From Roadside Stops to the Courthouse Stairs: Defendant Race and the Attorney’s Role in Routine Cases

Abstract

Latinos, especially those who recently immigrated, face many obstacles in navigating the political and judicial environment in the United States. While prior scholarship suggests that racial minorities are more likely to be stopped by law enforcement for traffic violations and face harsher penalties for major crimes, little research has explored whether a defendant’s characteristics are influential in routine traffic court cases. Using an original database, this paper examines disparate treatment in speeding ticket reductions. The results indicate that Latino defendants are less likely to receive meaningful reductions to their charges. However, attorney representation greatly lessens the likelihood of disparate treatment for Latino drivers. As traffic court proceedings often represent the only interaction most people have with the judicial system, these findings have significant implications for racial equality, the administration of justice, attorney representation, and public opinion of the judiciary.

The disparate treatment of racial minorities in the criminal justice system is one of the most politically salient issues in the United States today. Numerous studies show that Black and Latino drivers are more likely to be stopped for routine traffic violations than White drivers (e.g., Novak 2004; Warren, Tomaskovic-Devey, Smith, Zingraff, and Mason 2006). This trend, often referred to as “Driving While Black” for Blacks (Harris 1999) or “Driving While Brown” in the case of Latino drivers (Mucchetti 2005), has received even more attention with racially tinged police shootings in Ferguson, Missouri; Baton Rouge, Louisiana; and Falcon Heights, Minnesota, among others. Studies have also chronicled the unequal treatment of minorities in sentencing decisions for some crimes (e.g., Mustard 2001; Johnson 2003; Unah 2011).

Yet, less understood is whether this disparate treatment extends to the resolution of routine court matters. While most work done in the U.S. judicial system occurs in the trial courts (Rowland and Carp 1996; LaFountain, Strickland, Holt, and Lewis 2016), scholars rarely focus their research at the trial level, and even fewer studies examine “routine” cases. Of course, there are valid reasons for focusing on higher courts, such as appellate rulings’ wide-ranging policy implications, the long-term influence of precedential cases, and the difficulties in wading through thousands of trial court decisions, if the data are even available at all. However, actions and behaviors that occur at the trial level present a plethora of interesting questions worthy of study (Barclay and Chomsky 2014). From what influences the initial criminal charge or civil complaint, to the choice of an attorney, to plea bargains and settlement negotiations, to the outcome itself – each stage of trial court litigation raises unique questions shaping the administration of justice.

However, this vast range of promising questions also raises numerous and potentially daunting challenges for investigators. Parsing out which factors, actors, or even which stage of the proceedings are the most important to study poses a limiting factor for a researcher. Further complicating matters is that much of the decision making occurs outside of the judge and jury (Schanzenbach and Tiller 2007) or before a case even enters the courtroom with police contact (Knowles, Persico, and Todd 2001) and prosecutor charging decisions (Rehavi and Starr 2014). Most cases in the criminal context result in a plea bargain, and civil cases are overwhelmingly settled between the parties without a trial (e.g., Kritzer 1991; Galanter 2004; Ostrom, Strickland, and Hannaford-Agor 2004; Durose, Farole, and Rosenmerkle 2009). This makes it even more difficult to obtain data for investigation.

Even with these obstacles, this project proceeds down the less traveled road of trial court research and explores whether disparities in treatment seen in other contexts, such as traffic stops and felony prosecutions, also occur during routine traffic ticket resolutions. As we know from Herbert Kritzer’s work on civil attorneys, there is much to learn from examining “the ordinary and modest cases” (1990, 3). In addition, going to court over a speeding ticket or smaller matters may be the only direct contact most members of the public ever have with a court system. For example, in North Carolina in 2016 there were 1,065,580 non-DWI traffic citations issued, translating to, on average, about one out of every ten people in the state receiving a ticket.¹ The aggregate effects of systematic disparate treatment based on race or other aspects could, collectively, raise major issues for the day-to-day lives of citizens and confidence in the courts (Micheslon 2003).

In examining a large sample of routine speeding tickets, we find that disparate treatment does exist, but perhaps not in the same way as in other parts of the judicial system. We find that Latino drivers receive harsher punishments and “worse” plea deals in these routine matters than those of other races, controlling for other factors such as age, gender, and legal factors. Unlike some other studies, we find that Black defendants did not receive significantly different treatment than did those of other races.² However, legal representation appears very important, particularly for Latino drivers, suggesting the importance of attorneys even in these routine cases.

We note that while the stakes might be lower than a felony case, minor cases have potential for real consequences: fines, increased insurance costs, license suspensions, and, in some cases, jail time. These direct consequences do not even scratch the surface of the downstream effects, with lost wages for court appearances, potential for detention based on unpaid fees, and the mark of a misdemeanor criminal record (Feely 1979; Micheslon 2003; Natapoff 2018). Studying case dispositions for this large segment of cases gives us greater insight into larger issues, such as systematic racial disparities and resource-related differences in treatment.

Another important factor to consider is that lower courts, especially traffic courts, also face little, if any, oversight. This is due to the strong power of prosecutors to resolve cases before factual adjudication and the general lack of review by higher courts. Prosecutors’ relatively complete discretion within many criminal justice contexts has been well documented (e.g., Gordon and Huber 2002; O’Neil 2003). However, the power for prosecutors and the role of defense attorneys in offering pleas may be even greater in resolving these minor cases. Misdemeanor cases resolve with a guilty plea in 95% of cases (Natapoff 2018).³ Without oversight, resolution of speeding tickets, for example, could display important trends that may be underlying the entire judicial system. Just as recent studies have examined routine traffic stops and have found differences in police practices based on race, age, and other factors (e.g., Pickerill, Moshier, and Pratt 2009; Rojeck, Rosenfeld, and Decker 2012; Ferrell 2015), an extensive examination of routine court cases might foster a better understanding of important trends that may otherwise be difficult to observe. Because appointed counsel are not available to all defendants in these minor cases, attorneys could also have a major impact on outcomes (Smith and Maddan 2022).

Lastly, we note that most scholarly work that does focus on trial courts tends to focus on judges’ behaviors (e.g., Kim, Schlanger, Boyd and Martin 2009; Epstein, Landes, and Poser 2013). This study, however, builds on recent scholarship (e.g., Boyd and Hoffman 2013; Barclay and Chomsky 2014; Metcalf 2016; Collins, Moyer, and Dumas 2017) exploring other court actors. As has been noted, “There is relatively little attention [by social scientists] directed at lawyers as interesting political phenomena in themselves” (Kritzer 2012, 8). This analysis adds to this emerging line of scholarship by focusing attention to the parties, attorneys, and racial inequality in the legal system.

Racial inequity within the political context

Social scientists have explored the causes and impact of racial discrimination in numerous political and social contexts, such as racial “priming” in political ads (Valentino, Hutchings, and White 2002), the influence of candidate race on electoral behavior (Matsubayashi and Ueda 2011), and the delivery of public services (Brown and Coulter 1983), to name just a few. Various theories have explored the possible causes of racial prejudice and often included both individual and institutional foundations for racially biased opinions (Borgida and Miller 2013). Some theories point to the idea of “group positioning” (Blumer 1958) in which historical contexts and the past positioning of dominant and inferior social groups dictates individual perceptions about members of those groups (Bobo and Hutchings 1996). This form of social dominance theory stems from “a more general tendency for humans to form and maintain [a] group-based hierarchy” that leads to “systematic institutional and individual discrimination” (Sidanius, Pratto, van Laar, and Levin 2004, 846–847). In effect, individuals use racial stereotypes as mental shortcuts or cues that can lead to prejudice and discrimination. These racial schemas allow individuals to form quick opinions with less effort when they encounter new members of diverse racial groups, which then influences behaviors through racial priming (White 2007).

In the context of disparate racial treatment, it is important to recognize that different minority groups face very different challenges. From a theoretical perspective, Blacks in the United States have a long history of slavery and racial discrimination. Both internally and externally, this group possesses a sense of “linked fate” based on their common experiences (Dawson 1995), leading to increased group consciousness and identification within the group (McClain et al. 2009).

This shared experience is, at least in part, due to the group’s perception as a racial threat (Behrens, Uggen, and Manza 2003). The basic premise is that scarcity creates competition among groups for power, economic benefits, and other resources. Racial minority groups are perceived as a threat by the racial majority, which provokes more hostile actions against the minority group (King and Wheelock 2007). For example, in his seminal work on the South, V.O. Key (1949) suggested that in areas of high Black concentrations, Whites strengthened efforts to prevent Black electoral participation due to the heightened threat. In other examples, racial threat theory has been used as a basis, for the loss of White voters for the Democratic Party in the South (Giles and Hertz 1994) and as an influence on the level of punishments for minority defendants (King and Wheelock 2007).

While Latinos may face similar racial threat perceptions as Blacks (Rocha and Espino 2009), they also face different obstacles that are frequently more challenging to study. Notably, Latinos are studied as a pan-ethnic group despite this being a function of administrative convenience, stemming from a 1970 census question that placed “Hispanics” as White with varying national origins (Rodriguez 2000). Latino political behavior is challenging to understand because Latinos, as a monolith, lack shared experiences (Beltran 2010).

One potentially binding experience, at least for some Latinos who recently migrated into the United States, is the experience of acculturation. Acculturation is the process in which individuals of different backgrounds and heritage learn the customs of the new culture in which they live (Alvarez-Rivera, Nobels, and Lersch 2014). The speed of acculturation, which may be conditioned by language differences and varying ethnic perceptions on government, may influence acceptance of societal norms. For example, several public health studies found that low levels of acculturation were significantly associated with lower use of health services (Solis, Marks, Garcia, and Shelton 1990; Lum and Vanderaa 2010). Other studies have found that acculturation influences the use of seat belts (Romano, Tippetts, Blackman, and Voas 2005) and voter registration and turnout (Xu 2005). Prior studies have examined the influence of how immigrants adapt to the surrounding culture, including the relationship of criminal activity and newly relocated individuals (Morenoff and Astor 2006).

Examining acculturation theory highlights the reality that Latinos may face language barriers in navigating the court system. In the political context, some adult Latinos rely on their children as “language brokers” in navigating the political system (Carlos 2021). Many studies have shown English proficiency is important to social integration, and limited proficiency can have a negative impact on income (Rosenstone and Hansen 1993; Andersen 2010; Bleakley and Chin 2010). In the courtroom, indigent defendants are generally entitled to court-appointed interpreters only during formal proceedings, when court is in session (Rahel 2013). For informal, off-the-record exchanges, there is no such assistance from an interpreter. This can disadvantage defendants, especially those unrepresented by counsel, in the plea negotiations process, which requires informal, out-of-court discussions with prosecutors (Rahel 2013). For something like routine traffic tickets, knowing that a prosecutor is even open to a reduction in one’s charge (which most generally are, to quickly move through their dockets) can be a key factor in case outcomes. If understanding the “rules of the game” differs by racial groups, disparate outcomes could result that may appear racially motivated but are, in fact, representative of a lack of the defendants’ knowledge of the courts.

With these theories in mind, numerous prior studies have found disparate outcomes based on race within the criminal justice system (Birch 2015; Smith and Maddan 2022). Although policing practices are not the focus of this article, the trends noted by other scholars concerning traffic stops, ticketing, and fines may transfer to the courthouse as well. Prior research shows that minorities are more likely to be ticketed (Farrell et al. 2004) and more likely to be fined for traffic violations when the law enforcement officer has that discretion (Makowshy and Stratman 2009). Part of the purpose of this study is to determine if these ticketing trends continue when those routine cases are resolved in court. For example, Johnson (2003) found that Black and Latino defendants were more likely to receive sentences above the proscribed sentencing guideline punishments than were Whites in Pennsylvania courts. Similar findings for Black defendants have been found at the federal district court level (Mustard 2001), particularly in light of Supreme Court decisions that lessened the mandatory nature of federal sentencing guidelines (Ulmer, Light, and Kramer 2011).

In addition, defendant characteristics may influence prosecutors’ decisions very early in the judicial processes (Sommers, Goldstein, and Baskin 2014). In “low information” cases, where prosecutors and judges have limited information about a defendant’s criminal history, a defendant’s race can end up serving as a proxy for inherent criminality (Berdejo 2018). For example, in their extensive examination of misdemeanor marijuana charges in New York City, Kutateladze, Andiloro, and Johnson (2016) found that minority defendants were less likely to receive reductions for plea bargains than were White defendants. While prior studies show disparate treatment based on the defendant’s race, these studies generally involve more substantial criminal activities, with little prior attention paid to more routine cases, such as traffic violations.

Attorneys: The answer to disparate treatment?

Numerous prior studies have noted the importance of attorneys as a potential factor in case outcomes (e.g., Haynie and Sill 2007; McAtee and McGuire 2007; Dumas, Haynie, and Daboval 2015; Smith and Maddan 2022). Attorneys, as repeat players in the judicial system (Galanter 1974), could serve as an important tempering factor, moderating the influence of defendant characteristics such as race. This may be particularly true if the representing attorney is a frequent participant within that specific jurisdiction. Often referred to as the court “workgroup” (Galanter 1974; Croyle 1983), lawyer relationships and knowing the unique local norms may influence outcomes beyond the facts of the particular case (Kautt 2002; Ulmer 2005). As Lynch and Omori point out in their examination of drug trafficking cases, trial courts may be considered “semi-autonomous set of systems governed by the same formal rules, statutes, and procedural policies, while also embedded in localized legal cultures … shaped by regionally specific historical contingencies and norms” (2014, 412). Attorneys who know these local norms and “going rates” (Ulmer 2005, 259) for certain types of crimes might help mitigate any disparate treatment based on the defendant’s race, age, residency, or other factors. Particularly for issues of acculturation where some racial groups may be less aware of the “rules” and “norms” of court, an attorney may diminish the institutional disadvantages that may face racial minorities.

Non-racial factors determining court outcomes

To examine racial factors and attorney’s role in minor criminal case outcomes, it is important to note the prior explorations of how criminal cases might be resolved, particularly before trial. Other personal traits of the defendant, such as gender and age, have also been found to influence sentencing (Birch 2015). Several prior studies also confirm that female defendants generally receive lighter sentences than male defendants (Mustard 2001; Johnson 2003; Starr 2015), although these differences may be dependent on the type of crime involved (see Rodriguez, Curry, and Lee 2006). Gender has also been found to play a role in whether a driver even receives a citation, as prior research shows that female drivers are less likely to be ticketed when stopped than male drivers (Lundman 1979; Makowsky and Stratman 2009; Farrell 2015). The age of the defendant has been viewed as another characteristic that could influence sentencing, with prior studies generally finding that older defendants receive lower sentences after controlling for other factors (Curry and Corral-Camacho 2008; Doerner and Demuth 2010; Birch 2015). The older the driver, the less likely they will receive a ticket as well (Farrell 2015), perhaps based on a policy of deterrence as younger drivers are statistically more likely to be involved in automobile accidents (Massie, Campbell, and Williams 1995).

Another important factor in sentencing at trial for all case types involves the defendant’s prior criminal record. Clearly, for cases that involve structured sentencing and sentencing guidelines, a defendant’s prior record is built into the normal sentencing range (Ulmer et al. 2011). Yet, previous studies also show that defendants with prior records may be less likely to be offered a reduced charge in exchange for a guilty plea (Kutateladze et al. 2016; Metcalfe 2016). Thus, although not an innate characteristic such as gender, age, or race, a defendant’s prior convictions may influence the processes and options in pending litigation. Even without a formal structured sentencing system for speeding tickets, a defendant’s prior driving record may be an important factor as to whether someone is offered a “good deal.” Prosecutors, as public servants, may want to reward good driving records; thus, those drivers may receive better plea bargains or dismissals.

Prior research also points to disparate treatment for defendants who live outside the court jurisdiction in regard to those who receive tickets. Research has shown that defendants who are not local to the area where charged may be more likely to be ticketed and are subjected to higher fines than citizens with closer geographic ties to where the ticket is adjudicated (Farrell 2015). Research by Makowsky and Stratman (2009) suggests that both economic and political motivations are involved in charging and fining out-of-town drivers. In their examination of traffic stops in Massachusetts, the authors proposed that fines for out-of-town drivers are used to increase local revenues, while elected leaders, such as sheriffs, prosecutors, and judges, are insulated from the political fallout over these higher monetary penalties. Given these incentives to lump additional burdens on out-of-town drivers when stopped, it could be that local defendants may be treated differently and receive better pleas than non-local defendants.

Jurisdictional-level factors might also have an impact on case outcomes.⁴ For example, prior research finds that Democratic trial judges are more likely to give lighter sentences than Republican judges (Schanzenbach and Tiller 2007; Fischman and Schanzenbach 2011; Yang 2014). For our purposes, judges are scarcely involved in traffic ticket resolutions, as most of the discretion lies within the prosecutor to offer or agree to a reduction. However, partisan considerations could influence how a prosecutorial district sets policies concerning pleas. While prior studies on prosecutors are somewhat mixed on partisanship (see Unah 2011), Democratic district attorneys may be more likely to reduce tickets than districts with Republican elected district attorneys, if Democratic district attorneys’ behaviors are similar to their judicial counterparts. Some prior studies also suggest that being “tough” on routine traffic violations may reduce future traffic accidents and serious injuries (Hingson, Howland, and Levenson 1988; de Figueiredo et al. 2001). Therefore, prosecutors and law enforcement in areas with a history of a high numbers of accidents may attempt to more strongly enforce traffic laws to reduce future accidents.

Assessing court outcomes in routine traffic cases

By focusing on defendant characteristics and the moderating effect of attorneys in routine cases, we are able to expand our knowledge of trial court interactions. As noted above, several important factors have been shown to influence criminal case outcomes. Race, in particular, can serve as a proxy for other information in judicial and prosecutorial decision-making (Berdejo 2018). Building on these prior studies of trials and some of the recent work that examines traffic stops, we would expect that, if the same implications hold, disparate impact may occur based on race in routine traffic ticket resolution. We thus would hypothesize that for Black drivers, Latinos, and other minority defendants:

H1(a): Black defendants will be less likely to receive a reduction than White defendants.

H1(b): Latino defendants will be less likely to receive a reduction than White defendants.

H1(c): Other non-White defendants will be less likely to receive a reduction than White defendants.

However, we also know that in most cases the presence of an attorney can influence case outcomes (Dumas, Haynie, and Daboval 2015; Metcalfe 2016). As Marc Galanter (1974) posited in his seminal work, lawyers can overcome obstacles faced by the parties, such as the unequal knowledge of legal norms. It is possible that having legal representation can serve as an equalizer, lessening the impact of other factors such as race, such that:

H2: Disparate treatment based on race will be mitigated for defendants represented by attorneys.

It is possible that the influence of a defendant’s race may be conditional on the attorney representation in a manner that is not identifiable through standard modeling. For this reason, an interactive term is included for each racial category to determine the conditional effects of race and attorney representation (Friedrich 1982; Kam and Franzese 2007). These conditional effects may portray a more nuanced analysis than an additive model alone.

The data for this analysis stem from the official electronic cataloguing software operated by the North Carolina Administrative Office of the Courts. A sample of all criminal cases was selected from 20 of the 100 counties in the state from a single calendar year. The counties were selected from representative categories based on the county population, geographic location, and urbanization. Geographically, North Carolina counties are divided among 16 Regional Councils of Governments (COGs),⁵ which provided one basis for geographic selection of counties. In addition, the North Carolina Rural Center identifies 78% of North Carolina counties as “rural” based on U.S. Census population density.⁶ Based on these criteria, to achieve a representative sample a county from each COG was included in the data, and 75% (15 of 20) of the counties are rural.

The cases from these 20 counties were randomly selected using a random number generator and criminal docket numbers. About 740 of these selected cases involved speeding tickets and were included in this analysis.⁷ Information from the cases was coded, including such factors as the race, gender, age, and home address of the defendant, whether an attorney represented the defendant, the defendant’s plea, the verdict, the sentence, information on the defense attorney (if present), and other case-level information.⁸

In examining case outcomes concerning speeding violations, one issue involves how to define and measure a “good deal” for each case. Winning and losing may be a bit more difficult to determine when all cases are plea bargained. Also, given the distinctive legal structures of every state’s DMV regulations, much of whether a defendant deems a plea successful may depend on whether the defendant could face a revocation of her license, whether she might have her insurance costs increased based on the verdict, or whether a high fine is imposed.

Given the intricacies of the North Carolina drivers’ points system and speeding laws, two separate dependent variables are tested concerning the final verdict for each case. The first dependent variable analyzes the reduction by simply subtracting the miles per hour over the speed limit as charged from the miles per hour over the speed limit at conviction. For this dependent variable, if the charge was reduced to a non-moving violation, dismissed, or granted a “prayer for judgment continued,” then the dependent variable was noted as the full amount of the original charge. So, for example, if a defendant was charged with driving 55 miles per hour in a 35 miles per hour zone and had the ticket reduced to 44 in a 35, the dependent variable for this observation would be “11.” If, however, the same charge was dismissed, the value for this observation would be “20.” The analyses for this dependent variable appear in Table 2 (below).

However, merely examining the amount of the reduction may underestimate the advantages a defendant receives from that reduction, particularly given the state’s laws. For this reason, a second dependent variable was created, a dichotomous variable that measures whether the defendant receives one of three potential “best outcomes” based on state laws.⁹ This category of best outcomes includes a dismissal of the case, a plea to a nonmoving violation,¹⁰ or a “Prayer for Judgement Continued” (PJC)¹¹ from the presiding judge. All of these options are considered ideal because, although there may be a fine, there are no license or insurance points, which could have significant consequences, such as higher insurance rates and/or the suspension of one’s driving privileges.¹² A dichotomous dependent variable was created that includes whether the defendant received one of these best three options (codes as a “1”) or they pled to a more severe penalty, coded as “0.” These dependent variables and the codification of all independent variables is listed in Table 1, below, with the results of models using this dependent variable included in Table 3, below.¹³

Table 1. Coding of Included Variables

The various independent variables are also detailed in Table 1. A variable is also included for out-of-state defendants. For the other jurisdictional variables, we use the number of per capita accidents within the county¹⁴ and the partisanship of the elected district attorney for the county. Determining the prior number of speeding tickets involved separate searches using defendant’s name in the same electronic database to find prior speeding ticket charges in the county. Overall, just under one-fourth (23%) of the defendants had prior speeding tickets within the county, with prior violations ranging from a single previous ticket to one driver (with a particularly heavy foot) who received 14 prior speeding tickets. The remaining 76% of defendants were first-time speeding offenders within the county. For the racial and gender variables, we also include North Carolina statewide Census data to assist in assessing the sample’s representation of the state as a whole (United States Census Bureau 2020). We note that while the sample is fairly similar to the state’s racial composition, males are somewhat overrepresented in our sample based on statewide demographics.

Results

Included in Table 2: Plea Reduction (MPH) are three models, each examining the overall reduction from the original charge to the plea using ordinary least squares regression.¹⁵ Model 1 displays the results from our most basic model, examining just race and attorney representation on case outcomes. Model 2 adds an interactive term for race and attorney representation. Model 3 displays our full model with all independent variables included.¹⁶

Table 2. Plea Reduction (MPH) Models

Note: Base term (excluded category) for race is White/Caucasian. Robust standard errors clustered by county.

* p < .05

^** p < .01

^*** p < .001 (two-tailed).

Concerning our race variables in Table 2, Model 1, we see some indication of racial disparity in case outcomes. The negative and statistically significant coefficient shows that Latino drivers received less of a reduction than Whites, amounting to a reduction that is around 3 mph less than that of a White defendant. Building on Model 1, Model 2 adds interactive terms for the racial cohorts and attorney representation. Here we see evidence of conditional effects based on the race/attorney interactive terms for Latino defendants across all models, suggesting the possibility that the presence of an attorney is more influential for Latino defendants than those of other races, after controlling for other factors. The effect of a Latino with an attorney is positive across all models, indicating that Latinos receive better deals. In Model 3, with other control variables included, Latino drivers are expected to plead to case outcomes over 4 miles per hour higher than White drivers. In Models 1, 2, and 3, the positive and statistically significant coefficient for attorney representation shows the consistent increase in the reduction. We also note that in models (not shown) where the excluded category is Latino defendants, we actually see across almost all models in Tables 2 and 3 that Latino defendant get “worse” plea deals than all other races, not just White defendants.¹⁷

Table 3. Plea Reduction (Best Outcome) Models

Note: Base term (excluded category) for race is White/Caucasian. Robust standard errors clustered by county.

* p < .05

^** p < .01

^*** p < .001 (two-tailed).

The full model (Model 3) includes all other variables of interest, including specifics of the case, defendant characteristics besides race, and jurisdictional factors. We see no difference based on age, gender, or whether the defendant was from North Carolina or out of state, despite prior studies finding them significant. The legal factors indeed mattered, as those with more tickets in the county (Prior Tickets) received fewer reductions as expected in Model 3. We see a somewhat unexpected result with the level of charge in Model 3, as that variable is statistically significant, but in the opposite direction than one might think. This unanticipated result could be from the fact that with a higher charge more room exists for negotiations, and thus a prosecutor is more likely to give “something” for a plea, even if shaving off a few miles per hour from the original charge had little overall impact on the outcome. We also see that the presence of a Democratic elected district attorney increases the plea reduction but that the rate of accidents in the county was not statistically significant.

We note, however, that a mere analysis of the overall reduction in charge based on the miles-per-hour may not accurately represent the disadvantage of Latino drivers based on state law. To this end, we also conduct analyses exploring the likelihood of receiving one of the three best outcomes for drivers: complete dismissals, a plea to a nonmoving violation, or the granting of a prayer for judgment continued. As mentioned above, these three outcomes represent no license points or insurance points; thus the long-term consequences are very limited for these outcomes. Similar to Table 2, we build our models in Table 3: Plea Reduction (Best Outcome) from most basic (Model 4) to the full model including all independent variables (Model 6).

The results for the models in Table 3: Plea Reduction (Best Outcome) largely mirror the analyses in Table 2: Plea Reduction (MPH). Again, we see consistently across all models that Latinos are less likely than Whites and other racial cohorts to receive the best outcome.¹⁸ Also, attorney representation is positive and statistically significant across all models, indicating the importance of legal assistance even in these routine cases. In Model 5, which adds the interactive terms, once again we see the conditional relationship between Latino defendants and attorney representation. This conditional relationship is not present for the other racial cohorts.

Model 6 provides some further insights as we add all the other potentially influential factors. Latino defendants, again, are less likely to receive a plea reduction and are more than 25% less likely to receive a plea reduction than a White defendant (0.44 predicted probability for Latinos and 0.61 for White defendants). Attorneys also continue their strong, substantive effect on outcomes with those represented by attorneys having a 0.79 predicted probability of a sentence reduction as opposed to 0.50 for those unrepresentative. Rather than a coin-flip possibility at a better outcome, individuals represented by a lawyer have a 63% increase in the likelihood they will receive the “best outcome” in their case, mirroring our theoretical expectations.

The most striking result, however, comes with the interactive effect between the Latino defendant and attorney representation variable. Latino defendants unrepresented by a lawyer stand a 1 in 5 (0.19 predicted probability) chance of receiving a best outcome reduction. However, a Latino defendant represented by a lawyer quadruples their chance of a reduction, with an increased predicted probability of a “best outcome” of 0.91. Figure 1 plots these stark differences based on representation among Latino defendants.

Figure 1. Predicted Probability of “Best Outcome.” Note: Bars represent the 95% confidence interval.

Unlike the overall reduction models in Table 2, our “best outcome” models in Table 3 do show significant results in that male defendants were less likely than female defendants to receive one of the best outcomes, as were out-of-state defendants. Consistent with expectations, those drivers charged with driving at a higher speed were less likely to receive the best plea outcome. However, prior tickets in the county did not appear influential.

Discussion

These results shine a light on two important aspects of our court system: racial discrimination in the lower courts and the important role attorneys play in achieving better outcomes. Lower courts, especially state courts, receive little attention in political science research. Despite the scarcity of research, these courts wield a great deal of power over individuals, frequently deciding cases and changing lives. Facing a case in a lower court as a defendant can be punishment in and of itself, even before the case reaches its ultimate disposition (Feely 1979). Misdemeanor cases can lead to large intrusions on liberty, with jail or probation sentences, permanent criminal records, unpayable fines and fees, immigration consequences, and a loss of respect for the political system at large (Natapoff 2018). This negative perception can lead to less civic engagement and lower participation in society (Brayne 2014).

The strong, negative effect on unrepresented Latinos exposes a startling reality. While Latinos are not a monolith, and we do not have evidence of the recency of potential immigration, this points to a larger problem faced by these defendants in the court system: language barriers. A lack of access to interpreters beyond the formal court appearance may be leading to large deficiencies in due process in the courtroom. This disparate treatment is brought into further relief given how much attorney representation improves outcomes for these defendants.

Negative interactions, especially those seemingly motivated by racial bias, can lead to a long-term erosion of legitimacy and support for the courts and criminal justice system writ large. Courts rely on institutional legitimacy for their long-term support. Building this institutional legitimacy is not a one-off endeavor; it requires well-received decisions over a long period of time (Scherer and Curry 2010). Without this institutional legitimacy, courts could face challenges in parties complying with their decisions altogether.

Second, this study further refines our understanding of the role of attorneys in the legal system. The findings suggest a potential important equalizer for some racial disparities: the presence of an attorney. While a deeper investigation beyond our scope is needed to determine the exact cause of why Latino drivers receive harsher treatment compared to others, the conditional effects of attorney representation suggest that acculturation (Alvarez-Rivera, Nobels, and Lersch 2014) may be a main factor.

While we lack immigration data or first-generation status information for the Latino drivers including in this study, it is possible that this group has less experience with the court system than those of other races. Perhaps the disparate treatment is due to a lack of knowledge about the “rules of the game” more than intentional discrimination or subconscious racism. However, we do not see the conditional effects present for Black defendants with attorneys and some evidence of disparate treatment (Table 3, Model 6), suggesting the mitigating factor of having an attorney may not remove disparate treatment in the same way for all races.

The presence of an attorney appears to diminish this disparate impact based on race at least in some instances.¹⁹ Perhaps the prosecutor never even realizes the defendant’s race, age, or gender as they deal only with defense attorneys. In addition to the need for expanded access attorneys, possibly through legal aid, if a lack of information is the true issue, other options such as clear procedures and policies for prosecutor’s offices, more court-provided resources for less knowledgeable court participants, and easier access to the written and unwritten “rules” of the game may help lessen some racial disparities. Given our results for Latino drivers and that language may be an issue, this additional help may extend to more court interpreters, multi-language forms, and expanding community efforts in Latino communities to help them gain an understanding of court procedures.

In addition, this paper also adds to this line of research with the geographic component, noting the potential disparate treatment of out-of-state defendants. As others suggest (Makowsky and Stratman 2009), this finding could perhaps be due to economic and political incentives in placing higher economic burdens on those who have little political influence in the area. However, there could be less nefarious reasons for this apparent disparate treatment of non-local defendants. For example, the benefits of seeking a ticket reduction may not be worth the costs to travel back to the jurisdiction and non-local defendants try to simply pay it off, or reductions may not be benefit them based on their state’s motor vehicle laws. While we cannot make assumptions based on this study alone, again future projects may be needed to explore this aspect more fully.

We embrace and engage with some of the limitations of our project. Foremost, this project addresses cases from only one state. While we note that our sample is representative of statewide racial demographics and that we have no a priori reason to presume the North Carolina trial court system is an outlier or that relationships would be different in other states, it may limit this study’s generalizability as applied to other states and crimes. That said, studies from a single state can also offer advantages, including an increase in internal validity, the ability to include more refined measurements of important concepts, and the ability to contextualize data in ways not available in multistate studies (Nicholson-Crotty and Meier 2002). For example, the ability to develop accurate plea rankings for speeding tickets, as done in this study, may prove impossible across multiple states given each jurisdiction’s unique motor vehicle and insurance laws. In addition, this project joins numerous other examples of innovative works using data from just one state, such as an exploration of how registration rules influence turnout in Wisconsin (Burden and Neiheisel 2013), the influence of public financing on judicial behavior in North Carolina (Hazelton, Montgomery, and Nyhan 2016), and the factors that influence public perceptions of judicial fairness in Mississippi (Overby, Brown, Bruce, Smith, and Winkle 2005).

Another important caveat from our results: they only observe behavior once a case is before a court. This omits the selection process by police officers who choose which motorists to stop and receive citations. Our models do not account for these decisions, creating some sample-selection bias. While many are making headway into statistically disentangling sample-selection bias (Knox, Lowe, and Mummolo 2020; Clark et al. 2020), officers’ choices upstream from the courts still makes a difference in outcomes. It is likely, given that police stop Black motorists more often than White motorists (Knowles, Perisico, and Todd 2001) and that officers are more likely to use force against Black civilians than White civilians (Clark et al. 2020), that our sample population could contain some bias against Black drivers “baked in” to the population and does not fully capture the potential bias this group faces. This may be a potential shortcoming in our model that could bolster our results even further, much like how prosecutors’ charging decisions may underestimate racial disparities in sentencing (Rehavi and Starr 2014).

Currently, researchers are exploring numerous issues involving significant factors and aspects of our judicial system. This examination of routine speeding tickets may appear to be on a much smaller scale than, for example, the long-term implications of Supreme Court precedent, the ability of the courts to check the other branches of government, or the role of race in death penalty cases. However, if we want to know what trial courts and trial lawyers actually “do,” how the courts that conduct the vast majority of work in our system (i.e., state trial courts) complete their business, and what normal citizens may experience in what for many people is their only interaction with the court system, then we sometimes need to get “down in the weeds” with studies such as this. Only through this type of research can we gain a full understanding of all facets of our judiciary.