Can Racial Diversity among Judges Affect Sentencing Outcomes?

How does racial diversity impact institutional outcomes and (in)equality? Discussions about diversity usually focus on how individuals’ identities shape their behavior, but diversity is a group-level characteristic. Scholars must, therefore, consider the relationship between group composition and the individual decisions that shape institutional outcomes. Using felony data from a large U.S. court system, I explore the relationship between racial diversity among the judges comprising a court and individual judges’ decisions. I find that as the percent of Black judges in a courthouse increases white judges are less likely to render incarceration sentences in cases with Black defendants. Increases in racial diversity decrease the Black–white gap in the probability of incarceration by up to 7 percentage points. However, I find no relationship between judge’s racial identities and disparities in their decisions. This study highlights the importance of conceptualizing diversity as a group characteristic and the relationship between institutional context and outcomes.


Introduction
Most residents of the United States will never meet their congressional or state representatives, but many will have contact with public officials in the criminal justice system at some point.The decisions these public officials make directly affect those individuals' lives-they may mean the difference between a night at home and a night in jail or the difference between freedom and imprisonment.The environment in which these public officials work, from the building to employee policies to their colleagues, must influence how they approach their jobs, which, in turn, affects the lives of the countless individuals' they interact with each day.This paper examines how the composition of a judge's group of colleagues can influence her sentencing decisions in criminal cases and how this influence differs based on a judge's own personal characteristics.
Millions of people are arrested in the U.S. every year, and the vast majority of these individuals are charged and tried in state, rather than federal, courts.1For example, the Illinois circuit courts 2 hears almost as many felony cases in one year as all of the federal district courts combined (Illinois Courts 2013; U.S. Federal Courts 2013).State trial court judges' sentencing decisions have far-reaching political consequences for the lives of those accused (Weaver and Lerman 2010;Lerman and Weaver 2014;White N.d.b) and also for their families and communities (Burch 2013;Walker 2014;Walker and García-Castañon 2017;White N.d.a).
The majority of defendants are people of color; and people of color (specifically Blacks and Latinos) are more frequently incarcerated than Whites in the U.S. (Carson 2018).The citizens who appear most frequently in the largest state criminal courts are those for whom the criminal justice system is likely their primary, if not only, contact with government.Citizens develop their perceptions about government and decide whether and how to participate based on these interactions (Lerman and Weaver 2014).
...exposure to criminal justice institutions has implications for how individual citizens understand the political world they inhabit-how they see and seek to make claims on government, how they understand their political standing, and whether they perceive their group as a valued political equal in an ongoing scheme of cooperation (Lerman and Weaver 2014, p. 9).I argue that increasing the representation of Black judges on the bench increases fairness in felony sentencing, not because of Black judges' behavior, but because their presence alters their peers' behavior.These shifts in individual judges' behavior increase the court's ability, as a whole, to provide equitable sentences to defendants across racial groups.The representative bureaucracy literature demonstrates that the level of descriptive or "passive" representation within the government workforce influences how policies are implemented and administered, especially as those policies relate to women and people of color (including : Meier 1993;Meier, Wrinkle and Polinard 1999;Riccucci and Meyers 2004;Riccucci and Van Ryzin 2017;Meier 2018).As is the case in business settings, this relationship may depend on representation of the minority group reaching a "critical mass" within the bureaucracy (Kanter 1977;Meier 1993;Atkins and Wilkins 2013).And the resulting changes in the organization's performance may not be due solely to the behavior of minority-group member bureaucrats; majority-group member bureaucrats may change their behavior as well (Lim 2006;Hong 2017).
A significant literature in judicial politics examines how judges' decisions differ with their racial identities.However, we know relatively little about how changes in de-scriptive representation on the bench affect sentencing outcomes for defendants in the nation's busiest courts or about how judges' racial identities affect their colleagues' behavior.Previous work has examined the affect of group composition on appellate judges' decisions (including: Boyd, Epstein and Martin 2010;Kastellec 2013), andSchanzenbach (2005) found that the racial composition of a federal district could influence overall criminal sentencing within the district.This paper makes the novel contribution that group composition can influence solitary trial judges-both White and Black-as well.
I find that, as they gain Black colleagues, Black and White judges become less likely to incarcerate Black defendants, and White judges become more likely to incarcerate White defendants, shrinking the Black-White incarceration gap by up to eleven percentage-points.I suggest potential mechanisms for these different responses.For Black judges, having Black colleagues may alleviate pressures associated with tokenism,3 while White judges might face pressure not to appear discriminatory as they gain Black colleagues.However, only judges who work in close proximity with their Black colleagues respond to shifts in racial diversity; and both Black and White judges are less responsive to racial diversity in election years, when they may be consumed with more immediate electoral pressures.
The findings in this paper have extremely important policy implications.First, they suggest that increasing racial diversity on the bench may help decrease racial disparities in sentencing.Second, since Black defendants appear in criminal court and are sentenced to incarceration more frequently than White defendants, increasing racial diversity among judges may lead to lower incarceration rates overall.Finally, the findings highlight the potential impact of policies geared toward improving racial diversity among judges.In the sections that follow, I provide a theoretical grounding for this paper, in-troduce the state court system and data, present the analyses, discuss the findings, and conclude.

Theoretical Grounding: Racial Diversity among Judges
Most scholarship investigating the implications of racial diversity within the judiciary has focused on the extent to which an individual judge's racial identity influences her behavior.These studies are typically interested in how non-White judges differ from White judges.When it comes to Black judges, specifically, scholars have shown that they rule more "liberally" than White judges in cases related to race, including affirmative action, Voting Rights Act, and racial discrimination cases (Including: Sen 2017;Boyd 2016;Kastellec 2013;Cox and Miles 2008a).There is less consensus about the relationship between judicial characteristics and case outcomes in criminal courts (Grossman et al. 2016;Abrams, Bertrand and Mullainathan 2012;Spohn 2009).In older studies, Uhlman (1978) finds little difference in Black and White judges' rulings, but Welch, Combs and Gruhl (1988) find that Black judges are more equitable in their decisions to incarcerate Black and White defendants than their White Counterparts.Spohn (1990), on the other hand, finds that Black judges' rulings are somewhat harsher than White judges'.However, differences in behavior between non-White and White judges is only one way that shifts in racial diversity among judges might affect case outcomes.Studies of three-judge Courts of Appeals panels show that one judge's characteristics can affect how the other two judges vote and, ultimately, the panel's final decision.These studies find that adding a Black judge to a panel with two non-Black members, increases the likelihood that the panel will decide in favor of affirmative action and the Voting Rights Act, because Black judges vote differently than non-Black judges, and because non-Black judges' alter their votes when they are deciding on panels with Black judges (Cox and Miles 2008a;b;Kastellec 2013). 4The panel effects studies offer important contributions to our understanding of the relationship between racial diversity and judging, but there are relatively few Black appellate court judges, so these studies do not evaluate how Black judges' behavior (in addition to White judges') might change based on the racial identities of the judges with whom they serve on panels.
Importantly, there is some literature that addresses the relationship between racial diversity and judicial decision making in federal trial courts with a focus on criminal sentencing (Collins et al. 2010;Schanzenbach 2005).Schanzenbach (2005) measures the proportion of Black, women, and Democratic-appointed judges in each of the federal judicial districts to assess the impact of race on sentencing.He interacts this measure with a measure of defendants' race to identify the relationship between the level of racial diversity among a district's judges and sentencing based on a defendant's race.He finds that as Black judicial representation in a district increases, there is not much change in how Black defendants are sentenced (Schanzenbach 2005).
With a focus on gender rather than racial diversity among judges, Collins et al. (2010) usefully incorporate critical mass theory into their analyses of gender and decision making.The authors argue that the extent to which women judges will behave differently than men is a function of how many other women they count as colleagues, which measures their ability to escape the pressures of tokenism (Kanter 1977).These authors study outcomes in different types of district court cases, including criminal cases; they model the ideological direction of a judicial decision based on whether the judge is a woman and the number of women working in her city.Federal judicial districts sometimes have more than one physical location.Judges who work in the same city within their district are more likely to interact with one another than with judges in the same district who work in different locations.Collins et al. (2010) note that these opportunities for interaction can play an important role in critical mass theory, and incorporate this improvement over district-level measures of judicial characteristics into their models.The authors find that as women's representation in a district's location increases, women are less likely to decide against defendants in criminal cases (Collins et al. 2010).
Drawing from the concepts of social pressure, tokenism, and critical mass theory, I argue that proximity to one's colleagues allows for racial diversity among those colleagues to affect an individual judge's decisions, even when that judge decides cases alone.Critical mass theory is a useful concept for thinking through the mechanism that could drive Black judges' response to gaining more Black colleagues, but it is not as helpful for evaluating White judges' response to increases in Black judicial representation.The scholarship addressing appellate court panel composition discusses some of the mechanisms that might drive the shift in non-Black judges' votes when they serve on panels with a Black colleague, including the possibility that the presence of a Black judge may alter White judges' behavior or the possibility that information shared during deliberation changes with the composition of the panel, for example (Boyd, Epstein and Martin 2010;Kastellec 2013).However, although trial court judges may discuss their caseloads with their colleagues, they do not deliberate those cases (Collins et al. 2010).
The presence mechanism is more useful, but unlike in the panel context, Black judges are not randomly present or absent in trial courthouses. 5All judges are present hearing cases in their coutrooms and interacting with one another professionally.Therefore, this study carefully considers how Black judges' presence might lead to changes' in White judges' decision making in a trial court setting.I present two potential mechanisms for explaining why White judges might respond to the level of racial diversity among their colleagues and one potential mechanism to explain why Black judges might respond below.The analyses presented later in this paper cannot definitively identify the mechanisms underlying the relationship between racial diversity among judges and individual judges' sentencing decisions.However, they do provide some support for two of the mechanisms-social pressure and the alleviation of tokenism-which are discussed in more detail below.
Contact.White judges' opinions of and behavior towards Black defendants may change as they come into contact with more Black colleagues.Scholars have found that intergroup contact can often lead to negative outcomes, including "exclusionary attitudes" and discrimination toward out-group members (Enos 2014;Forbes 1997).However, Enos (2014) finds that repeated contact decreases the magnitude of the negative outcomes associated with intergroup contact, and earlier work has identified the benefits of longer term intergroup contact for decreasing out-group prejudice under certain conditions (Allport 1979).A group of psychologists find, for example, that when intergroup contact is positive, White Americans who have positive interactions with Black Americans have more empathy toward Black Americans and are more supportive of collective action in support of Black Americans (Selvanathan et al. 2017).
Most judges have repeated contact with their colleagues, and these interactions are likely positive, at least when compared with the negative sentiments they are more likely to associate with defendant interactions.For example, in the court system from which the data for this paper come, criminal trial judges interact with other judges on their assigned judicial teams where they review team members' progress through case assign-ments and receive information from the presiding judge.Colleagues might also chat after parking in the judges' parking lot or while enjoying their morning coffee in the presiding judge's chambers before heading to their respective courtrooms (Bogira 2005, pp.33). 6  If this mechanism underlies the relationship between Black judges' representation within a court and White judges' behavior, I would expect to observe White judges becoming less punitive toward Black defendants as they gain Black colleagues, but only when they work in close enough proximity to come into contact and have positive interactions with their Black colleagues.However, there is no clear implication for their behavior toward White defendants if this mechanism underlies the relationship, so I would expect their behavior toward White defendants to remain unchanged.
Social Pressure.Institutional factors may create social pressures that lead White judges to alter their sentencing decisions.As I mentioned above, court's often track trial judges' productivity.The particular court under study in the analyses that follow, shares reports on the court's and each of its judges' progress through cases with all of the court's judges.So, each judge is aware of other judges' sentencing records and aware that her colleagues are aware of her record.7As they gain Black colleagues, White judges, knowing that their colleagues could become aware of their case records, may feel pressure to alter their sentencing behavior so as not to appear discriminatory against Black defendants and preferential toward White defendants in front of their Black colleagues.
Just as fear of one's neighbors learning about one's participation in an election can encourage the socially desirable act of voting (Gerber, Green and Larimer 2008), fear of appearing biased in front of one's colleagues may encourage sentencing decisions perceived to be more socially acceptable to a group of colleagues that has become more racially diverse.I expect that, for judges, the knowledge that they will have to see and interact with their colleagues may also be an important factor for this potential mechanism.Fear of her colleagues reviewing a discriminatory sentencing record will be much more salient to a judge if she has to see those colleagues in person on a regular basis.
A social pressure mechanism should lead White judges to render decisions that do not appear to be discriminatory and that do appear to be fair.The empirical implications of this mechanism underlying the relationship between racial diversity and White judges' sentencing behavior would be greater leniency in White judges' sentencing of Black defendants and, potentially, increased harshness in their sentencing of White defendants, again, only when they work in close proximity with their Black colleagues.
Alleviation of Tokenism.When there are low levels of Black judicial representation the Black judge(s) on a court may experience tokenism.In these circumstances, Black judges may feel pressure to conform to White judges' typical behavior, which is more punitive toward Black defendants (Asch 1956;Steffensmeier and Britt 2001).Tokenism pressures occur due to the scrutiny and visibility that come with being a member of an extreme minority group (Cook and Glass 2015;Yoder 1991;Kanter 1977).
However, the pressures of tokenism may subside as more Black judges join the bench and Blacks become less of an extreme minority group.As the pressures alleviate, Black judges may feel less scrutinized and more able to sentence Black defendants less harshly.If the alleviation of tokenism is the mechanism underlying the relationship between racial diversity on the bench and Black judges' sentencing decisions, then the expectation would be that Black judges should become less punitive toward Black defendants as they gain Black colleagues.The alleviation of tokenism does not suggest that Black judges' behavior toward non-Black defendants should change as they gain more Black colleagues, since tokenism, if it is an important factor, is unlikely to make them feel that their behavior towards non-Black defendants is scrutinized in the same way as their behavior towards Black defendants.Criminal Divisions judges, like all judges in the Circuit Court of Cook County, gain their seats on the bench in one of three ways.The first, and most common way, is through partisan elections; elected judges are referred to as Circuit Judges.Historically, these have been county-wide elections, but in 1992 the Illinois General Assembly established residential subcircuits for the election of judges. 9The legislation seems to suggest that eventually county-elected judgeships will be replaced by subcircuit-elected judgeships; however, there are currently both county-and subcircuit-elected judges serving on the Circuit Court of Cook County.Once elected, these judges have the same responsibilities, and they all serve the entire county.Circuit judges serve six-year terms and at the end of each term they run in uncontested nonpartisan retention elections.

The Circuit Court of Cook County, Criminal Division
The second route to the bench is through the associate judge process.Associate judges are appointed to the court by current Circuit judges for four-year terms.Associate Judges are similar to Circuit Judges, but they are not permitted to preside over felony cases without approval from the Illinois Supreme Court. 10Finally, the state's supreme court may appoint judges to fill vacancies between elections, but these judges must run in the next election if they wish to retain their seat.The majority of judges in Cook County are elected Circuit Judges.County's 32,000 felony defendants were found not guilty.
Each case is identified by a case number, and the following information is included for each case: initiation date; defendant's last name, date of birth, race, and sex; judge's name; statute under which the offense fell; description of charge; the class of the original charge; final disposition and date it was rendered; length of minimum and maximum sentence; courthouse and number of the courtroom where the case was heard.11I dropped all cases initiated after 2013, to allow enough time for a final disposition to be reached, and all cases initiated prior to 1995 when the court began randomly assigning cases to judges (Abrams, Bertrand and Mullainathan 2012).12When a defendant has multiple charges associated with his case, those cases share a case number.A defendant is often only convicted and sentenced for one of these charges while the others are dismissed.I drop all cases with both duplicated case numbers and dismissed charges for the analyses that follow. 13The final dataset includes just under 450,000 cases, 78 percent of which were heard in Chicago's Leighton courthouse.Figure 1 presents the distribution of all felony cases by year.0 10,000 20,000 30,000 40,000 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 8 1 9 9 9 2 0 0 0 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 2 0 1 2 2 0 1 3 The defendant in the vast majority of these cases-just under 76 percent-was Black, while White and Latino defendants were charged in 11 and 12 percent of cases, respectively (see Table 1).The data also include a wide variety of offenses.In Illinois, judges' discretion with respect to sentencing is limited within the terms established by the Illinois General Assembly's "Penalties for Crimes in Illinois", the state's sentencing guidelines. 14Judges decide whether to give a defendant a jail or prison term and, if so, for how long within the usual incarceration term for the felony class of the offense in question.M refers to murder, for which the usual prison term is anywhere from four to 60 years, and the option of probation instead of prison is not allowed (except for second degree murder).Class X is the next most serious class of felony, and it includes aggravated battery with a firearm, for example.Probation instead of incarceration is not an option for those convicted of Class X felonies, and the usual prison term is six to thirty years.
Probation instead of incarceration is allowable for all of the less serious offenses, and the usual prison and probation terms decrease as the seriousness of the charges decreases.There were a total of 74 Criminal Division judges during this period, with 38-40 judges serving in the Division each year.The number of cases per judge varies from 338 to 11,419 with an average of 3,801 cases per judge. 16Each judge has a permanent courtroom.17However, floating judges fill in for the permanent judges when permanent judges are unable to come to work and sometimes render the final dispositions in cases.
The Cook County data include the courtroom to which a case was assigned and the name of the sentencing judge.In Chicago's Leighton Building, felony cases are randomly assigned to trial courtrooms, and therefore (permanent) judges, after defendants are formally charged.Regardless of whether a floating judge happens to render the sentence in a given case one day, the permanent judge will have been present for the majority of the pre-trial and trial phases, and will have ultimate responsibility for a case.To distinguish permanent from floating judges, I identify the judge deciding cases most frequently in each courtroom.

Effects of Having Black Colleagues on Judges' Sentencing Decisions
This section includes the results of analyses of the relationship between judges' colleagues' characteristics and judges' decisions to sentence felony defendants to incarceration.The unit of analysis is a judge-case.The dependent variable is a dichotomous variable that takes a value of one if a case ended with an incarceration sentence and zero otherwise.
Figure 4 shows the incarceration rate for each year included in the data.The incarceration rate declined fairly steadily until 2011, when it began to increase again.However, the main independent variable, the proportion of Black judges in Cook County's Criminal Division has not steadily increased over time, as shown in Figure 3, suggesting that any observed response to the level of racial diversity on the bench is not merely a reflection of the passing of time and related changes to judges' beliefs that might come with it.
18 Similar analyses to those presented below measured racial diversity using the proportion of judges who were not White and produced similar results.
I measure the independent variable annually, in the first set of models presented, and monthly throughout the rest of the paper.It is likely that individual judges respond differently to changes in the representation of Black judges among their colleagues, so the models that follow are within-judge analyses.1 9 9 6 1 9 9 7 1 9 9 8 1 9 9 9 2 0 0 0 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1 2 0 1 2 2 0 1 3 The results presented below come from logistic regression models of the decision to incarcerate the defendant in a case, and they include random effects for each Criminal Division judge to account for unobserved factors affecting an individual judge's decisions, including her educational and professional background, for example, and the likelihood that her decisions across cases may be related.I use random effects models, because the measure of racial diversity-the proportion of Black judges in the Criminal Divisionvaries little within a year (and does not vary at all within a year in the first set of models presented).Also, there will be little within-unit variation for any judges who only appear in the data for a short period of time, including judges who retired shortly after 1995, for example.When assessing the effects of variables with little within-unit variation, especially when there is minimal correlation between the variable and unit effects, random effects models are often recommended (Clark and Linzer 2015). 19However, the main results are consistent across fixed effects and random effects specifications. 20 The analyses include measures for the following defendant characteristic: race, age, gender, and whether the defendant has previously been charged with a felony (within the timeframe covered by the dataset).They also account for the seriousness of the charge with indicators for each but the least serious felony class. 21Finally, I account for the length of the case with a measure of the number of years from case initiation to disposition.A case's initiation date is the date that the defendant is assigned to a trial judge's courtroom, and the disposition date is the date the case is completed, either with a plea, bench trial, or jury trial.
I include case length, because it is related to the seriousness of the offense.The court case for a low-level theft is likely resolved more quickly than an aggravated assault case, for example.The data support this, as cases for more serious offense are longer than those for less serious offenses.Case length is also one way to measure the quality of a defendant's attorney, which is especially helpful since the data do not include useful measures of this concept.Journalistic and ethnographic accounts of the goings on at the Leighton Criminal Court in Chicago confirm that defense attorneys with heavier caseloads 19 Correlation between unit effects and the main independent variable, as well as related interaction terms is below 0.1. 20The random and fixed effects are calculated for the permanent judges, and not the floating judges.Fixed effects models and models with clustered standard errors are included in the appendix. 21The models exclude murder cases, as the judges have the least discretion in sentencing in these cases and they are often not assigned to judges randomly.Removing these cases does not alter the results of the analyses; however, the observed associations are stronger, as should be expected.
and those who might be less concerned with the well-being of their clients are more likely to encourage a defendant to accept a plea agreement more quickly than other defense attorneys who often charge higher fees for their services (Bogira 2005;Van Cleve 2016).
Black defendants are more likely to be represented by a public defender or an attorney who chargers lower fees, and the data show that Black defendants' cases are shorter than White defendants' cases (Van Cleve 2016).
Results from random effects logistic regression models of judges' sentencing behavior are included in Table 4.The first column presents the results for cases heard in Chicago.The second column includes just those cases heard in the two suburban courthouses.In general, the control variables are related with incarceration in the ways we might expect.Black defendants are more likely to be incarcerated than White defendants (the reference category for defendant race), as are Latino defendants.Defendants of other races include those who were identified as Asian, Middle Eastern, and other races and ethnicities in the court's data.Older defendants and those with previous felony charges are more likely to be incarcerated than younger defendants and those without a felony history, female defendants are less likely to be incarcerated than male defendants, and White judges are more likely to render incarceration sentences than judges of other races.
The coefficients for the case length variable and its interaction with felony class are both in the expected direction based scholars' prior observations of felony cases in Chicago.As case length increases, so does the likelihood of incarceration for low-level felonies.However, when the cases for more serious felonies run longer, they are less likely to end with incarceration sentences.
The model includes the measure of racial diversity among judges-the proportion of Black judges on the bench (each year) as well as an interaction term with the indicator that a defendant is Black.The coefficient for the racial diversity measure is statistically Standard errors in parentheses * p < 0.05, * * p < 0.01, * * * p < 0.001 significant and positive, and the interaction term is statistically significant and negative for cases heard in Chicago (first column of Table 4) and also for cases heard in the suburban courthouses (second column), suggesting that increasing racial diversity among the court's judges is associated with an increase in the likelihood of incarceration for cases with White defendants, a weaker increase in the likelihood of incarceration for cases with Black defendants heard in the suburbs, and a decrease in the likelihood of incarceration for cases with Black defendants heard in Chicago.
The size of the interaction term coefficient is larger when considering just those cases heard in Chicago, which we should expect if, as I theorize, contact and interaction with one's colleagues drives this relationship.Throughout this entire period, none of the Criminal Division judges in Skokie and only one of the judges in Bridgeview were Black.
So, if the suburban judges were to respond to racial diversity in the Criminal Division, it would be with respect to colleagues based in Chicago whom they rarely, if ever, see.
Predicted probabilities measuring the substantive effect of racial diversity among judges, and how it varies with defendant race, illustrate this relationship more clearly.
Predicted probabilities of incarceration from the models in Table 4 are presented in Figure 5.Each graph shows the probability of a case-with a 25 year old male defendant who has no prior felony cases-ending with an incarceration sentence by the race of the defendant as the proportion of Black judges on the bench increases. 22Each point on the lines represents the probability of incarceration for the related level of diversity and the vertical lines at each point are 95-percent confidence intervals.For cases heard in Chicago's Leighton courthouse (left-side panel of Figure 5), the probability of incarceration decreases for cases with Black defendants as the court's judges become more racially diverse and increases at a similar rate for cases with White defendants.The relationship between racial diversity among judges and incarceration is so strong that it nearly closes the gap in incarceration between cases with Black and White defendants, decreasing it by roughly 15 percentage-points.To better illustrate this result, the left panel of Figure 6 shows the change in the gap in the probability of incarceration for cases with Black versus White defendants in Chicago as the proportion of Black judges in the Criminal Division increases.The middle line plots the difference in probability, while the lines on either side of it represent the 95-percent confidence interval.
.  or shifts in judges' attitudes about race over time.There is not a great deal of judicial turnover within the Criminal Division of the Circuit Court of Cook County.Once they receive a permanent position within the Division, most judges remain until their retirement or until they receive a position on a higher court.There is also much less turnover of Criminal Division judges in the suburban courthouses than in the Leighton Building, because there are so many fewer judges in the suburbs.However, the remainder of the paper focuses only on cases heard in Chicago, where the first set of analyses showed that judges do respond to the level of racial diversity in their group of colleagues, and uses a monthly measure of racial diversity among judges.Shifts in the proportion of Black judges on the bench within a month should capture actual changes in the court's roster of permanent judges, but it is possible that some may capture floating judges filling in for permanent judges during long absences.However, even in this case, the floating judge will likely be around long enough to become part of the professional environment within the Leighton Building.
Table 5 presents random effects logit models of the decision to incarcerate in felony cases herd in Chicago with judge-level random effects similar to those in Table 4, but the racial diversity variable is measured monthly.The model in the first column does not control for time.The coefficients for the racial diversity measure and its interaction with the indicator for a Black defendant are significant and similar to those in the first column of 4, though slightly larger.Columns 2 and 3 of Table 5 do include controls for time; Column 2 incorporates year fixed effects and Column 3 includes a year trend.The coefficient for the main racial diversity measure is much larger in Column 2 with year fixed effects than in either the model with no time measure (Column 1) or the model with a year trend measure (Column 3).This is likely because the proportion of Black judges varies one to four times most years year and never more than six times in one year.A year trend is also more appropriate than year fixed effects, in this case, since there is a general downward trend in incarceration during this period, and because individuals feelings about and responses to racial diversity in this setting are more likely to change gradually over time (or not at all) than they are to change in response to yearly shocks to the institution.In fact, there have not been many institutional changes likely to impact all judges' perceptions of or responses to their Black colleagues.For example, there were only three Criminal Division presiding judges during this period and only two Chief judges of the Circuit; and, as mentioned, the Division experiences fairly low turnover among judges.
The right panel of Figure 6 shows the substantive effects of racial diversity among judges on sentencing from Column 3 of Table 5.As the proportion of Black judges on the bench increases, the gap in the probability of a case with a Black versus White defendant ending with an incarceration sentence decreases.The substantive results derived from the model with the monthly measure and year trend are quite similar to those obtained from the model using the annual measure of racial diversity among judges (shown in the left panel of Figure 6).
To provide further evidence that the results so far are not merely the reflection of judges' changing attitudes toward race and incarceration over time, I constructed a model that attempts to control for this possibility with a year trend interacted with the indicator for a Black defendant (model not shown).Overall, the results of this model are similar to those presented so far.The coefficients for the trend variable and the interaction with defendant race are both negative, suggesting that judges are becoming less punitive over time, especially in cases with Back defendants.However, as Figure 7 shows, including these measures does not mute the relationship between the proportion of Black judges on the bench and individual judges' decisions to incarcerate.The results so far are in line with the theory that racial diversity among judges leads to more equitable treatment of defendants.As expected, this relationship is only present among judges who work in the same building; and as Black judicial representation increases, judges are more likely to render incarceration sentences in cases with White defendants and less likely to do so in cases with Black defendants.However, I also argue that Black and White judges are likely to respond differently to variations in Black judicial representation, and I consider this and other factors related to potential mechanisms underlying the relationship uncovered here in the following subsection.

Judges' Racial Identities and Responses to Racial Diversity
The potential mechanisms introduced previously indicate that Black and White judges should respond similarly to increased Black judicial representation when rendering sentences for cases with Black defendants, but that there is no reason to expect Black judges to alter their behavior in cases with White defendants.Also, only the social pressure mechanism (and not the contact mechanism) suggests that White judges should become more concerned with appearing fair or non-discriminatory in their sentencing and therefore more punitive in cases with White defendants.The fourth and fifth columns of Table 5 show the results of the models for just those cases heard by White judges in Chicago and those heard by Black judges in Chicago, respectively.
The coefficient for the proportion of Black judges on the bench is statistically significant in the model of cases heard by White judges (Column 4), but not in the model of cases heard by Black judges (Column 5), and the coefficient for the interaction term is significant in both models.As the predicted probabilities from each of these models show in Figure 8, both White and Black judges become less punitive in cases with Black defendants as racial diversity on the bench increases.However, only White judges are more likely to render incarceration sentences in cases with White defendants, behavior that is in line with social pressure as a mechanism.Black judges' decreased punitiveness toward Black defendants and the lack of change in their behavior toward White defendants provides some support for the alleviation of tokenism as the mechanism underlying this relationship. .

Discussion and Conclusion
Diversity with respect to background, ideology, gender, and race can greatly alter any environment, and the analyses presented above show that there is much more at stake when considering racial diversity among trial court judges than scholars typically acknowledge.
First, the impact of such diversity extends beyond the behavior of any one judge whose identity might make the bench more racially diverse.The results suggest that the presence of Black judges alters individual judges' behavior, making both Black and White judges less punitive in cases with Black defendants and increasing equality in sentencing outcomes for Black and White defendants.However, this only occurs when judges work in close proximity with their Black colleagues.Second, racial diversity among members of the judiciary has the potential to do more than instill a sense of institutional legitimacy among those most likely to appear in criminal courts as defendants (Scherer and Curry 2010); it may have the ability to impact the lives and freedom of individuals who interact with the institution directly.Sentencing disparities have been well documented; however, the results of the analyses presented here show that increasing the number of judges who look like the majority of defendants could reduce those defendants' likelihood of being imprisoned.Finally, these findings have important policy implications as the country continues to struggle to address disparate outcomes for people of color in all phases of the criminal justice process.This paper moves beyond previous scholarship evaluating the relationship between racial diversity among judges and judicial decision making by providing an individual The data used in this paper come from the Circuit Court of Cook County in Illinois, which is one of the largest unified court systems in the U.S. The Circuit is composed of six municipal districts, with the First Municipal District covering the entire city of Chicago.The chief judge presides over the Circuit, which includes non-judicial offices (such as Adult Probation, Jury Administration, etc.) and the judiciary.The judiciary comprises three departments: the County Department, the Juvenile Justice and Child Protection Department, and the Municipal Department.The Criminal Division is part of the County Department, and this division is responsible for hearing all of the county's felony cases.The judges of the Criminal Division hear cases in three of the county's courthouses: the George N. Leighton Criminal Court Building in Chicago, the Skokie Courthouse, and the Bridgeview Courthouse.Skokie and Bridgeview are municipalities in suburban Cook County.Felonies originating (occurring in) the city of Chicago are heard in the city's Leighton Criminal Building, and felonies originating outside of the city limits are heard by a Criminal Division judge in either Skokie or Bridgeview.Since 1995, cases have been randomly assigned to judges in Chicago's Leighton Building. 8Criminal cases are not necessarily assigned using a similar process in the suburban courthouses.The judges in Skokie and Bridgeview share these municipal courthouses with the other judges of those municipal districts who hear civil, traffic, ordinance violation, and misdemeanor cases.The Division's presiding judge and 32 trial judges have courtrooms in Chicago's Leighton Building and there are usually three to six Criminal Division trial judges with courtrooms in the Skokie and Bridgeview district courthouses.

Figure 1 :
Figure 1: Number of Felony Cases Initiated per Year in the Circuit Court of Cook County, Criminal Division

Figure 2 :
Figure 2: Incarceration Rates by Defendant Race

Figure 4 :
Figure 4: Pct.Cases with Incarceration Sentences per Year

Figure 5 :Figure 6 :
Figure 5: Probability of Incarceration by Defendant Race as Black Judicial Representation Increases

Figure 7 :
Figure 7: Change in Difference between Probability of Incarceration for Black v. White Defendants as Black Judicial Representation Increases (when controlling for time x defendant race)

Figure 9 :
Figure 9: Direct Effect of Judge Race on Sentencing

Figure 10 :
Figure 10: Change in Difference between Probability of Incarceration for Black v. White Defendants as Black Judicial Representation Increases (in retention election years)

Table 2
shows the distribution of cases by felony class from most to least serious.Class 14 See: http://www.ilga.gov/commission/lru/2005PFC.pdf for details on felony classes and sentencing guidelines in Illinois.

Table 1 :
Distribution of Cases by Defendant Race

Table 2 :
Distribution of Cases by Felony Class from Most to Least Serious See Figure2).Surprisingly, cases with Latino Defendants do not appear to receive incarceration sentences that much more frequently than those with White defendants.However, it is important to note that the distribution of cases each year by race (included in the appendix) suggests that officials began identifying Latino defendants as White and Black in the 2010s.

Table 3 :
Characteristics of Criminal Division Judges Figure 3: Proportion of Black Judges in the Criminal Division per Year are not the only judges who contribute to the court's racial diversity.However, Black judges make up the second largest racial group in the Circuit Court of Cook County'sCriminal Division, and they are an important group when considering racial disparities in sentencing.Additionally, as the table shows, there were very few Latino judges in the Criminal Division during this period.18Theproportion of Black judges on the bench each year, shown in Figure3, rose fairly steadily from 1995 through 2002 and then decreased through 2008, after which it increased and leveled off, but never reached the level it did in 2002.The sections that follow analyze whether and the extent to which this measure of racial diversity is related to individual judges' sentencing decisions in felony cases.

Table 4 :
Random Effects Logit Moldels of Racial Diversity and Sentencing

Table 5 :
Racial Diversity (Monthly) and Sentencing in Chicago * p < 0.05, * * p < 0.01, * * * p < 0.001 Diversity and Sentencing by Judge RaceAs they gain Black colleagues, pressure not to appear discriminatory may encourage White judges to sentence Black and White defendants more equitably.If, when they are members of an extreme minority group, Black judges feel increased pressure to exhibit behavior that is more in line with White judges, that pressure might lead them to sentence Black defendants harshly, even if that harshness is not in line with how they might behave otherwise.However, Black judges may feel less visible and more comfort-able showing leniency toward Black defendants as they gain more Black colleagues, but have no reason to change their behavior toward White defendants.