decision (Leiber, 2003; Leiber & Fox, 2005; Leiber & Johnson, 2008; Leiber & Mack, 2003; Pope & Feyerherm, 1992; Rodriguez, 2010), and because those who are detained are more likely to be formally processed, disparities at the detention decision create a state of cumulative disadvantage as juveniles progress through the system (Bortner & Reed, 1985; Frazier & Bishop, 1985; Kempf-Leonard, 2007; Leiber, 2012; Leiber & Fox, 2005; Pope & Feyerherm, 1992). Indeed, Steinhart (2006) warns that high override rates may unintentionally increase disproportionate minority contact; thus override criteria must be race neutral.

There is a gap in the literature concerning the use of overrides in risk assessment generally and in JDAI sites specifically. The purpose of the current study was to address this void in the literature by examining the use of overrides in the predispositional detention decision in one juvenile court that has been implementing JDAI for over 5 years. Based on data collected on delinquency offenses postimplementation of JDAI from 2005 to 2009, this study addressed two specific research questions: (1) how often are overrides used? and (2) how are legal and extralegal (i.e., age, race, gender) factors related to the use of overrides?

Literature Review

Risk Assessment and Overrides

Risk assessment instruments have been increasingly used in predispositional detention decisions as well as in the dispositional phase of the juvenile justice decision-making continuum, in order to increase objectivity and consistency of decision making (Hoge, 2002). According to Hoge (2002), screening instruments have the potential to dramatically reduce the influence of personal prejudices and biases in decision making, which are often reflections of the overutilization of decision maker discretion. The potential to reduce bias has obvious implications for the effective management of disproportionate minority contact.

Despite the intended objectivity of risk assessment instruments, overrides create avenues through which discretion, subjectivity, and bias, may be reincorporated into the detention decision. According to Steinhart (2006), “an override is a decision to detain or release a child in contravention of the risk score and outcome recommended by the risk assessment instrument” (p. 22). Jurisdictions with high override rates are problematic in that the integrity of the risk assessment process may be compromised (Steinhart, 1994; Virginia Department of Juvenile Justice, 2004). Though no stated limit has been imposed on release overrides (i.e., instances where the risk assessment instrument recommended detention but instead the youth was released), the National Council on Crime and Delinquency has proposed that detain overrides (i.e., instances where the risk assessment instrument recommended release but detention was administered) should be at or less than 15% of youth whose risk assessment instrument score indicated release (Steinhart, 2006). During initial implementation of an risk assessment instrument, however, it is common to find rates that are in excess of 50% of youth who qualify for release (Steinhart, 2006).

Although there is a paucity of literature specifically examining overrides, some states have addressed the practice of overrides in their risk assessment instrument evaluations. A few studies have found that risk assessment instruments work as intended when they are followed. For example, a study of San Francisco’s risk assessment instrument, developed by the National Council on Crime and Delinquency, showed that when the risk assessment instrument indicated and actual decisions matched, positive outcomes were generally reported. For example, only 3 of the 227 youth referred to the Juvenile Probation Department during the study period whose indicated and actual decision was release were rearrested during the follow-up period (indicating a 94% success rate). Additionally, there was no incidence of failure to appear at a court hearing for those juveniles who were released based on their risk assessment instrument score (Steinhart, 1994). Similarly, in a 2007 assessment of public safety outcomes associated with the Virginia Detention Assessment Instrument, Reiner, Miller, and Gangal (2007) found that compliance with the Detention Assessment Instrument generally conferred positive outcomes. For instance, the new offense rate was only 6.5% when the indicated decision matched the actual decision. However, there was an 11.1% new offense rate when a mitigating override was used. The new offense rate associated with aggravating overrides was 4.9%. The study ultimately found ‘sound validation of the public safety outcomes of the Detention Assessment Instrument’ (Reiner, Miller, & Gangal, 2007, p. 37). These two studies suggest that the risk assessment instruments were operating as intended with regard to identifying those youth who are a threat to public safety, and that overrides should be used sparingly.

Although overrides should be used in a limited fashion, there are specific instances when the use of an override may be warranted. For example, an override may be necessary in cases when a parent or guardian is unwilling or unable to care for the youth, when the youth is a danger to himself or herself or others, or when the youth is involved with a gang (Steinhart, 2006). Additionally, overrides may be justified when a youth has had marginal involvement in an offense or when there are significant substance abuse or mental health issues. Steinhart (2006) notes, however, that aggravating and mitigating override criteria may be particularly problematic as they may unintentionally introduce disparities that cause disproportionality. For instance, the inability of family to provide supervision may be an aggravating factor that disproportionately affects minority youth, leading to higher rates of detained minorities.

Some research has examined the impact of overrides on disproportionality. In 2003, the Virginia Department of Juvenile Justice (2004) conducted an evaluation report of the Detention Assessment Instrument. This report was intended to assess the initial implementation phase of the Detention Assessment Instrument over a 12-month period in 35 Court Service Units (CSUs).¹ The statewide override percentage was 54.7%.² Related to disproportionate minority contact, the study found that of the 3,937 instances when an override to a more restrictive placement was authorized, 53.2% of these cases were non-White offenders, indicating disproportionality. On the other hand, Simpson’s (2010) evaluation of the risk assessment instrument implementation on detention practices in a rural parish in Louisiana found that risk assessment instruments were not consistently being used in the detention screening process and that officers continued to subjectively evaluate risk which resulted in overrides. Interestingly, the study found that overrides tended to favorably impact minority youth. According to Simpson (2010), ‘black youth were not harmed by the use of mandatory and administrative overrides and were actually helped by the use of informal overrides as law enforcement consistently chose to release these youth’ (p. 54).

Due to the dearth of research on overrides specifically, it is useful to reference research on sentencing guideline departures for theoretical and empirical guidance. Like risk assessment, sentencing guidelines are used to increase objectivity in sentencing decisions. They also allow for both upward and downward ‘departures’ to allow judges to consider aggravating and mitigating factors. The majority of prior research has found that legal characteristics, such offense seriousness and prior record, are the primary determinants of sentencing departures (Johnson, Ulmer, & Kramer, 2008). However, other research has found that departures are influenced by extralegal variables, such as race, gender, and age. For example, Johnson (2005) found that males, minorities, and young offenders were more likely to receive upward departures. Similarly, many studies have found that minority status and being male reduce the likelihood of receiving a downward departure (see Engen, Gainey, Crutchfield, & Weis, 2003; Johnson, 2005; Johnson et al., 2008).

Attribution theory has been used to understand the decision-making process that occurs in sentencing guideline departures in the criminal justice system (see Albonetti, 1991). Johnson (2005) notes that under the constraints of time and sentencing variability, judges and other courtroom personnel may use ‘decision-making shortcuts, or patterned responses, that link extralegal characteristics to the likelihood of future criminality’ (p. 767). In this way, stereotypes or preconceived notions of culpability associated with demographic characteristics of offenders, such as race, age, and gender, may infiltrate the decision-making process. Similarly, in the juvenile justice literature, the symbolic threat thesis has been used to explain that stereotypes based on race, age, and/or gender may impact case workers perceptions of juveniles; specifically, Black males may be seen as more threatening to public safety (for an overview of the symbolic threat thesis, see Leiber & Fox, 2005). Accordingly, males and older youth may also seem more culpable than females when assessing the motivations behind delinquency and therefore may be more likely to receive severe outcomes. Attribution theory and the symbolic threat hypothesis may help us to understand disparities found in override rates.

Study Site

As part of JDAI’s focus on the reduction of utilization of secure detention, risk assessment instruments are used at all JDAI sites “to guide the intake officer in making the critical decision whether to detain or release an arrested youth” (Steinhart, 2006, p. 5). In Virginia, the Detention Assessment Instrument, developed by Virginia Department of Juvenile Justice in 2002, is used to guide detention decision making (see Appendix).³ The Detention Assessment Instrument is assumed to help prevent the detention of juveniles who are not serious or chronic offenders, who do not represent a community threat, and who are not at risk for failure to appear. It is also presumed to ‘improve consistency and equity in detention utilization, both within each Court Service Unit and across different Court Service Units’ (Virginia Department of Juvenile Justice, 2004, p. 9). The CSU being studied became a JDAI site in October 2005. From 2004 to 2007, there was a 27% reduction in average daily detention population, a 27% reduction in length of stay, and a 39% reduction in predispositional detention admissions.

A Detention Assessment Instrument is completed by an intake officer on every juvenile charged with a delinquent offense.⁴ Scoring is based on a number of legal variables, including the seriousness of the current offense, prior record, and supervision status.⁵ The score generated by the instrument indicates that the juvenile is either (a) released, (b) assigned to a detention alternative (e.g., supervision, house arrest), or (c) securely detained. Importantly, the Detention Assessment Instrument allows for intake staff to ‘override’ the indicated decision; that is, intake staff have the discretion to make an exception and detain or release a child in ‘contravention’ of the outcome recommended by the Detention Assessment Instrument (Steinhart, 2006). Our analysis examines three types of overrides: mandatory, discretionary aggravating, and discretionary mitigating.⁶ Mandatory overrides are upward overrides used in the following instances: (a) use of a firearm in current offense, (b) escapee from secure placement, and (c) local court policy.⁷ Discretionary overrides may be aggravating⁸ or mitigating.⁹

Data and Method

Research Design and Analytic Plan

This study used data compiled by the Virginia Department of Juvenile Justice. The data set includes the entire population of juveniles who were processed through the CSU’s intake department in one Juvenile and Domestic Relations Court from October 2005 to June 30, 2009. Status offenders were removed from the sample because they are not eligible for detention. Additionally, 1,885 cases did not have Detention Assessment Instrument information available and were therefore removed from the original sample yielding a final sample size of 4,683 cases.¹⁰

First, we provide descriptive statistics (see Table 1 for descriptive statistics on the full sample and by each type of override). Second, we provide information on the override rates by year because research suggests that override rates may be high initially and then decrease over time. Third, multinomial logistic regression is used to determine the influence that legal and extralegal variables have on the likelihood of receiving an override versus the Detention Assessment Instrument indicated decision. We analyzed the extent to which override decisions are made equitably with regard to race, gender, and age while controlling for relevant legal variables.

Table 1.

Descriptive Statistics.

	Full sample	Aggravating Overrides	Mitigating Overrides	Mandatory Overrides
N (% of Sample)	4683	208 (4.4)	343 (7.3)	104 (2.2)

Age	Mean = 15.4	Mean = 15.72	Mean = 15.46	Mean = 15.91
	SD = 1.55	SD =1.64	SD =1.46	SD =1.53
	Range =8-21	Range =11-19	Range =10-19	Range =11-20
Race
White	.13	.14	.15	.11
Black	.84	.82	.82	.88
Other	.03	.04	.03	.02
Gender
Female	.29	.3	.23	.09
Offense Seriousness
Felony	.32	.14	.48	.25
Offense Type
Person	.42	.37	.33	.06
Drug	.04	.04	.12	.07
Property	.26	.17	.32	.14
Other	.28	.42	.23	.73
Additional Charges
Yes	.43	.38	.49	.43

Prior Referrals (square root)	Mean =1.72	Mean = 2.31	Mean = 2.35	Mean = 1.96
	SD =1.33	SD =1.39	SD = 1.30	SD = 1.24
	Range =0-6.40	Range = 0-6.24	Range = 0-5.74	Range = 0-4.58
Prior Felonies
Yes	.31	.42	.56	.52
Supervision Status
Yes	.26	.45	.45	.39
History of Escape/ Runaway
Yes	.08	.29	.06	.17
History of Failure to Appear
Yes	.08	.13	.13	.06

Variables

Dependent variable

Overrides are operationalized as a categorical variable coded as either mitigating (=1), mandatory (=2), or aggravating (=3); as such, multinomial logistic regression was used in order to analyze the data (Detention Assessment Instrument indicated decision is the reference category). Of the 4,683 cases, 13.9% involved overrides. Roughly 52% of all overrides were mitigating, 16% were mandatory, and 32% were aggravating.

Independent variables

Age is a continuous variable measuring actual age in years and the average age was about 15 years old. Race is operationalized as a series of dummy variables, representing Black (1 = yes; 0 = no), White (1 = yes; 0 = no), and Other (1 = yes; 0 = no) juvenile offenders, with 84% Black, 13% White, and 3% other.¹¹ Gender is dichotomous (1 = female, 0 = male), with 70.7% of the sample being male and 29.3% female.

Control variables

We include a number of control variables including: current offense (seriousness and type), additional charges at current intake, prior record, supervision status, history of escape or runaway, and history of failure to appear.

Offense seriousness is measured by whether the current offense is a felony (1 = yes; 0 = no) or not. Offense type is a series of dummy variables (1 = yes, 0 = no) representing (1) person, (2) drug, (3) property, and (4) other, with ‘other’ serving as the reference category. The presence of additional charges at the time of the current referral is operationalized as either having additional felony, misdemeanor, or probation/parole violation offenses or not having additional felony, misdemeanor, or probation/parole violation charges (1 = yes; 0 = no). Prior record was measured with two variables. The first, prior referrals, measures the number of times a juvenile has been processed through intake in the past. Due to skewness, we used the square root of this measure in all analyses. Second, prior felonies is a dummy variable indicating whether the juvenile had a history of prior felony offenses (1 = yes; 0 = no). Current supervision status is dummy coded (1 = yes; 0 = no) and measures whether the juvenile was currently under court supervision or not currently under court supervision. History of escape or runaway is dummy coded (1 = yes; 0 = no) and indicates whether the juvenile has a past history of runaway or escape. Finally, history of failure to appear is dummy coded (1 = yes, 0 = no) and indicates whether the juvenile has ever failed to appear to a court hearing.

Results

Overrides Rates by Year

According to Steinhart (2006), detain overrides should be at or less than 15% of youth whose indicated decision is release. However, it is common to find rates that are much higher than that in the early implementation of a risk assessment instrument. Since JDAI began in 2005, we examined total override rates for 4 years, beginning in fiscal year (FY) 2006 (table available upon request). In FY 2006, there were 343 total overrides (32.5% of cases); in FY 2007, it decreased to 122 (8.3% of cases), and in FY 2008, it was the lowest at 83 (7.9% of cases). In FY 2009, there were 107 overrides (11.8% of cases). Each type of override generally decreased each year, with the exception of mandatory overrides, which generally increased over time.

Predictors of Detention Assessment Instrument Overrides

Table 2 displays the results of the multinomial regression analysis.¹² The pseudo r² for the model is .163. Regarding receiving an aggravating override versus the Detention Assessment Instrument indicated decision, offense seriousness, prior referrals, history of escape/runaway, and age are significant. In particular, the odds of receiving an aggravating override decreased by over 60% for those charged with a felony offense compared to those charged with a misdemeanor or other offense. Those with prior referrals were significantly more likely to receive an aggravating override and those who had previous escape/runaway attempts had more than 3 times increased odds of receiving an aggravating override. Additionally, older youth were more likely to receive an aggravating override, with increased odds of 13% for each additional year of age. This change is nonlinear, so while a 15 year old would have about 28% increased odds of receiving an aggravating override compared to a 13-year-old juvenile, a 17-year-old youth would have 84% increased odds of receiving an aggravating override compared to a 13 year old (see also Lottes, Adler, & DeMaris, 1996).

Table 2.

Multinomial logistic regression predicting overrides (N=4683).

	Aggravating Override	Mitigating Override	Mandatory Override
Variable	Versus Indicated Decision	Versus Indicated Decision	Versus Indicated Decision
Offense Seriousness
Felony	-.99 (.37)**	.54 (1.71)**	.78 (2.19)*

Offense Type
Person	.02 (1.02)	-.12 (.89)	-3.29 (.04)**
Drug	.11 (1.11)	1.02 (2.76)**	-1.15 (.32)*
Property	-.22 (.81)	.15 (1.16)	-2.19 (.11)**

Additional Charges
Yes	.11 (1.12)	.095 (1.10)	.41 (1.51)

Prior Referrals	.15 (1.17)*	.16 (1.17)**	-.28 (.75)*

Prior Felonies
Yes	-.012 (.99)	.65 (1.91)**	.73 (2.08)*

Supervision Status
Yes	.45 (1.57)*	.67 (1.95)**	.24 (1.27)

History of Escape/Runaway
Yes	1.44 (4.23)**	-.52 (.60)**	1.03 (2.81)**

History of Failure to Appear
Yes	.03 (1.03)	.09 (1.09)	-.74 (.48)

Age	.12 (1.13)*	-.09 (.92)*	.17 (1.19)*

Race
Black	-.25 (0.78)	-.40 (.67)*	.13 (1.14)
Other	-.06 (.94)	-.14 (.87)	-.52 (.59)

Gender
Female	-.09 (.92)	.16 (1.17)	-1.23 (.29)**
-2 Loglikelihood	4065
Pseudo R-Square	.163

Note. Numbers represent coefficients and (odds-ratios)

*p<0.05 **p<0.01; p-values computed for two-tailed significance tests

When comparing mitigating overrides to receiving the indicated decision, offense seriousness, offense type, prior intake referrals, prior felonies, supervision status, history of escape/runaway, age, and race all proved to be significant predictors. The odds of receiving a mitigating override increased almost 2 times for those charged with a felony and 3 times for those charged with a drug offense, compared to those charged with misdemeanors and disorderly conduct offenses, respectively. The odds of receiving a mitigating override increased almost 2 times for both having prior felonies and being on probation (supervision) at the time of offense. Juveniles with prior intake referrals had increased odds of receiving a mitigating override, while those with a history of escape/runaway had decreased odds of receiving a mitigating override versus receiving the indicated decision. Furthermore, older juveniles and African American juveniles each had decreased odds of receiving a mitigating override versus receiving the indicated decision. Specifically, African American youth had decreased odds of receiving a mitigating override of about 33%, compared to White youth. Additionally, each additional year in age corresponds to about an 8% decrease in odds of receiving a mitigating override.

Finally, turning to mandatory overrides, offense seriousness, offense type, prior intake referrals, prior felonies, history of escape/runway, age, and gender were significant predictors. Those juveniles charged with felonies had 2 times increased odds of receiving a mandatory override, while those charged with person, property, and drug offenses had significantly decreased odds of receiving a mandatory override, compared to those charged with other offenses. Juveniles with prior intake referrals were less likely to receive a mandatory override while those with prior felonies were more likely to receive a mandatory override versus receiving the indicated decision. For juveniles with a history of escape/runway, the odds of receiving a mandatory override increased almost 3 times. Older juveniles were significantly more likely to receive a mandatory override, with each additional year in age resulting in about 19% increased odds, while females’ odds of receiving a mandatory override decreased by about 70% compared to males.

Discussion

This study aimed to investigate how overrides were used in the predispositional detention decision in one juvenile court in Virginia that has been implementing JDAI for over 5 years. Offense seriousness, prior record, and history of escape and runaway were consistent predictors of each type of override, but there were also differences by age, race, and gender. Specifically, older juveniles were more likely to receive aggravating and mandatory overrides and less likely to receive mitigating overrides. Additionally, African Americans were roughly 33% less likely to receive a mitigating override than White youth and females were about 70% less likely to receive mandatory overrides than male juveniles.

Those charged with a felony were less likely to receive an aggravating override, which makes sense because some felonies (alone) carry detainable scores (i.e., score of 15); thus aggravating factors are not considered. Those with prior referrals or on supervision, those with a history of escape/runaway and older juveniles were more likely to get an aggravating override. Recall that aggravating overrides are generally used for offenders who have a particularly violent history, significant mental health, or substance abuse problems, or because there is no capable or willing guardian. In terms of age, older juveniles were more likely to receive an aggravating override. Some research indicates that younger juveniles are viewed as less culpable and that detention is used more often for older juveniles (see Bishop & Frazier 1992, 1996; Guevara, Herz, & Spohn, 2006; Leiber & Mack, 2003; Leiber, Brubaker, & Fox, 2009). Juvenile justice staff may view younger youth as less blameworthy due to factors such as immaturity and inexperience and therefore more amenable to rehabilitation (Champion, 2001).

Those charged with a drug offense were more likely to receive a mitigating override. Some drug offenses carry a high score (e.g., felony drug carries score of 12 while only 15 is needed for detention) but as they may be nonviolent, those juveniles may be deemed as a low risk by intake workers and not in need of secure detention. Those charged with a felony, those with prior referrals, prior felonies, and those on supervision were more likely to get a mitigating override. Again, this makes sense because those with felonies (and other point accumulating characteristics like prior felonies, etc.) are more likely to earn a score high enough to indicate secure detention without receiving an override, so if mitigating factors (e.g., marginal involvement in offense, willing/able guardians) are present, they are more likely to receive this type of override. Those with escape/runway were less likely to get a mitigating override. Older juveniles and African Americans were less likely to get a mitigating override.

African Americans are not enjoying the leniency afforded to White juveniles with regard to mitigating overrides. Recall that mitigating overrides are used in cases where the juvenile was marginally involved in the offense, the parent or guardian is willing and able to provide adequate supervision, the child regularly attends school, or the offense is less serious than indicated by the charge. There are a number of possible explanations for this finding. Non-Whites may be differentially involved in delinquency; for example, they may be less likely to be deemed marginally involved in the offense (Lauritsen, 2005; Tracy, 2005). Certainly, research shows that risk factors linked to violent crime are greater for non-Whites, and while White and non-White youth display similar rates for offending for less serious crimes, Black youth are disproportionately involved in more serious violent crime. Additionally, socially disorganized areas experience greater police presence and increased arrest rates (Chapman, Desai, Falzer, & Borum, 2006). Thus, legal measures, such as gang involvement and prior record, may be confounded by social factors that disproportionately impact non-Whites.

It is also plausible that extralegal factors, such as school attendance and parental supervision, are to blame. Non-Whites may be less likely to report good school attendance and are therefore less likely to benefit from overrides granted for this purpose. Similarly, according to Steinhart (2006), overrides are often determined by the ability and willingness of a family member or guardian to provide supervision, and this may vary by race. Accordingly, he warned that overrides must be monitored so that they do not unintentionally disproportionately impact minorities. This is a policy that while race neutral on its face appears to be having a disparate impact on African American juveniles.

Recall that mandatory overrides are reserved for juveniles who have escaped from secure placement, use of a firearm in commission of the crime, or local court policy. The only local court policy in place during the time of the study was possession of a weapon (this refers to any instrument that may be used as a weapon). There were very few escapes from secure confinement, so for the present purpose, mandatory overrides were used for weapons—possession or use, which may or may not be a felony, may or may not be in conjunction with a person offense, and so on. Juveniles charged with felonies were more likely to get a mandatory override while those charged with person, drug, or property offenses were significantly less likely to receive a mandatory override. Those with prior referrals were less likely to get mandatory overrides, but those with prior felonies and history of escape/runaway were more likely to get mandatory overrides. Finally, all else constant, younger juveniles, and females were less likely to get mandatory overrides. This finding would suggest that younger juveniles and females were less likely to get arrested for weapons offenses and parallels recent trends. In a national report on juvenile offenders and victims, Snyder and Sickmund (2006) found that in 2003 females made up only 11% of total juvenile weapons arrests and older youth (ages 16–17) made up the majority of juvenile weapons arrests (64%). Additionally, much research on detention has found that females enjoy leniency compared to males (Bishop & Frazier, 1992, 1996; Johnson & Scheuble, 1991; Kurtz, Linneman, & Spohn, 2008; McCord, Widom, & Cromwell, 2001).

The findings with regard to race, gender, and age provide support for attribution theory and the symbolic threat hypothesis. It is possible that, despite the use of the screening instrument, bias is impacting the decision to grant overrides. It is plausible that discretionary override decisions by caseworkers are influenced by stereotypes about dangerousness and culpability based on demographic characteristics. Bridges and Steen (1998) found that probation officers used different ‘attributions’ to explain the delinquent behavior of White and Black youth. For example, delinquency by Black youth was attributed to internal factors (e.g., lack of morality and discipline) and delinquency by White youth was attributed to external factors (e.g., impoverished neighborhood). Such causal attributions influenced perceptions about likelihood of recidivism and therefore resulted in harsher punishments for Black juveniles. Given the low proportion of White juveniles in the juvenile court in this jurisdiction (13%), it is plausible that Whites are more likely to benefit from positive attributions; in other words, rather than negatively stereotyping minorities, positive attributions are leading to the increased use of mitigating overrides for Whites.

Policy Implications

The present study suggests several policy implications. JDAI (and other such initiatives) aim to ensure equity in decision making; if decision makers are overriding decisions indicated by validated risk assessment instruments, the potential exists to introduce subjectivity and bias. When overrides occur with any regularity, risk assessment instruments should be revisited to ensure that the instrument sufficiently captures characteristics that are integral to the detention decision. To the extent possible, patterns in overrides should be incorporated into the Detention Assessment Instrument to reduce the need for overrides and therefore eliminate the opportunity for bias and stereotypes to influence detention decisions.

Practitioners should address the use of overrides for drug-related offenses. Individuals with drug-related offenses were 3 times more likely to receive a mitigating override in this CSU. This may suggest that caseworkers feel that drug-related offenses do not warrant predispositional detention. If drug-related offenders are more likely to receive an override, the Detention Assessment Instrument should be changed to reflect the local court culture. For instance, the Detention Assessment Instrument may incorporate a nonviolent drug offender category which carries a lower score than violent drug offenses. Presently, the high level of overrides associated with drug-related crimes suggests inconsistency and subjectivity in decision making when dealing with this class of offender.

Regarding race, African American youth were less likely to receive a mitigating override. It is likely that mitigating factors vary by race and result in different override outcomes for African Americans. However, one study in Virginia found that ‘African American and Caucasian youth have similar risk patterns’ (Charlottesville/Albemarle Commission on Children and Families Task Force on Racial Disparity and Disproportionality in Youth Services, 2011, p. 6) across a variety extralegal factors. If this is true in the current study site, it could suggest that the differential use of overrides between racial groups is due to bias. Indeed, racial disparities have been present throughout the history of the justice system (Gabbidon & Green, 2005). Future studies should be directed toward unraveling the reasons behind such differences.

Finally, policy could be amended to reduce the differential use of overrides by age. This CSU may consider revising the Detention Assessment Instrument to assign a risk category to age as opposed to the subjectivity inherent in overrides. For instance, the screening instrument used in Louisville metro area adds 2 points to the total Detention Assessment Instrument score if a felony is committed by a youth that is 16 or older. For those that are under 16 and commit a felony only 1 point is added (see Dedel & Davies, 2008).

Limitations and Future Research

A key limitation in the current study was the unavailability of information related to the reason an override was given. Thus, it was impossible to determine which types of mitigating/aggravating factors were most important in making override recommendations. Future research should compile the staff cited reasons for overrides. The Virginia Department of Juvenile Justice (2004) recommended that ‘intake narratives should be available for all overrides to secure detention, and should include a description of the steps considered to avoid placement into secure detention’ ¹³ (Virginia Department of Juvenile Justice, 2004, p. 3). This practice ensures accountability in decision making; it also provides a written record with which to troubleshoot potential areas of concern.

This study was also limited because it did not control for a variety of extralegal variables that have been found to be important in predicting juvenile justice outcomes in past research, such as family structure (Bishop, Leiber, & Johnson, 2010; Leiber, 2003, 2012; Leiber, Bishop, & Chamlin, 2011; Leiber & Johnson, 2008; Pope, 1995) and community characteristics (see Chapman et al., 2006; Lauritsen, 2005; Rodriguez, 2010). Future research on overrides should compare similarly situated youth across a variety of legal and extralegal attributes (Kempf-Leonard, 2007).

Finally, this study was unable to address more than one decision-making point. Past research has addressed the importance of examining more than one decision point (Bortner & Reed, 1985; Frazier & Bishop, 1985; Frazier & Cochran, 1986; Kempf-Leonard, 2007; Leiber, 2012; Leiber & Fox, 2005; Pope & Feyerherm, 1990, 1992). The decision to override a youth from a less restrictive placement to secure detention may result in more severe outcomes throughout the juvenile justice process. Future research should investigate the impact of overrides on later justice outcomes. Despite limitations, the current study was one of the first examinations of the use of overrides in detention risk assessment. Our findings help clarify the patterns and predictors of different types of overrides. Thus, our study addresses a void in the literature and should serve as a springboard for future research.

Footnotes

Appendix

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This research was partially supported by a summer research grant from the College of Arts and Letters at Old Dominion University.

Notes

References

Albonetti

C. A.

(1991). The symbolic punishment of white-collar offenders. In Bridges

Myers

(Eds.), Inequality, crime, and social control (pp. 269–282). Boulder, CO: Westview Publishers.

Bishop

D. M.

(2005). The role of race and ethnicity in juvenile justice processing. In Hawkins

D. F.

Kempf-Leonard

(Eds.), Our children, their children: Confronting racial and ethnic differences in American juvenile justice (pp. 23–84). Chicago, IL: The University of Chicago Press.

Bishop

D. M.

Frazier

C. E.

(1992). Gender bias in juvenile justice processing: Implications of the JJDP Act. The Journal of Criminal Law and Criminology, 82, 1162–1186.

Bishop

D. M.

Frazier

C. E.

(1996). Race effects in juvenile justice decision making: Findings of a statewide analysis. The Journal of Criminal Law and Criminology 86, 392–414.

Bishop

D. M.

Leiber

Johnson

(2010). Contexts of decision making in the juvenile justice system: An organizational approach to understanding minority overrepresentation. Youth Violence and Juvenile Justice, 8, 213–233.

Bortner

M. A.

Reed

W. L.

(1985). The preeminence of process: An example of refocused justice research. Social Science Quarterly, 66, 413–425.

Bridges

G. S.

Steen

(1998). Racial disparities in official assessments of juvenile offenders: Attributional stereotypes as mediating mechanisms. American Sociological Review, 63, 554–570.

Champion

(2001). The juvenile justice system: Delinquency, processing, and the law. Upper Saddle River, NJ: Prentice-Hall.

Chapman

J. F.

Desai

R. A.

Falzer

P. R.

Borum

(2006). Violence risk and race in a sample of youth in juvenile detention. Youth Violence and Juvenile Justice, 4, 170–184.

10.

Charlottesville/Albemarle Commission on Children and Families Task Force on Racial Disparity and Disproportionality in Youth Services. (2011). Final report. Charlottesville/Albemarle Commission on Children and Families (CCF). Retrieved from http://www.ccfinfo.org/PDFs/disparities_final_report.pdf.

11.

Chesney-Lind

(1977). Judicial paternalism and the female status offender: Training women to know their place. Crime and Delinquency, 23, 121–130.

12.

Dedel

Davies

(2008). Detention risk assessment instrument: Pilot test results final report to Kentucky Department of Juvenile Justice. Portland, OR: One in 37 Research.

13.

Engen

R. L.

Gainey

R. R.

Crutchfield

R. D.

Weis

J. G.

(2003). Discretion and disparity under sentencing guidelines: The role of departures and structured sentencing alternatives. Criminology, 41, 99–130.

14.

Frazier

C. E.

Bishop

D. M.

(1985). The pretrial detention of juveniles and its impact on case dispositions. The Journal of Criminal Law and Criminology, 76, 1132–1152.

15.

Frazier

C. E.

Cochran

J. C.

(1986). Detention of juveniles: Its effects on subsequent juvenile court processing decisions. Youth and Society, 17, 286–305.

16.

Gabbidon

Greene

H. T.

(2005). Race, crime, and justice. New York, NY: Routledge.

17.

Gottfredson

S. D.

Moriarty

L. J.

(2006). Statistical risk assessment: Old problems and new applications. Crime and Delinquency, 52, 178–200.

18.

Guevara

Herz

Spohn

(2006). Gender and juvenile justice decision making: What role does race play? Feminist Criminology, 1, 258–282.

19.

Hoge

R. D.

(2002). Standardized instruments for assessing risk and need in youthful offenders. Criminal Justice and Behavior, 29, 380–396.

20.

Johnson

Ulmer

J. T.

Kramer

(2008). The social context of guideline circumvention: The case of federal district courts. Criminology, 46, 711–783.

21.

Johnson

B. D.

(2005). Contextual disparities in guideline departures: Courtroom social contexts, guidelines compliance, and extralegal disparities in criminal sentencing. Criminology, 43, 761–796.

22.

Johnson

D. R.

Scheuble

L. K.

(1991). Gender bias in the disposition of juvenile court referrals: The effects of time and location. Criminology, 29, 677–699.

23.

Kempf-Leonard

(2007). Minority youths and juvenile justice: Disproportionate minority contact after nearly 20 years of reform efforts. Youth Violence and Juvenile Justice, 5, 71–87.

24.

Kurtz

D. L.

Linneman

Spohn

(2008). Investigating racial disparity at the detention decision: The role of respectability. Southwest Journal of Criminal Justice, 5, 140–157.

25.

Lauritsen

J. L.

(2005). Racial and ethnic differences in juvenile offending. In Hawkins

D. F.

Kempf-Leonard

(Eds.), Our children, their children: Confronting racial and ethnic differences in American juvenile justice (pp. 83–104). Chicago, IL: The University of Chicago Press.

26.

Leiber

(2003). The contexts of juvenile justice decision making: When race matters. Albany: State University of New York Press.

27.

Leiber

(2012). Race, pre- and post detention, and juvenile justice decision making. Crime and Delinquency, 1–23. doi: 10.1177/0011128709345970.

28.

Leiber

Bishop

Chamlin

(2011). Juvenile justice decision-making before and after the implementation of the disproportionate minority contact (DMC) mandate. Justice Quarterly, 28, 460–492.

29.

Leiber

Brubaker

S. J.

Fox

K. C.

(2009). A closer look at the individual and joint effects of gender and race on juvenile justice decision making. Feminist Criminology, 4, 333–358.

30.

Leiber

Fox

K. C.

(2005). Race and the impact of detention on juvenile justice decision making. Crime and Delinquency, 51, 470–497.

31.

Leiber

Johnson

J. D.

(2008). Being young and black: What are their effects on juvenile justice decision making? Crime and Delinquency, 54, 560–581.

32.

Leiber

Mack

K. Y.

(2003). The individual and joint effects of race, gender, and family status on juvenile justice decision-making. Journal of Research in Crime and Delinquency, 40, 34–70.

33.

Lottes

I. L.

Adler

M. A.

DeMaris

(1996). Using and interpreting logistic regression: A guide for teachers and students. Teaching Sociology, 24, 284–298.

34.

McCord

Widom

C. S.

Cromwell

N. A.

(Eds.). (2001). Juvenile crime, juvenile justice. Washington, DC: National Research Council and Institute of Medicine.

35.

Mendel

R. A.

(2011). No place for kids: The case for reducing juvenile incarceration. Baltimore, MD: The Annie E. Casey Foundation.

36.

Pope

C. E.

(1995). Juvenile justice in the next millennium. In Klofas

Stojkovic

(Eds.), Crime and justice in the year 2010 (pp. 267–278). Belmont, CA: Wadsworth.

37.

Pope

C. E.

Feyerherm

(1990). Minority status and juvenile justice case processing: An assessment of the research literature (Part I). Criminal Justice Abstracts, 22, 327–385.

38.

Pope

C. E.

Feyerherm

(1992). Minorities and the juvenile justice system: Full report. Washington, DC: Office of Juvenile Justice and Delinquency Prevention.

39.

Reiner

Miller

J. B.

Gangal

(2007). Public safety outcomes of Virginia’s detention assessment instrument. Juvenile and Family Court Journal, 58, 31–41.

40.

Rodriguez

(2010). The cumulative effect of race and ethnicity in juvenile court outcomes and why preadjudication detention matters. Journal of Research in Crime and Delinquency, 47, 391–413.

41.

Schwalbe

C. S.

(2008). A meta-analysis of juvenile justice and risk assessment instruments: Predictive validity by gender. Criminal Justice and Behavior, 35, 1367–1381.

42.

Simpson

(2010). Do objective measures reduce the disproportionate rates of minority youth placed in detention: Validation of a risk assessment instrument? (Doctoral dissertation). University of New Orleans, New Orleans, La.

43.

Steinhart

(1994). Objective juvenile detention criteria: The California experience. In Schwartz

Barton

(Eds.), Reforming juvenile detention: No more hidden closets (pp. 47–68). Columbus: Ohio State University Press.

44.

Steinhart

(2006). Juvenile detention risk assessment: A practice guide to juvenile detention reform. Baltimore, MD: Annie E. Casey Foundation.

45.

Snyder

H. N.

Sickmund

(2006). Juvenile offenders and victims: 2006 national report. Washington, DC: US Department of Justice.

46.

Tracy

P. E.

(2005). Race, ethnicity, and juvenile justice: Is there bias in post-arrest decision making? In Hawkins

D. F.

Kempf-Leonard

(Eds.), Our children, their children: Confronting racial and ethnic differences in American juvenile justice (pp. 300–350). Chicago, IL: The University of Chicago Press.

47.

U.S. Bureau of the Census. (2000). Virginia department of criminal justice services. Juvenile justice system disproportionate minority contact database. Retrieved from http://www.dcjs.virginia.gov/juvenile/dmc/index.cfm#3

48.

Virginia Department of Juvenile Justice (2004). 2003 Evaluation Report of the Detention Assessment Instrument: Implementation Phase. Richmond, VA:Virginia Department of Juvenile Justice.

49.

Virginia Department of Criminal Justice Services, Juvenile Justice System Race by Locality Database. Retrieved from http://www.dcjs.virginia.gov/juvenile/dmc/#1