Assessing Variations in Juvenile Court Processing in Urban Versus Rural Courts: Revisiting “Justice by Geography”

Abstract

Drawing on Feld’s (1991) “justice by geography” thesis, we examined whether juvenile court outcomes and case-level influences on those outcomes varied across urban and rural courts. Using a sample of 60,068 juvenile referrals across 66 counties in one state, we estimated direct effects of urbanism on detention, petition, adjudication, and judicial placement, as well as cross-level interactions between urbanism and several case-level factors for each outcome. We found limited support for the hypotheses. First, findings indicated that odds of detention were significantly greater in more urban courts, but indicated no differences in other outcomes. Second, findings also indicated greater extralegal differences (race, sex, and age) in more urban courts—contrary to hypotheses. Taken together, findings highlight the localized yet complex nature of juvenile justice processing and emphasize the need for additional multilevel research assessing the role of other contextual factors as potential sources of variation across macrosocial units.

Keywords

juvenile court justice by geography geographical disparities juvenile justice

It has been long observed that the administration of justice varies across different geographical contexts (see, e.g., Eisenstein et al., 1988; Myers & Talarico, 1987). In the juvenile justice system, such variation may be especially likely given the emphasis placed upon individualized treatment of delinquent youth. As Sampson and Laub (1993) note, “a fundamental fact is that the juvenile court is organized at the local (i.e., county) level, giving rise to potentially important community-level variations in juvenile justice” (p. 287). The implication, that similar cases may be processed differently according to jurisdictional context, has been referred to as “justice by geography” (Feld, 1991; Krisberg et al., 1984).

Almost three decades ago, Feld (1991) observed that “despite statutes and rules of statewide applicability, juvenile justice administration varies consistently with urban, suburban, and rural social structure and context” (p. 156). Despite this important observation, few studies have rigorously assessed how juvenile justice processing differs in urban and rural settings. More often, urbanism has been treated as a control variable in research assessing other contextual hypotheses, such as minority threat (Zane, 2018) or economic threat (Rodriguez, 2010). Additionally, while some studies have investigated how racial/ethnic and sex differences in processing differ across urban and rural courts (e.g., Bray et al., 2005; Rodriguez, 2008), fewer studies have investigated the differential influence of legal factors across court types. Following Feld (1991), however, we should expect that urban and rural courts differ substantially in terms of their degree of emphasis on due process and formal rationality, and in turn, how legal and extralegal variables influence decision-making in different courts.

The present study seeks to contribute to the “justice by geography” literature by assessing how juvenile justice processing differs across urban and rural juvenile courts located within one state. Specifically, we first examine whether urban juvenile courts are more punitive than rural courts with respect to four major stages of juvenile court processing. Second, we assess whether the effects of key case-level variables—race/ethnicity, sex, age, prior record, offense severity, and preadjudication detention—vary across urban and rural courts.

Background

Weber’s (1954) work on law, legal procedure, and the link between urbanization and bureaucratization provides one framework for thinking about how the administration of justice may vary across social contexts. Weber evaluated legal systems on the basis of two dimensions: the level of formality associated with a legal system and its application of the law, and the degree to which the system’s application of law is rational (Savelsberg, 2006). A legal system’s formality has been defined as its adherence to “strict procedures and evidence rules, often coupled with the relative autonomy of legal institutions or legal personnel,” whereas rationality “refers to a legal system’s systematic, logical derivation of abstract legal propositions that legal experts can reconstruct and consistently affirm over time, rendering the system in general calculable and predictable” (Shamir, 2007, p. 2). In turn, Weber distinguished formally rational legal systems from substantively rational legal systems. Formally rational systems are bureaucratically organized with special emphasis on the uniform application of formal rules of legal procedures in order to produce consistent outcomes. Substantively rational systems, by contrast, are less formal, more autonomous, and not as bureaucratically organized. Instead, these systems are characterized by increased discretion and reliance on informal rules or norms, often based in custom and tradition. As a result, similar cases may receive different outcomes when processed in substantively rational systems.

Drawing on this framework, early studies sought to assess whether court decision-making varied according to levels of bureaucratization (see, e.g., Dixon, 1995; Hagan, 1977; Tepperman, 1973). Much of this research compared the influence of legal, extralegal, and case processing variables on court outcomes in urban versus rural courts, where differences were attributed “to the rationalization and bureaucratic administration of justice found in urban courts as opposed to the lack of bureaucratic administration in rural courts” (Dixon, 1995, pp. 1164–1165). The assumption was that urban criminal courts were more formally rational than rural courts, and decision-making in these courts would be largely guided by legally relevant factors such as offense characteristics and prior record. By contrast, rural courts were viewed as less formal and bureaucratized, reflecting Weber’s notion of a substantive rational system. Here, decision-making would depend on legal as well as non-legal and informal considerations such as social status (see Bridges et al., 1987; Dixon, 1995). A key distinction, then, was that outcomes would be more predictable in urban courts in terms of known legal variables (in contrast to other, more difficult-to-measure factors).

Empirical findings from early studies testing these propositions were largely mixed. For instance, Tepperman (1973) found that sex differences were more pronounced in rural than in urban juvenile courts, while Tittle and Curran (1988) found race differences in counties with average levels of urbanism, but not in low or high urban courts. Focusing on criminal court outcomes, Hagan (1977) found that non-White offenders received more severe sentences than White offenders, and these differences were significantly greater in rural courts. Austin (1981) examined sentencing outcomes for criminal defendants in Iowa and found that older and non-White offenders were more likely to receive a prison sentence in rural courts. On the other hand, several studies found greater racial disparities in urban courts (Bridges et al., 1987; Kempf & Austin, 1986; Myers & Talarico, 1986a) and one study found larger sex differences in more urban criminal courts (Myers & Talarico, 1986a). Still, other early studies found no differences across urban and rural courts (Britt, 2000; Myers & Talarico, 1986b).

Justice by Geography in Juvenile Courts

One application of Weber’s (1954) framework draws upon the unique history of the juvenile court (see Engen et al., 2002). Feld (1991) observed that the U.S. Supreme Court’s landmark decision in In re Gault (1967) extended a number of due process guarantees to juvenile defendants and thus required juvenile courts to adopt a more formal approach akin to the criminal justice system. One unintended consequence of these expanded rights, however, was a conflict with the juvenile court’s original treatment-oriented mission. As Feld (2003) puts it, this due process revolution “shifted the focus of delinquency proceedings from a social welfare inquiry into a quasi-criminal prosecution” (pp. 774–775). In essence, this shift reflected an evolution from a substantively rational system, guided by the traditional principles of parens patriae, to a more formally rational system guided by principles of due process. Feld (1991) argued that there was reason to expect this shift to occur more prominently in urban courts. Specifically, higher population density, ethnic heterogeneity, and crime rates in urban communities weakened social cohesion and uniformity, leading to a greater reliance on formal social control as well as procedural formality in those courts (Feld, 2003). Conversely, the traditional approach of the juvenile court as a substantively rational system might remain viable in more rural settings, where population homogeneity, residential stability, and social cohesion might allow for greater reliance on informal social control.

This is what Feld (1991) refers to as “justice by geography” (p. 157; see also Krisberg et al., 1984). According to the hypothesis, case processing outcomes will vary between urban and rural juvenile courts in two major ways. First, Feld (1991) predicted that as a result of becoming more formal and routinized, urban courts will be more punitive than rural courts. Rural courts will have greater room for leniency and discretion as a result of being more informal and traditional in their orientation (i.e., substantive rationality). Second, it was predicted that case-processing decisions in urban courts would be largely guided by legal, offense-related considerations, while rural courts would retain the traditional welfare-oriented emphasis of the juvenile court. As such, rural court decision-making is expected to involve greater individualized consideration of each case, including non-legal factors.

To date, only a small handful of studies have tested Feld’s (1991) “justice by geography” hypothesis. Much of this research focused on racial/ethnic differences in urban versus rural courts, and findings have been mixed. In one study, Kempf-Leonard and Sontheimer (1995) found that Black youth faced greater odds of formal intake processing and petition in urban courts, and greater odds of detention in suburban (but not urban) courts than in rural courts. DeJong and Jackson (1998), on the other hand, found that Black youth were more likely to receive secure placement in rural courts, but no significant race differences in referral outcomes. Other tests of the “justice by geography” thesis similarly reached mixed results (e.g., Blackmon et al., 2015; Maupin & Bond-Maupin, 1999; Shook & Goodkind, 2009; Taylor et al., 2012). One limitation of this research, however, is that studies did not employ multilevel models to test the effects of urbanism on case processing outcomes.

More recently, however, studies using more sophisticated, multilevel modeling techniques have examined the “justice by geography” thesis and reached similarly mixed conclusions. While some recent studies have found greater likelihood of detention, formal petition, adjudication of delinquency, and secure placement in urban courts (e.g., Lowery et al., 2018; Peck & Jennings, 2016; Rodriguez, 2010), more research has not identified any association between county-level urbanism and juvenile court outcomes (Armstrong & Rodriguez, 2005; Bray et al., 2005; Freiburger & Jordan, 2011; Leiber & Peck, 2019; Peck et al., 2019; Lowery & Burrow, 2019; Zane et al., 2020). Additionally, some findings provide only partial support. For example, Rodriguez (2008) found that youth processed in urban counties were more likely to be detained and petitioned, but were also more likely to be informally processed at intake.

A less examined aspect of the “justice by geography” hypothesis is whether the influence of legal and extralegal factors varies across rural and urban settings. Most prior research on this question has focused on the effects of race on case processing in urban versus rural courts—again with mixed findings. While some earlier studies found that Black youth received harsher dispositions in rural courts (DeJong & Jackson, 1998; Kempf et al., 1990), more recent studies have found that minority youth are treated more harshly in urban (Rodriguez, 2008) as well as in suburban courts (Shook & Goodkind, 2009; Taylor et al., 2012). Other studies have not identified any racial or ethnic differences in processing outcomes in rural versus urban courts (Bray et al., 2005; Freiburger & Jordan, 2011). Contrary to Feld (1991), Rodriguez (2008) found that some other extralegal factors were only associated with juvenile court outcomes in urban court. Specifically, males were more likely to be formally petitioned in urban courts only, and older youth were more likely to be detained in urban courts only.¹

In sum, there is a relatively small body of research testing Feld’s (1991) “justice by geography” hypothesis in the context of juvenile justice, with mixed findings to date. Prior research is limited in several ways. First, the majority of prior research has treated urbanism as a control variable rather than assessing how urbanism influences juvenile justice processing, either directly or by moderating the influence of case-level factors. That is, there have been few direct tests of the “justice by geography” perspective (see, e.g., Bray et al., 2005; Shook & Goodkind, 2009). Second, much of this work does not employ multilevel modeling techniques, which are necessary for testing contextual hypotheses such as “justice by geography” (see, however, Bray et al., 2005). Several statistical problems emerge when traditional regression models are employed in the context of multilevel data. Most notably, statistical dependency within clusters produces correlated residual errors such that standard errors will be underestimated. Traditional regression models also employ the incorrect degrees of freedom for level-2 predictors such as the main explanatory variable of interest—urbanism. Additionally, testing interactions between urbanism and case-level factors requires a multilevel strategy since “the single-level regression model assumes de facto that individual predictors exert the same effect in each aggregate grouping” (Johnson, 2010, p. 623). In short, contextual hypotheses cannot be adequately tested without multilevel methods.

A third issue with prior research is that the few multilevel studies that have investigated interactions between case-level factors and urbanism have focused on extralegal factors (e.g., race/ethnicity) but have not examined the differential influence of legal factors (e.g., prior record) across types of court.² Yet it stands to reason that the influence of legal factors on juvenile court outcomes provides one of the best tests of formal versus substantive rationality across courts. Finally, prior studies have not properly tested interactions in the context of nonlinear variables (see Ai & Norton, 2003; Brambor et al., 2006). The current editors of the American Sociological Review have recently advised as follows: “The case is closed: don’t use the coefficient of the interaction term to draw conclusions about the statistical interaction in categorical models such as logit, probit, Poisson, and so on” (Mustillo et al., 2018, p. 1282). Instead, when examining the interaction between a categorical (e.g., race/ethnicity) and continuous variable (e.g., urbanism), the interaction effect is best represented by the “second differences” (i.e., second derivatives) of the marginal effects of the categorical variable across different values of the continuous variable (Long & Mustillo, 2018; Mize, 2019). The present study seeks to contribute to the empirical literature on “justice by geography” by addressing these limitations in prior work.

Current Study

The “justice by geography” hypothesis posits that, following In re Gault (1967), juvenile courts have diverged in how they process cases, leading to differences in overall punitiveness as well as what factors influence final outcomes. Specifically, it has been argued that while rural courts largely retained the traditional approach of the original juvenile court, urban courts shifted to a greater emphasis on due process and formal rationality, more akin to criminal court (Feld, 1991). On the one hand, this suggests that the more legalistic orientation is less constrained by the original juvenile court’s rehabilitative mission and is, by extension, more punitive. On the other hand, embracing a more formally rational orientation should also reduce individualized discretion and produce fewer extralegal disparities (e.g., race). Instead, legal factors will be more predictive of outcomes in “criminalized” urban juvenile courts, reflecting their increased reliance on just deserts philosophy and “the principle of the offense” (see Feld, 1993). As a result, in urban jurisdictions there might be a more equitable but also more punitive court, while in rural jurisdictions there may be a less punitive but also less equitable one (see Zimring, 2014).

The current study tests these hypotheses. First, it is predicted that urban courts will be more punitive than rural courts with respect to preadjudication detention, formal petition, adjudication of delinquency, and judicial disposition. We hypothesize as follows:

H_1A: Odds of detention will be significantly higher in more urban courts.

H_1B: Odds of petition will be significantly higher in more urban courts.

H_1C: Odds of adjudication will be significantly higher in more urban courts.

H_1D: Odds of placement will be significantly higher in more urban courts.

Second, it is expected that the influence of legal factors will be greater in urban courts, while the influence of extralegal factors will be greater in rural courts. We predict that the influence of three extralegal factors—race/ethnicity, sex, and age—will be negatively moderated by urbanism, such that effects are greater in rural courts.³ We hypothesize as follows:

H_2A: The association between detention and extralegal factors—race/ethnicity, sex, and age—will be negatively moderated by urbanism.

H_2B: The association between petition and extralegal factors—race/ethnicity, sex, and age—will be negatively moderated by urbanism.

H_2C: The association between adjudication and extralegal factors—race/ethnicity, sex, and age—will be negatively moderated by urbanism.

H_2D: The association between placement and extralegal factors—race/ethnicity, sex, and age—will be negatively moderated by urbanism.

Third, we predict that the influence of three legal factors—prior record, offense severity, and detention—will be greater in urban courts. We hypothesize as follows:

H_3A: The association between detention and legal factors—prior record and offense severity—will be positively moderated by urbanism.

H_3B: The association between petition and legal factors—prior record, offense severity, and detention—will be positively moderated by urbanism.

H_3C: The association between adjudication and legal factors—prior record, offense severity, and detention —will be positively moderated by urbanism.

H_3D: The association between placement and legal factors—prior record, offense severity, and detention—will be positively moderated by urbanism.

Method

The current study draws on case-level and county-level data to test the multilevel hypotheses outlined above. Case-level information on all juvenile court dispositions in the state of Florida in 2017 was provided by the National Juvenile Court Data Archive (NJCDA), a research organization maintained by the National Center for Juvenile Justice and supported by a grant from the Office of Juvenile Justice and Delinquency Prevention.⁴ This was merged with county-level demographic and socio-economic data from the 2010 decennial Census Summary Files and the American Community Survey (ACS) 5-year estimates for 2017, as well as county-level juvenile arrest rates for 2017 from the Uniform Crime Reports (UCR). Cases with missing data for case-level variables were dropped from the analyses (n = 2,273) as were referrals for which the race/ethnicity category was “other” (n = 250). Additionally, we dropped counties that represented extreme outliers in terms of urbanism. For the full detention sample, this resulted in an additional 4,258 dropped cases.⁵ (Sensitivity analyses retained the dropped counties.) All counties included in the analyses have a minimum of five observations. The final sample size consisted of 60,068 referrals across 66 counties.

Measures

Dependent variables

The current study examines four stages of the juvenile court process: preadjudication detention, petition, adjudication, and judicial disposition. All outcomes were coded dichotomously to facilitate interpretation. The first outcome, preadjudication detention, was coded as “1” for youth referrals detained prior to adjudication and “0” for those released while awaiting petition or adjudicatory hearing. Approximately 26% of referrals (n = 15,577) were detained. The second outcome of interest, petition, was coded as “1” for referrals formally processed (i.e., petition of delinquency) and “0” for those informally processed (i.e., dismissal at petition or diversion). Fifty-four percentage of referrals were formally processed (n = 34,974). Third, adjudication was coded as “1” for cases receiving an adjudication of delinquency or adjudication withheld and “0” for cases dismissed at adjudication. The available sample size at this stage of court processing included 31,122 petitioned referrals across 65 counties. Approximately 82% (n = 25,450) of petitioned referrals were adjudicated delinquent (or withheld). Finally, judicial disposition was coded as “1” for placement in a residential facility (i.e., commitment) and “0” for community supervision (i.e., probation). The available sample size at this stage of processing consisted of 25,046 adjudicated referrals across 65 counties, and approximately 25% of these referrals resulted in placement in a residential facility (n = 6,181). Table 1 provides descriptive statistics for all variables included in the analyses.

Table 1.

Descriptive Statistics.

Variables	n (%)
Outcomes
Detention ^a	15,577 (25.93)
Petition ^b	34,974 (54.37)
Adjudication ^c	25,450 (81.77)
Secure placement ^d	56,181 (24.68)
Case-level
White	20,026 (33.34)
Black	30,751 (51.19)
Hispanic	9,291 (15.47)
Male	45,820 (76.28)
Age, mean (SD)	15.48 (1.57)
Violent offense	14,654 (24.40)
Property offense	18,914 (31.49)
Drug/alcohol offense	5,795 (9.65)
Weapon offense	1,161 (1.93)
Probation violation	14,551(24.22)
Public order	2,260 (3.76)
Obstruction of justice	406 (0.68)
Violation of local ordinances	1,789 (2.98)
Other offense	538 (0.90)
Felony	21,681 (36.09)
Prior referrals	28,507 (47.46)
Prior adjudications	3,904 (6.50)
County-level
Urbanism, persons per square mile (×100), mean (SD)	2.93 (3.69)
Caseload pressure, mean (SD)	342.35 (193.23)
Juvenile referral rate (per 1,000), mean (SD)	17.02 (5.79)
Percent non-White, mean (SD)	20.81 (9.82)
Percent youth, mean (SD)	19.76 (3.14)
Concentrated disadvantage, mean (SD)	0 (1)
North region, n (%)	41 (61.12)
Central region, n (%)	12 (18.18)
South region, n (%)	13 (19.70)

^aN₁ = 60,068; N₂ = 66. ^b N₁ = 64,326; N₂ = 67. ^c N₁ = 31,122; N₂ = 65. ^d N₁ = 25,046; N₂ = 65.

Case-level variables

The key independent variables of interest are race and ethnicity, sex, age, prior record, offense severity, and detention. The race/ethnicity of the defendant was measured as a factor variable comprised of three categories: Black, Hispanic, and White (reference category). The full sample of juvenile court referrals was approximately 33% White (n = 20,026), 51% Black (n = 30,751), and 15% Hispanic (n = 9,291) defendants. Defendant sex was measured as a binary variable, coded “1” for male defendants who made up approximately 76% of referrals (n = 45,820). Age at time of referral was measured as a continuous variable ranging from 10 to 20 years of age, with a mean (SD) age of 15.48 (1.57). (Although the upper age of jurisdiction in Florida is 17, the extended age of jurisdiction is age 20.) In the analysis testing cross-level interactions, age was recoded as a factor variable to compare differences between “children” (ages 10–13), “true juveniles” (ages 14–15), and “young adults” (ages 16–20) (Mears et al., 2014, p. 174). Defendants ages 14–15 serve as the reference category. Prior referral was measured dichotomously to differentiate between youths with any prior referrals (coded as “1”) and those with no prior referrals to juvenile court. Overall, approximately 47% of referrals included in the final sample involved youth with at least one prior referral (n = 28,507). Offense severity was measured using a dichotomous indicator for felony (coded as “1”) and non-felony (coded as “0”) offenses. Altogether, 36% of referrals included in the final sample involved a felony offense (n = 21,681). Pre-adjudication detention is also included as an independent variable for non-detention outcomes (see above).

In addition to the main independent variables of interest, the analysis also controls for type of offense and prior adjudications. Type of offense was measured using a series of dummy variables for each offense type. These include: violent offenses (24.4%), property offenses (31.5%), drug and alcohol offenses (9.7%), public order offenses (3.8%), violations of probation or aftercare (24.2%), weapon-related offenses (1.9%), obstruction of justice offenses (.7%), violations of local or municipal laws (3%), and other offenses (.9%). Property offense serves as the reference group. Finally, prior adjudication was measured as youths with an adjudication of delinquency or adjudication withheld (coded as “1”) from 2007 to 2016. Approximately 7% of referrals had at least one prior adjudication (n = 3,904).⁶

Contextual variables

The main contextual variable of interest in the current study is county-level urbanism. Here, we measured urbanism as a continuous variable of county population density (recoded as 100 persons per square mile). Unlike other contextual variables in the analysis, this was based on 2010 data as this was the last decennial Census prior to 2017. Across counties included in the analyses, levels of population density ranged from .1 (i.e., 10 persons per square mile) to 14.45 (i.e., 1,445 persons per square mile). The sample mean (SD) was 2.93 (3.69), which corresponds to an average of 293 persons per square mile.

Additionally, several contextual variables were also included as control variables. First, we control for caseload pressure, which was measured as the total number of referrals in a circuit divided by the number of judges (see Ulmer & Johnson, 2017).⁷ Across the sample, this ranged from 37.7 to 949.8 referrals per judge, with a mean average (SD) of 342.4 (193.23). Second, we control for referral rate with a measure that captures the number of referrals in a county divided by the number of youth (i.e., number of referrals per 1,000 youth in county). This ranged from 6.5 to 34 referrals per 1,000 youth, with a mean average (SD) of 17.02 (5.79). Third, we control for county-level percent non-White population, which was measured as a continuous variable. Across the entire sample, percent non-White ranged from 6.9% to 58.6%, with a mean average (SD) of 20.81 (9.82). Fourth, youth population density was measured as the percentage of the county population that was under the age of 18. This ranged from 7.4% to 27.5%, with a mean average (SD) of 19.76 (3.14). Fifth, we control for concentrated disadvantage at the county level. We used principal component analysis to create a weighted index score of concentrated disadvantage based on percent of households below poverty, percent of female-headed households, percent of families receiving assistance via the Supplementary Nutrition Assistance Program, and the median household income (reverse coded). Principal components analysis suggested a one-factor solution with all factor loadings above .42 and a first eigenvalue of 3.03 (all others below .67); approximately 76% of the variance among the items was accounted for by the first factor. Prior to standardization, concentrated disadvantage scores ranged from −3.53 (indicating low disadvantage) to 4.08 (representing high disadvantage), with a mean value of 0 (SD = 1.75). Lastly, we control for region using a categorical variable to compare differences between juvenile courts in the North, Central, and South regions. Central region serves as the reference category in the analyses.

Analytic Strategy

To examine the effects of case-level predictors on juvenile court outcomes across urban and rural counties, a multilevel modeling approach that accounts for the nested nature of the data was necessary (see Raudenbush & Byrk, 2002). The final sample size varied based on the dependent variable, with fewer cases as the referral travels further into the court system from the detention decision at intake (i.e., total referrals) to judicial disposition (i.e., delinquent juveniles). This non-random reduction in sample size raises the possibility of endogenous selection bias (see Elwert & Winship, 2014). Some research employs the Heckman two-step procedure to account for potential selection effects when looking at sequential criminal justice outcomes, but this requires valid exclusion restrictions to be included in the model (Bushway et al., 2007). Here, no exclusion criteria could be identified. In this context, an alternative strategy is simply to examine the data generating process and inspect variables of interest for glaring selection problems (see generally Stolzenberg & Relles, 1997).⁸ Moreover, interpretations of estimates are made relative to the samples utilized at each stage of process.

Given the binary coding of each dependent variable, we employed multilevel logistic regression models in the subsequent analyses using random effects analysis. The first stage of the analyses involved estimating unconditional (i.e., null) variance component models to determine whether each of our outcomes varied significantly across counties in the sample. Second, random intercept models were estimated to examine the direct effects of case-level and contextual predictors on each outcome (H₁). For ease of interpretation, non-binary control variables were grand-mean centered. Third, we estimated random coefficients models with cross-level interactions between county-level urbanism and case-level predictors (the latter specified as random effects), testing whether these case-level influences varied across urban and rural courts (H₂ and H₃). With nonlinear dependent variables, the product term in regression output does not represent a valid test of the interaction, however (Mustillo et al., 2018). As such, marginal effects were estimated to assess differences in the probability of each outcome by race/ethnicity, sex, age, prior record, offense severity, and detention at different values of urbanism. To test interactions, we estimated second differences in the marginal effects of case-level factors across levels of urbanism. Given the large number of second differences estimated to test the cross-level interactions, we follow Mize (2019) in using figures of predicted probabilities to display the findings. (Full output available from authors upon request.) All analyses used two-sided hypothesis testing with an α of .05 and were performed using Stata 16.

Results

To determine whether juvenile court outcomes varied across counties in the sample, we began by estimating an unconditional multilevel regression model for each outcome of interest. The analyses (not shown here) suggest that the odds of detention, petition, adjudication, and placement vary significantly across counties. First, for detention, the level-2 variance component (ψ) of .16 and corresponding intraclass correlation (ρ) of .05 indicated that before any predictors were included in the model, approximately 5% of the variation in detention outcomes was attributable to differences between counties. Second, for petition, the level-2 variance component (ψ) of .29 and intraclass correlation (ρ) of .08 indicated that before including any predictors in the model, approximately 8% of the variation in petition outcomes was due to differences between counties. Third, for adjudication, the level-2 variance component (ψ) of .37 and intraclass correlation (ρ) of .10 indicated that before including any predictors in the model, approximately 10% of the variation in adjudication outcomes was attributable to differences between counties. Finally, for placement, the level-2 variance component (ψ) of .14 and intraclass correlation (ρ) of .04 indicated that before including any predictors in the model, 4% of the variation in placement outcomes was attributable to differences between counties. A multilevel modeling approach is thus warranted for all outcomes of interest.

Direct Effects

Next, we assessed the direct relationship between juvenile court outcomes and urbanism (i.e., H₁). Table 2 provides the results from the random intercept models regressing detention, petition, adjudication, and placement on urbanism.

Table 2.

Random Intercept Models.

Variables	Detention^a	Petition^b	Adjudication^c	Placement^d
Fixed effects	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Intercept	.02*** (.01, .03)	.41*** (.28, .61)	1.99*** (1.24, 3.20)	.03*** (.02, .05)
Case-level
Black^e	1.37*** (1.30, 1.44)	1.26*** (1.21, 1.32)	1.01 (.94, 1.08)	1.49*** (1.38, 1.62)
Hispanic^e	1.14*** (1.07, 1.23)	1.10** (1.04, 1.16)	1.13* (1.01, 1.25)	1.16* (1.04, 1.31)
Male	1.61*** (1.52, 1.70)	1.17*** (1.12, 1.22)	1.26*** (1.17, 1.35)	1.93*** (1.75, 2.12)
Age 10–13^f	.63*** (.58, .67)	.78*** (.74, .83)	.53*** (.48, .59)	.77*** (.69, .87)
Age 16–20^f	1.04^† (1.00, 1.09)	1.05* (1.01, 1.09)	.81*** (.76, .87)	.96 (.89, 1.03)
Violent offense^g	2.35 *** (2.23, 2.49)	.68*** (.65, .71)	.88** (.81, .97)	.97 (.88, 1.06)
Drug/alcohol offense^g	.85*** (.77, .92)	1.02 (.96, 1.09)	.83** (.74, .94)	.64*** (.55, .75)
Weapon offense^g	6.08*** (5.31, 6.96)	1.17* (1.03, 1.33)	1.38* (1.08, 1.75)	.94 (.77, 1.16)
Probation violation^g	.55*** (.51, .59)	1.36*** (1.29, 1.44)	.92 (.84, 1.01)	2.49*** (2.22, 2.79)
Public order offense^g	.97 (.85, 1.12)	.40*** (.36, .44)	.74** (.61, .91)	.87 (.65, 1.17)
Obstruction offenses^g	1.24* (1.00, 1.55)	1.02 (.83, 1.25)	1.20 (.81, 1.76)	.75 (.54, 1.06)
Violations^g	2.55*** (2.27, 2.87)	1.00 (.92, 1.13)	.73*** (.61, .87)	1.18 (.95, 1.46)
Other offenses^g	1.13 (.92, 1.39)	.76** (.64, .91)	.70* (.51, .95)	1.37^† (.89, 1.79)
Felony	3.82*** (3.63, 4.03)	1.74*** (1.67, 1.82)	1.43*** (1.32, 1.56)	1.98*** (1.80, 2.18)
Prior referrals	4.52*** (4.31, 4.73)	2.14*** (2.06, 2.22)	1.57*** (1.48, 1.68)	2.32*** (2.15, 2.51)
Prior adjudications	.70*** (.65, .76)	.93* (.86, 1.00)	.83** (.76, .96)	1.31*** (1.17, 1.46)
Detention	–	2.58*** (2.47, 2.70)	1.10*** (1.02, 1.18)	4.58*** (4.25, 4.95)
County-level
Urbanism	1.03* (1.00, 1.07)	.99 (.97, 1.02)	.99 (.96, 1.02)	.98 (.94, 1.02)
Caseload pressure	1.00^† (1.00, 1.01)	1.00 (1.00, 1.01)	1.00 (1.00, 1.01)	1.00* (1.00, 1.01)
Referral rate	1.02^† (.99, 1.03)	1.02^† (1.00, 1.04)	1.03* (1.00, 1.06)	.99 (.98, 1.02)
Percent non-White	.99 (.98, 1.01)	1.02* (1.00, 1.04)	1.02* (1.00, 1.04)	.98* (.97, 1.00)
Percent youth	1.01 (.97, 1.04)	1.07** (1.02, 1.12)	1.02 (.97, 1.08)	1.04 (.99, 1.09)
Concentrated disadvantage	.99 (.87, 1.12)	.85* (.73, 1.00)	1.05 (.86, 1.28)	1.06 (.90, 1.26)
North region^h	1.43* (1.02, 2.00)	1.41 (.90, 2.21)	1.58^† (.93, 2.68)	.95 (.60, 1.48)
South region^h	1.83** (1.30, 2.59)	1.04 (.65, 1.66)	2.16** (1.24, 3.20)	.98 (.62, 1.56)
Random effects
Random intercept variance component	.10	.21	.27	.17
Intraclass correlation	.03	.06	.08	.05

Notes. ^a N₁ = 60,068; N₂ = 66. ^b N₁ = 64,326; N₂ = 67. ^c N₁ = 31,122; N₂ = 65. ^d N₁ = 25,046; N₂ = 65. ^e Reference category: White. ^f Reference category: Age 14–15. ^g Reference category: Property offense. ^h Reference category: Central region. † p < .10 * p < .05 ** p < .01 *** p < .001 (two-tailed).

First, findings indicate that the odds of being detained prior to adjudication were 3% higher in urban counties than in rural counties (OR = 1.03). Additionally, several case-level variables of interest were associated with detention. Among legal factors, odds of detention were higher for juveniles with prior referrals (OR = 4.52) and felony offenses (OR = 3.82). Among extralegal factors, odds of detention were higher for Black (OR = 1.37) and Hispanic (OR = 1.14) referrals than for White youth. Additionally, odds of detention were higher for male referrals (OR = 1.61) and lower for younger juveniles than true juvenile offenders (OR = .63).

Second, odds of petition were not significantly higher in more urban counties (OR = .99, p = .69). Several case-level variables of interest were associated with petition, however. Among legal factors, odds of petition were higher for juveniles with prior referrals (OR = 2.14) and felony offenses (OR = 1.74), as well as those who received preadjudication detention (OR = 2.58). Among extralegal factors, odds of petition were higher for Black (OR = 1.27) and Hispanic (OR = 1.10) referrals than for White youth. Additionally, odds of petition were higher for male referrals (OR = 1.17), while odds of petition were lower among younger juveniles (OR = .78) and higher among young adults (OR = 1.05) compared to true juvenile offenders.

Third, odds of adjudication were not significantly higher in more urban counties (OR = .99, p = .53). Several case-level variables were associated with adjudication, however. Among legal factors, odds of adjudication were higher for defendants with prior referrals (OR = 1.58), felony offenses (OR = 1.43), and preadjudication detention (OR = 1.10). Among extralegal factors, odds of adjudication were higher among Hispanic defendants than White defendants (OR = 1.13), with no significant differences between Black and White youth. Odds of adjudication were also higher among male defendants (OR = 1.26). Additionally, odds of adjudication were much lower among younger juvenile defendants (OR = .53), as well as among young adult defendants (OR = .81), compared to true juvenile offenders.

Fourth, odds of placement were not significantly higher in more urban counties (OR = 1.00, p = .85). Several case-level variables of interest were associated with placement, however. Among legal factors, odds of placement were higher among defendants with prior referrals (OR = 2.32), felony offenses (OR = 1.98), and preadjudication detention (OR = 4.58). Among extralegal factors, odds of placement were higher among Black (OR = 1.49), and Hispanic (OR = 1.16) defendants than White defendants. Odds of placement were also higher among male defendants (OR = 1.93), and lower among younger juvenile offenders (OR = .77) compared to true juvenile offenders.

Interaction Effects

In addition to testing the direct relationship between urbanism and likelihood of more punitive juvenile court outcomes above (i.e., H₁), we were interested in whether urbanism moderated the relationship between juvenile court outcomes and legal and extralegal case-level variables (i.e., H₂ and H₃). Below, we report the findings for each outcome.⁹

Preadjudication detention

Figure 1 presents separate panels with the predicted probabilities of detention by each case-level variable of interest across levels of county urbanism. Findings indicated that urbanism moderated the relationship between probability of detention and race (panel A), age (panel C), and prior referrals (panel D).

Figure 1.

Predicted probabilities of detention by case-level variable across urbanism.

First, findings indicate that race interacted with urbanism but not in the hypothesized direction. As shown in panel A, Black-White differences in the probability of detention were significantly greater in more urban courts. Where urbanism was lowest (i.e., 0 persons per square mile), the predicted probability of detention was .23 for Black youth and .21 for White youth, a 2% difference (p < .05). The probability of detention for Black youth rose steadily as urbanism increased. The probability of detention also increased for White youth, but at a slower rate. At the highest levels of urbanism (i.e., 1500 persons per square mile), the predicted probability of detention was .33 for Black youth and .24 for White youth, a 9% difference (p < .001). The difference between these differences (i.e., second difference)—7%—was statistically significant (p < .001), indicating that urbanism positively moderated the (positive) association between race and detention. The difference in the probability of detention for White and Hispanic youth, which increased from 2 to 4% across values of urbanism, was not significant.

Second, findings indicated that age interacted with urbanism, but not in the hypothesized direction. As shown in panel C, true juveniles were more likely to be de detained than children across all values of urbanism—but this was especially pronounced in more urban courts. Where urbanism was lowest, the predicted probability of detention was .19 for children and .23 for true juveniles, a 4% difference (p < .001). The probability of detention for true juveniles rose steadily as urbanism increased, from .23 to .30. The probability of detention also increased for children, but at a slower rate. At the highest levels of urbanism, the predicted probability of detention was .30 for true juveniles and .21 for children, a 9% difference (p < .001). The second difference for low versus high urbanism—5%—was statistically significant (p < .05), indicating that urbanism positively moderated the (positive) association between age and detention. There was no significant difference, however, in the probability of detention for true juveniles compared to young adults (i.e., ages 16–20) across values of urbanism.

Third, findings indicated that prior referrals interacted with urbanism in the hypothesized direction. As shown in panel D, youth with prior referrals were more likely to be detained than youth without referrals across all values of urbanism—but this was especially pronounced in more urban courts. Where urbanism was lowest, the predicted probability of detention was .32 for youth with prior referrals and .13 for youth without any priors, a 19% difference (p < .001). The probability of detention for youth with a prior record rose steadily as urbanism increased, while the probability of detention only increased slightly for youth without a prior record. At the highest levels of urbanism, the predicted probability of detention was .44 for youth with prior referrals and .17 for youth without any priors, a 27% difference (p < .001). The second difference for low versus high urbanism—8%—was statistically significant (p < .01), indicating that urbanism positively moderated the (positive) association between prior record and detention.

The findings do not indicate that urbanism moderated the association between detention and other case-level variables, however. While the difference in the probability of detention decreased slightly across levels of urbanism for sex (panel B) and offense severity (panel E), the difference in differences was not significant in either case.¹⁰

Petition

Figure 2 presents separate panels with the predicted probabilities of petition by each case-level variable of interest across levels of county urbanism. Findings indicated that urbanism moderated the relationship between probability of petition and age (panel C) in the hypothesized direction.

Figure 2.

Predicted probabilities of petition by case-level variable across urbanism.

As shown in panel C, children–juvenile differences were more pronounced in less urban courts. Where urbanism was lowest, the probability of petition was .48 for children and .56 for true juveniles, an 8% difference (p <.001). The probability of petition for children increased steadily as urbanism increased, whereas the probability of petition for true juveniles slightly decreased. At the highest levels of urbanism, the predicted probability of petition was .55 for children and .54 for true juveniles, and the second difference for low versus high urbanism—9%—was statistically significant (p <.01), indicating that urbanism positively moderated the association between age and petition. The findings do not indicate that urbanism moderated the association between petition and other case-level variables, however. While the difference in probabilities increased slightly from low to high urbanism for ethnicity (panel A), sex (panel B), and offense severity (panel E), and decreased for prior referrals (panel D) and detention (panel F), second differences were not significant.

Adjudication

Figure 3 presents separate panels with the predicted probabilities of adjudication by each case-level variable across levels of county urbanism. Findings did not indicate that urbanism moderated the relationship between adjudication and any case-level variables of interest. While the difference in probabilities of adjudication decreased slightly from low to high urbanism for ethnicity (panel A), ages 16–20 (panel C), prior referrals (panel D), and detention (panel F), and increased for race (panel A), sex (panel B), ages 10–13 (panel C) and offense severity (panel E), second differences were not significant.¹¹

Figure 3.

Predicted probabilities of adjudication by case-level variable across urbanism.

Judicial placement

Figure 4 presents separate panels with the predicted probabilities of placement by each case-level variable across levels of county urbanism. Findings indicated that urbanism moderated the relationship between probability of secure placement and defendant sex (panel B) and age (panel C).

Figure 4.

Predicted probabilities of disposition by case-level variable across urbanism.

First, findings indicated that sex interacted with urbanism but not in the hypothesized direction. As shown in panel B, sex differences in the predicted probability of secure placement were significantly greater in urban rather than rural courts. Where urbanism was lowest, the predicted probability of placement was .28 for males and .24 for females, a 4% difference (p < .01). At the highest levels of urbanism, the predicted probability of placement was .24 for males and .06 for females, an 18% difference (p < .001). The second difference for low versus high urbanism—14%—was statistically significant (p < .001), suggesting that urbanism positively moderated the (positive) association between sex and placement.

Second, findings indicated that age interacted with urbanism but not in the hypothesized direction. As shown in panel C, young adults were less likely than true juveniles to receive secure placement in more rural courts, but significantly more likely to receive secure placement in more urban courts. Where urbanism was lowest, the predicted probability of placement was .25 for young adults and .28 for true juveniles, a 3% difference (p < .05). At the highest levels of urbanism, the predicted probability of placement was .33 for young adults and .27 for true juveniles, a 6% difference (p < .10). The second difference for low versus high urbanism—8%—was statistically significant (p < .05), indicating that urbanism positively moderated the (negative) association between age and placement. There was no significant difference, however, in the probability of placement for true juveniles compared to children (i.e., ages 10–13) across values of urbanism.

The findings do not indicate that the association between placement and other case-level variables was moderated by urbanism, however. While the difference in probabilities of placement increased from low to high urbanism for race/ethnicity (panel A) and offense severity (panel E), and decreased for prior referrals (panel D) and detention (panel F), second differences were not significant.

Discussion

Guided by Feld’s (1991) “justice by geography” thesis, the current study assessed whether juvenile court outcomes, as well as the influence of key legal and extralegal factors on those outcomes, varied across urban and rural counties. First, we hypothesized that the odds of detention, formal petition, adjudication of delinquency, and secure placement would be higher in urban than in rural courts (H₁). Second, we hypothesized that the association between juvenile court outcomes and three extralegal factors—race/ethnicity, sex, and age—would be negatively moderated by county urbanism (H₂). Third, we hypothesized that the association between juvenile court outcomes and major legal factors—prior record, offense severity, and preadjudication detention—would be positively moderated by county urbanism (H₃). Our findings indicate limited support for the research hypotheses.

Direct Effects of Urbanism (H₁)

First, findings provide limited support for the hypothesis that urban courts are more punitive than rural courts across juvenile court outcomes. On the one hand, findings indicate that the odds of preadjudication detention are significantly higher in more urban counties (H_1A). This finding lends some support to Feld’s (1991) contention that the “formal and due process-oriented” nature of urban courts “is associated with greater severity in pre-trial detention” (p. 162). On the other hand, we found no significant differences in the odds of petition, adjudication, or secure placement across counties according to level of urbanism (H_1B–H_1D). This indicates that geographical context (i.e., urbanism) may be more influential at some stages of juvenile court processing than others. An organizational perspective suggests that some of these stages may be more “loosely” or “tightly” coupled based on the relationships among various system actors. Hagan (1989) first adapted this terminology from organizational theorists and applied it to the criminal justice system: “In connotative terms, loose coupling is meant to evoke the image of entities (e.g., court systems) that are responsive to one another, while still maintaining independent identities and some evidence of physical or logical separateness” (p. 119). One of the key takeaways from this coupling framework is that tightly coupled systems will be more responsive to explicit organizational norms, while loosely coupled systems may be more willing to circumvent of such norms via “localized adaptation” (Ulmer, 2019, p. 488).

Here, detention arguably represents the most “loosely coupled” stage of processing under investigation: the detention decision is the result of input about dangerousness and public safety from the police, intake officer, prosecutor, and judge—with substantial variation in roles across jurisdictions (Barton, 2012). It may be the case that urban counties are more likely to detain referrals as a result of more formal intake processing at this loosely coupled stage of processing, as predicted by “justice by geography” (Feld, 1991). By contrast, it may be that decisions at more tightly coupled processing stages (petition, adjudication, and disposition) are already quite formal and due-process oriented—regardless of urbanism (see Bishop et al., 2010). This possibility remains largely unexamined, as prior assessments of “justice by geography” have typically been limited to only one (e.g., Bray et al., 2005; Lowery et al., 2018) or two case-processing outcomes (e.g., Rodriquez, 2008, 2010).

Interactive Effects of Urbanism and Extralegal Variables (H₂)

Second, findings provide limited support for the second set of hypotheses (H_2A–H_2D). It was hypothesized that the influence of defendant race/ethnicity, sex, and age on court outcomes would be more pronounced in less urban (i.e., rural) courts where, due to more informal processing, extralegal considerations may play a larger role. For defendant race/ethnicity, the findings indicate that Black and Hispanic youth were significantly more likely to be detained, petitioned, and committed to residential placement than White youth (while Hispanic youth were also more likely to be adjudicated delinquent). Contrary to hypotheses, however, Black-White differences in odds of detention were greater in more urban courts, while Hispanic-White differences in outcomes were not moderated by urbanism. For defendant sex, the findings indicate that male referrals were significantly more likely to be detained, petitioned, adjudicated, and committed. This relationship was only moderated by urbanism for one outcome, however, with the sex difference in odds of placement increasing in more urban courts. Moreover, as with race/ethnicity, this is contrary to the research hypotheses. For defendant age, compared to “true juveniles” (i.e., age 14–15), the findings indicate that the youngest juvenile offenders (i.e., age 10–13) were less likely to be detained, petitioned, adjudicated delinquent, and committed to residential placement, while the oldest juvenile offenders (i.e., age 16–20) were more likely to be petitioned but less likely to be adjudicated delinquent. This relationship was moderated by urbanism for several outcomes. For detention and secure placement, age differences were greater in more urban courts—contrary to hypotheses. For petition of delinquency, however, age differences were smaller in more urban courts—consistent with the research hypotheses. In sum, most of the interactions between extralegal factors and urbanism were in the unexpected direction—namely, that extralegal differences were more pronounced in more urban courts (see also Rodriguez, 2008).

One possible explanation for larger extralegal differences in urban courts may be that these courts operate in an organizational context in which informal relations among courtroom workgroups are prevalent, leading to increased discretion (Eisenstein et al., 1988). As Dixon (1995) argues, “work groups in highly bureaucratic and loosely coupled urban contexts characterized by low interaction, visibility, and accountability are more likely to produce informal legal cultures that use discretion in decision-making” (p. 1115). In turn, sentencing practices in these courts “are more individually determined, more discretionary, and more likely to invoke the use of legally irrelevant criteria” (p. 1116). Alternatively, it may be that courtroom officials in urban courts operate with limited and insufficient information regarding an offender’s rehabilitation potential or risk of reoffending (see, e.g., Albonetti, 1986, 1987; Ulmer, 2012). In this context, extralegal considerations such as defendant race, sex, and age may act as signals of culpability, dangerousness, or amenability to treatment (Bridges & Steen, 1998; Engen et al., 2002; Myers & Talarico, 1987).¹² Echoing a similar sentiment, Bishop and Leiber (2012, p. 462) observe that juvenile court decisions, especially in early processing, “often must be made within a matter of hours based on little information” such that “officials may rely on typifications—shorthand cues based on race and class stereotypes,” to which we could add age and gender stereotypes as well (see also Albonetti, 1991). There is a good deal of research on the relationship between organizational factors such as caseload pressures and criminal justice sanctions (e.g., Johnson, 2005, 2006; Johnson et al., 2008; Ulmer, 2019; Ulmer & Johnson, 2017), and future research could extend this to the moderating impact of organizational context on the influence of extralegal factors. This would stand in contrast to "justice by geography," which assumes that formality limits discretion and hence discrimination. Instead, it may be that formalism is associated with a greater need to rely on stereotypical attributions due to case processing pressures.

Interactive Effects of Urbanism and Legal Variables (H₃)

Third, findings provide limited support for the third set of hypotheses (H_3A–H_3D). It was hypothesized that the influence of legal factors would be more pronounced in urban courts where “criminalization” of the juvenile court has occurred and court processing is more formal and due-process oriented. For offense severity, the findings indicate that felony referrals were significantly more likely to be detained, petitioned, adjudicated, and committed to residential placement than non-felony referrals—but none of these associations were moderated by urbanism. Similarly, for preadjudication detention, the findings indicate that detained referrals were significantly more likely to be petitioned, adjudicated, and committed than non-detained referrals—but none of these associations were moderated by urbanism.

For prior record, the findings also indicate that across counties, youth with prior referrals were significantly more likely to be detained, petitioned, adjudicated delinquent, and committed than youth without prior referrals. This relationship was moderated by urbanism for one outcome, preadjudication detention. Specifically, findings indicated that the greater odds of detention for youth with prior referrals was more pronounced in more urban courts. This offers some support for Feld’s (1991) suggestion that the formal and bureaucratized nature associated with urban courts would lead to punishment decisions based on legal factors such as prior record. Still, the moderating effect of urbanism appears quite limited.

Limitations

Several study limitations warrant discussion. First, our analyses were limited to one state and, as such, the findings are not generalizable to all juvenile justice systems. Future assessments of “justice by geography” using data from multiple states will provide a framework for understanding the conditions under which urbanism is more or less likely to influence variation in juvenile court outcomes. The organizational context of different courts is especially important. Relatedly, a second limitation concerns the operationalization of procedural formality using county-level population density. The operationalization goes beyond the traditional urban-rural dichotomy prevalent in much prior research, yet county-level measures may still fail to capture the organizational richness of different courts (see Ulmer, 2019; Ulmer & Johnson, 2017). Future research is needed, including primary data collection, on the organizational nature of courts, including court communities and their local legal culture as well as characteristics of courtroom workgroup members (see, e.g., Arazan et al., 2019; Eisenstein et al., 1988; Farrell et al., 2009; Hester, 2017; Haynes et al., 2010; Metcalfe, 2016).

A third limitation is omitted variable bias. Despite controlling for important correlates of key case-level predictors and the outcomes analyzed, there are still many unmeasured characteristics that may be associated with outcomes as well as with variation in the effects of key predictors across different juvenile justice decision points. For example, the current study found that both Black and Hispanic youth were significantly more likely to be detained, formally petitioned, and committed to residential placement than their White counterparts. While such disparities may be the product of “focal concerns” of justice system personnel which lead to racially biased attributions (Harris, 2009), they may also be the product of other unobserved case characteristics, such as developmental maturity, substance abuse, family situation, attitudinal factors, victim characteristics, and strength of evidence (see, e.g., van Wingerden et al., 2016). This challenge confronts all research inquiries into disparities whose causes involve perceptual processes and decision-making that cannot be easily observed (Ulmer, 2012).

Conclusions

In light of our study findings, how should we continue to think about “justice by geography” in juvenile justice processing? First, our findings suggest that the juvenile courts examined in the present study are less representative of Weber’s notion of formal rationality as is predicted by Feld’s (1991) thesis. That is, it was not the case that legal variables were predictive of court outcomes but extralegal variables were not—irrespective of whether courts were urban or rural. The influence of legal as well as extralegal variables on court outcomes suggests that these systems are guided by other, less formal norms—more akin to what are called substantively rational systems. It may be that future research must focus more on these informal, more-difficult-to-measure aspects of juvenile court processing to explain differences in outcomes across different courts. However, given that we did not account for other background variables (e.g., family structure, school status), we stop short of concluding that the juvenile courts under examination were fully reflective of substantively rational systems. As mentioned earlier, it may be that the juvenile courts in our study instead are more reflective of Albonetti’s (1991) concept of bounded rationality—similar to theories of criminal justice decision-making under conditions of uncertainty.

Second, and even more importantly, the motivating concern of the present research—that juvenile justice outcomes might vary across geographical units—was supported, although the specific hypotheses about urban-rural differences largely were not. While such variation across counties (e.g., Rodriguez, 2008) as well as states (e.g., Zane et al., 2020) is well established, extant research has not sufficiently explained this variation. Just as “justice by geography” has received limited empirical support, so too have other leading contextual hypotheses such as minority threat (see Zane, 2018). Perhaps research on court-level variation must move beyond standard macro-level demographic variables that are typically examined in the literature, such as urbanism (or, in the case of minority threat, racial composition).

For example, it may be that theories that focus more on organizational characteristics of the juvenile court, its political economy as well as its internal culture, will better explain variation in juvenile punishment across counties. As Arazan and colleagues (2019) recently observed, “noticeably absent from the previous literature is discussion of accurately modeling courtroom organization theories that clearly discuss the court being embedded within the community and reflective of community norms and values” (p. 42). While the present research attempted to capture organizational differences among courts in terms of urbanism and case processing pressures, it may be that there are other organizational factors that require further examination, such as a focus on courtroom workgroups (e.g., Metcalfe, 2016) or court culture, often embedded in the larger political culture (see generally, Nelken, 2010). Lynch (2019, p. 20) argues for “re-inhabiting criminal sentencing research,” mainly through primary data collection aiming to capture the richness of courtroom workgroups in a more “bottom up approach” that examines how discretion operates in a dynamic context with social, legal, and political influences (see also Ulmer, 2019).

It is clear that future research must continue to pay attention to the “social structure and context” of juvenile court decision-making (Feld, 1991, p. 156), despite the failure, thus far, to explain adequately why juvenile court outcomes and disparities in processing vary across different contexts. To this end, our results more broadly illustrate the localized yet complex nature of the juvenile justice system and underscore the fact that no single, uniform juvenile justice system exists (see Zane et al., 2020). It is clear that justice varies by geography and future research should seek to investigate the role of other factors as potential sources of this variation. In the end, such efforts will contribute to the refinement of current perspectives on contextual variation in juvenile court processing and, in turn, will enhance our understanding of the broader context in which juvenile justice decision-making occurs.

Footnotes

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

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Jhon A. Pupo

Steven N. Zane

Notes

References

Aguinis

Gottfredson

R. K.

Joo

(2013). Best-practice recommendations for defining, identifying, and handling outliers. Organizational Research Methods, 16(2), 270–301.

Norton

E. C.

(2003). Interaction terms in logit and probit models. Economics Letters, 80(1), 123–129.

Albonetti

(1986). Criminality, prosecutorial screening, and uncertainty: Toward a theory of discretionary decision making in felony case processing. Criminology, 24(4), 623–644.

Albonetti

(1987). Prosecutorial discretion: The effects of uncertainty. Law & Society Review, 21(2), 291–313.

Albonetti

(1991). An integration of theories to explain judicial discretion. Social Problems, 38(2), 247–266.

Arazan

C. L.

Bales

W. D.

Blomberg

T. G.

(2019). Courtroom context and sentencing. American Journal of Criminal Justice, 44(23), 23–44.

Armstrong

G. S.

Rodriguez

(2005). Effects of individual and contextual characteristics on preadjudication detention of juvenile delinquents. Justice Quarterly, 22(4), 521–539.

Austin

T. L.

(1981). The influence of court location on type of criminal sentence: The rural-urban factor. Journal of Criminal Justice, 9(4), 305–316.

Barton

W. H.

(2012). Detention. In Feld

B. C.

Bishop

D. M.

(Eds.), The Oxford handbook of juvenile crime and juvenile justice (pp. 636–663). Oxford University Press.

10.

Bishop

D. M.

Leiber

M. J.

(2012). Racial and ethnic differences in delinquency and justice system responses. In Feld

B. C.

Bishop

D. M.

(Eds.), The Oxford handbook of juvenile crime and juvenile justice (pp. 445–484). Oxford University Press.

11.

Bishop

D. M.

Leiber

M. J.

Johnson

(2010). Contexts of decision making in the juvenile justice system: An organizational approach to understanding minority overrepresentation. Youth Violence and Juvenile Justice, 8(3), 213–233.

12.

Blackmon

Cain

D. S.

Livermore

(2015). Juvenile court dispositions in the deep south: Examining the concept of justice by geography. Journal of Social Sciences, 11(2), 82–90.

13.

Brambor

Clark

W. R.

Golder

(2006). Understanding interaction models: Improving empirical estimates. Political Analysis, 14(1), 63–82.

14.

Bray

T. M.

Sample

L. L.

Kempf-Leonard

(2005). Justice by geography: Racial disparity and juvenile courts. In Hawkins

D. F.

Kempf-Leonard

(Eds.), Our children, their children: Confronting racial and ethnic differences in American juvenile justice (pp. 270–290). University of Chicago Press.

15.

Bridges

G. S.

Crutchfield

R. D.

Simpson

E. E.

(1987). Crime, social structure and criminal punishment: White and non-white rates of imprisonment. Social Problems, 34(4), 345–361.

16.

Bridges

G. S.

Steen

(1998). Racial disparities in official assessments of juvenile offenders: Attributional stereotypes as mediating mechanism. American Sociological Review, 63(4), 554–580.

17.

Britt

C. L.

(2000). Social context and racial disparities in punishment decisions. Justice Quarterly, 17(4), 707–732.

18.

Bushway

S. D.

Johnson

B. D.

Slocum

L. 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.

19.

DeJong

Jackson

K. C.

(1998). Putting race into context: Race, juvenile justice processing, and urbanization. Justice Quarterly, 15(3), 487–504.

20.

Dixon

(1995). The organizational context of criminal sentencing. American Journal of Sociology, 100(5), 1157–1198.

21.

Eisenstein

Flemming

R. B.

Nardulli

P. F.

(1988). The contours of justice: Communities and their courts. Little Brown & Co.

22.

Elwert

Winship

(2014). Endogenous selection bas: The problem of conditioning on a collider variable. Annual Review of Sociology, 40, 31–53.

23.

Engen

R. L.

Steen

Bridges

G. S.

(2002). Racial disparities in the punishment of youth: A theoretical and empirical assessment of the literature. Social Problems, 49(2), 194–220.

24.

Farrell

Ward

Rousseau

(2009). Race effects of representation among federal court workers: Does black workforce representation reduce sentencing disparities? The ANNALS of the American Academy of Political and Social Science, 623(1), 121–133.

25.

Feld

B. C.

(1991). Justice by geography: Urban, suburban, and rural variations in juvenile justice administration. Journal of Criminal Law & Criminology, 82(1), 156–210.

26.

Feld

B. C.

(1993). Criminalizing the American juvenile court. Crime and Justice, 17, 197–280.

27.

Feld

B. C.

(2003). The politics of race and juvenile justice: The “due process revolution” and the conservative reaction. Justice Quarterly, 20(4), 765–800.

28.

Freiburger

T. L.

Jordan

K. L.

(2011). A multilevel analysis of race on the decision to petition a case in the juvenile court. Race and Justice, 1(2), 185–201.

29.

Hagan

(1977). Criminal justice in rural and urban communities: A study of the bureaucratization of justice. Social Forces, 55(3), 597–612.

30.

Hagan

(1989). Why is there so little criminal justice theory? Neglected macro- and micro-level links between organization and power. Journal of Research in Crime and Delinquency, 26(2), 116–135.

31.

Harris

(2009). Attributions and institutional processing: How focal concerns guide decision-making in the juvenile court. Race and Social Problems, 1(4), 243–256.

32.

Haynes

S. H.

Ruback

Cusick

G. R.

(2010). Courtroom workgroups and sentencing: The effects of similarity, proximity, and stability. Crime & Delinquency, 56(1), 126–161.

33.

Hester

(2017). Judicial rotation as centripetal force: Sentencing in the court communities of South Carolina. Criminology, 55(1), 205–235.

34.

In re Gault, 387 U.S. 1 (1967).

35.

Johnson

B. D.

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

36.

Johnson

B. D.

(2006). The multilevel context of criminal sentencing: Integrating judge- and county-level influences. Criminology, 44(2), 259–298.

37.

Johnson

B. D.

(2010). Multilevel analysis in the study of crime and justice. In Piquero

A. R.

Weisburd

(Eds.), Handbook of quantitative criminology (pp. 615–648). Springer.

38.

Johnson

B. D.

Ulmer

J. T.

Kramer

J. H.

(2008). The social context of guidelines circumvention: The case of federal district courts. Criminology, 46(3), 737–783.

39.

Kempf

K. L.

Austin

R. L.

(1986). Older and more recent evidence on racial discrimination in sentencing. Journal of Quantitative Criminology, 2(1), 29–48.

40.

Kempf

K. L.

Decker

S. H.

Bing

R. L.

(1990). An analysis of apparent disparities in the handling of Black youth within Missouri’s juvenile justice system. University of Missouri–St. Louis.

41.

Kempf-Leonard

Sontheimer

(1995). The role of race in juvenile justice in Pennsylvania. In Kempf-Leonard

Pope

C. E.

Feyerherm

W. H.

(Eds.), Minorities in juvenile justice (pp. 98–127). Sage.

42.

Krisberg

Litsky

Schwartz

(1984). Youth in confinement: Justice by geography. Journal of Research in Crime and Delinquency, 21(2), 153–181.

43.

Leiber

M. J.

Peck

J. H.

(2019). Clarifying the theoretical tenets of the symbolic threat perspective. Justice Quarterly, 37(6), 1040–1066.

44.

Long

J. S.

Mustillo

S. A.

(2018). Using predictions and marginal effects to compare groups in regression models for binary outcomes. Sociological Methods & Research. Advanced online publication. https://doi.org/10.1177/0049124118799374

45.

Lowery

P. G.

Burrow

J. D.

(2019). Concentrated disadvantage, racial disparities, and juvenile institutionalization within the context of attribution theory. Criminal Justice Studies, 32(4), 330–355.

46.

Lowery

P. G.

Burrow

J. D.

Kaminski

R. J.

(2018). A multilevel test of the racial threat hypothesis in one state’s juvenile court. Crime & Delinquency, 64(1), 53–87.

47.

(2018). Rural and urban differences in gender-sentencing patterns of Pennsylvania. Rural sociology, 83(2), 402–430.

48.

Lynch

(2019). Focally concerned about focal concerns: A conceptual and methodological critique of sentencing disparities research. Justice Quarterly, 36(7), 1148–1175.

49.

Maupin

J. R.

Bond-Maupin

L. S.

(1999). Detention decision-making in a predominantly Hispanic region: Rural and non-rural differences. Juvenile and Family Court Journal, 50(3), 11–23.

50.

Mears

D. P.

Cochran

J. C.

Stults

B. J.

Greenman

S. J.

Bhati

A. S.

Greenwald

M. A.

(2014). The “true” juvenile offender: Age effects and juvenile court sentencing. Criminology, 52(2), 169–194.

51.

Metcalfe

(2016). The role of courtroom workgroups in felony case dispositions: An analysis of workgroup familiarity and similarly. Law & Society Review, 50(3), 637–673.

52.

Mize

T. D.

(2019). Best practices for estimating interpreting, and presenting nonlinear interaction effects. Sociological Science, 6, 81–117.

53.

Mustillo

S. A.

Lizardo

O. A.

McVeigh

R. M.

(2018). Editors comment: A few guidelines for quantitative submissions. American Sociological Review, 83(6), 1281–1283.

54.

Myers

M. A.

Talarico

S. M.

(1986a). Urban justice, rural injustice? Urbanization and its effect on sentencing. Criminology, 24(2), 367–391.

55.

Myers

M. A.

Talarico

S. M.

(1986b). The social contexts of racial discrimination in sentencing. Social Problems, 33(3), 236–251.

56.

Myers

M. A.

Talarico

S. M.

(1987). The Social contexts of criminal sentencing. Springer.

57.

Nelken

(2010). Comparative criminal justice: Making sense of difference. Sage.

58.

Peck

J. H.

Jennings

W. G.

(2016). A critical examination of “being Black” in the juvenile justice system. Law and Human Behavior, 40(3), 219–232.

59.

Peck

J. H.

Leiber

M. J.

Beaudry-Cyr

(2019). Expanding the applicability of Sampson and Laub’s theory of inequality and social control: A multilevel examination. Criminal Justice and Behavior, 46(6), 902–919.

60.

Raudenbush

S. W.

Byrk

A. S.

(2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Sage.

61.

Rodriguez

(2008). A multilevel analysis of juvenile court processes: The importance of community characteristics. National Institute of Justice.

62.

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(3), 391–413.

63.

Sampson

R. J.

Laub

J. H.

(1993). Structural variations in juvenile court processing: Inequality, the underclass, and social control. Law & Society Review, 27(2), 285–311.

64.

Savelsberg

J. J.

(2006). Sociological theory in the study of sentencing: Lighthouse for a traveler between contents. In Deflem

(Ed.), Sociological theory and criminological research: Views from Europe and the United States (pp. 183–202). Elsevier.

65.

Shamir

(2007). Weber, Max (1864-1920). In Clark

D. S.

(Ed.), Encyclopedia of law & society: American and global perspectives (pp. 1574–1577). Sage.

66.

Shook

J. J.

Goodkind

S. A.

(2009). Racial disproportionality in juvenile transfer: The interaction of race and geography in pretrial detention for violent and serious offenses. Race and Social Problems, 1(4), 257–266.

67.

Stolzenberg

R. M.

Relles

D. A.

(1997). Tools for intuition about sample selection bias and its correction. American Sociological Review, 62(3), 494–507.

68.

Taylor

Guevara

Boyd

L. M.

Brown

R. A.

(2012). Race, geography, and juvenile justice: An exploration of the liberation hypothesis. Race and Justice, 2(2), 114–137.

69.

Tepperman

(1973). The effect of court size on organization and procedure. Canadian Review of Sociology, 19(4), 346–365.

70.

Tittle

C. R.

Curran

D. A.

(1988). Contingencies for dispositional disparities in juvenile justice. Social Forces, 67(1), 23–58.

71.

Ulmer

J. T.

(2012). Recent developments and new directions in sentencing research. Justice Quarterly, 29(1), 1–40.

72.

Ulmer

J. T.

(2019). Criminal courts as inhabited institutions: Making sense of difference and similarity in sentencing. Crime and Justice, 48, 483–522.

73.

Ulmer

J. T.

Johnson

B. D.

(2017). Organizational conformity and punishment: Federal court communities and judge-initiated guidelines departures. Journal of Criminal Law and Criminology, 107(2), 253–292.

74.

van Wingerden

Wilsem

Johnson

B. D.

(2016). Offender’s personal circumstances and punishment: Toward a more refined model for the explanation of sentencing disparities. Justice Quarterly, 33(1), 100–133.

75.

Weber

(1954). Max Weber on law in economy and society ( Rhenstein

, Trans). Harvard University Press.

76.

Zane

S. N.

(2018). Exploring the minority threat hypothesis for juveniles in criminal court: Static versus dynamic threat and diffuse versus targeted effects. Youth Violence and Juvenile Justice, 16(4), 418–441.

77.

Zane

S. N.

Mears

D. P.

Welsh

B. C.

(2020). How universal is disproportionate minority contact? An examination of racial and ethnic disparities in juvenile justice processing across four states. Justice Quarterly, 37(5), 817–841.

78.

Zane

S. N.

Singer

S. I.

Welsh

B. C.

(2021). The right to a good defense: Investigating the influence of attorney type across urban counties for juveniles in criminal court. Criminal Justice Policy Review, 32(2), 162–192.

79.

Zimring

F. E.

(2014). Minority overrepresentation: On causes and partial cures. In Zimring

F. E.

Tannenhaus

D. S.

(Eds.), Choosing the future for American juvenile justice (pp. 169–186). New York University Press.

Assessing Variations in Juvenile Court Processing in Urban Versus Rural Courts: Revisiting “Justice by Geography”

Abstract

Keywords

Background

Justice by Geography in Juvenile Courts

Current Study

Method

Measures

Dependent variables

Case-level variables

Contextual variables

Analytic Strategy

Results

Direct Effects

Interaction Effects

Preadjudication detention

Petition

Adjudication

Judicial placement

Discussion

Direct Effects of Urbanism (H1)

Interactive Effects of Urbanism and Extralegal Variables (H2)

Interactive Effects of Urbanism and Legal Variables (H3)

Limitations

Conclusions

Footnotes

Declaration of Conflicting Interests

Funding

ORCID iDs

Notes

References

Direct Effects of Urbanism (H₁)

Interactive Effects of Urbanism and Extralegal Variables (H₂)

Interactive Effects of Urbanism and Legal Variables (H₃)