The Usual, Racialized, Suspects: The Consequence of Police Contacts with Black and White Youth on Adult Arrest

Abstract

Research on race and policing indicates that Black Americans experience a greater frequency of police contacts, discretionary stops, and police harassment when stops occur. Yet, studies examining the long-term consequences of police contact with young people have not examined whether criminal justice consequences of police contact differ by race. We address this issue by examining whether police encounters with children and adolescents predict arrest in young adulthood and if these effects are the same for Black and White individuals. The paper uses longitudinal survey data from 331 Black and White respondents enrolled in the Seattle Public School District as eighth graders in 2001 and 2002. Our findings indicate that police encounters in childhood increase the risk of arrest in young adulthood for Black but not White respondents. Black respondents who experience contact with the police by the eighth grade have eleven times greater odds of being arrested when they are 20 years old than their White counterparts.

Police stops represent an entry point into the criminal justice system that, until recently, have received less scholarly attention than other stages of the criminal justice process. Recent research suggests that stops interpreted as harassing, unfair, or discriminatory damage public trust in police (Epp, Maynard-Moody, and Haider-Markel 2014; Fagan et al. 2010; Fagan and Tyler 2005; Meares 2014), encourage avoidance of police and other record-keeping institutions (e.g., hospitals and schools) (Brayne 2014; Brunson and Weitzer 2008; Gau and Brunson 2015; Remster and Kramer 2018), and negatively influence community member and individual mental health (Geller et al. 2014; Sewell, Jefferson, and Lee 2016). Scholars examining police stops and arrests using a labeling theory approach argue that stops may lead individuals who were contacted to engage in future illegal behaviors. This perspective posits that individuals who experience police stops align their identities with the criminal label and subsequently engage in illegal activities. Recent work in this tradition has extended from a singular focus on how labels are internalized to an examination of the processes by which labels erect structural barriers. Research suggests that police contacts and arrests create obstacles to education and employment (Bernburg and Krohn 2003; Kirk and Sampson 2013; Lopes et al. 2012; Wiley, Slocum, and Esbensen 2013), subsequently leaving individuals with fewer opportunities to engage in legal activities and ultimately increasing the risk of recidivism (Bernburg and Krohn 2003; Lopes et al. 2012).

Labeling theory, however, has been somewhat myopically focused on how criminal offending may increase future offending. Recent scholarship has acknowledged this and begun to unpack how criminal justice involvement may increase the risk of stops and arrests beyond what can be accounted for through engagement in illegal behaviors. Liberman, Kirk, and Kim (2014) find that arrested youth are more likely than their non-arrested peers to be arrested later—a finding that is not explained by differences in offending. The authors refer to the process by which arrest shapes law enforcement’s subsequent responses to youth as “secondary sanctioning”; they join others who have similarly found that, for young people, being stopped by police predicts future stops. McAra and McVie (2005) argue that young people with police contact become “usual suspects” and the focus of subsequent police attention, regardless of their law-abidingness. These studies underscore the importance of police contact for future contacts and arrests. However, they also leave important questions unanswered. Notably absent is an examination of how race influences the secondary sanctioning process. Any American study of policing must contend with the complex ways that experiences with the criminal justice system are shaped by race. It is well-documented that race differences exist in the number, type, and content of encounters that police have with the public (Brunson and Weitzer 2008; Crutchfield et al. 2012; Epp et al. 2014; Fagan et al. 2016; Hagan, Shedd, and Payne 2005; Meares 2014; Piquero 2008). We add an important perspective to both labeling theory and race and policing scholarship by centering our attention on how race and police contacts interact to produce different criminal justice consequences.

Three main questions orient this paper: 1) What effect, if any, does police contact or arrest by early adolescence (preceding the eighth grade) have on the chances of arrest in young adulthood (age 20)? 2) If there is an effect, is it explained by engagement in “secondary deviance”? 3) Are the effects of early police contact on young adult arrest the same for Black and White youth? To answer these questions, we analyze four waves of data from a longitudinal study of 331 Black and White respondents and their parents in Seattle, Washington. The data span six years, from the time when the youth respondents were in eighth grade until they were 20 years old.

RACE, POLICE STOPS, AND THEIR CONSEQUENCES

Proactive policing practices, such as hot-spot policing, stop-and-frisk policies, and “investigatory stops,” have become ubiquitous over the past four decades—expanding police discretion and increasing police contacts with the public. These types of proactive policies are scrutinized, in part, because they tend disproportionately to target Black and Latinx individuals and communities (Epp et al. 2014) and young people (New York Civil Liberties Union 2018; Rengifo and Pater 2017). Investigatory stops, for example, allow police to use minor transgressions, such as the absence of a functioning license plate light, to pursue evidence of criminal activity. These stops rely on police suspicion when determining whether to stop an individual and search for evidence of a crime (Epp et al. 2014; Roberts 1998). In a study of motorists in the Kansas City metropolitan area, Epp and colleagues (2014) find that while Black and White motorists are stopped at similar rates for traffic violations, Black drivers are nearly 2.7 times more likely than White drivers to be stopped in an investigatory stop. Similarly, evidence from New York City shows that police disproportionately target Latinx and Black individuals for stop-and-frisks (Fagan and Davies 2000; Fagan et al. 2016; Gelman, Fagan, and Kiss 2007). And yet, stop-and-frisks of Black and Latinx residents were less likely to lead to criminal arrests, leading some to argue that police have more relaxed standards when it comes to stopping residents of color (Gelman et al. 2007). In Seattle, scholars have found evidence that racial stereotypes linking Black individuals with drug use lead police to surveil racially heterogeneous spaces and draw police attention away from White individuals’ engagement in illegal drug sales (Beckett et al. 2005).

However, it is not only that people of color are more vulnerable to police contact. A substantial body of research suggests that Black and White Americans have different experiences of policing. Black Americans are more likely to report poor treatment when stops occur (Brunson and Weitzer 2008; Rengifo and Pater 2017; Weitzer and Tuch 2004). Young Black and Latino men report that the police consistently treat them as criminals—even as they engage in daily routines (Brunson 2007; Jones 2014; Rios 2011), and young people of color report more physical and verbal harassment at the hands of police (Brunson 2007; Brunson and Weitzer 2008; Jones 2014; Jones-Brown 2000; Rios 2011). Even when police act politely, highly discretionary stops send messages about assumed criminality and citizenship. Fairly conducted traffic-safety stops “tell a driver that they are an equal member of society and are held to the same rules as everybody else. Racially biased investigatory stops tell a driver that they look like a criminal and people like them are subject to arbitrary control befitting their subordinate status; they are not an equal member of society” (Epp et al. 2014:135). Investigatory stops of motorists or similar street stops convey strong symbolic messages about police views of criminality. When police officers do not provide legitimate reasons for contact, the belief that stops are deeply unfair or discriminatory is reinforced (Rengifo and Pater 2017).

The number, type, level of intrusiveness, and officer behavior during stops influence trust in police and police legitimacy (Epp et al. 2014; Gau and Brunson 2015; Tyler, Fagan, and Geller 2014; Tyler, Jackson, and Mentovich 2015). Studying the effects of the New York Police Department’s use of “stop and frisk,” Tyler et al. (2014) found the number and intrusiveness of stops young men experience or witness affects their views of police legitimacy—individuals with greater numbers of police contacts and more intrusive stops were less apt to view the police as legitimate. Tyler argues, therefore, that “procedural justice,” or the manner in which officers interact with the public, plays a crucial role in individuals’ assessments of police legitimacy. When police interactions leave people feeling as if they have been treated fairly, they are more likely to view the police as legitimate. However, when police interactions are perceived as harassing, disrespectful, or unfair, assessments of police legitimacy and trust for police dwindles (Meares, Tyler, and Gardener 2015; Sunshine and Tyler 2003; Tyler 2005). Importantly, accumulated negative interactions not only influence the contacted individual’s perceptions of the police but also affect the attitudes of others who hear stories of police harassment (Jones-Brown 2000).

Police contacts may also have mental health consequences for both those experiencing stops and residents living in neighborhoods with aggressive policing practices. Geller and colleagues (2014) Geller and colleagues (2014) found that young men stopped by police as part of New York City’s stop-and-frisk policy reported higher likelihoods of experiencing post-traumatic stress disorder (Geller et al. 2014). Men living in neighborhoods with aggressive stop-and-frisk activities were more likely to experience psychological distress, regardless of whether they were stopped by police (Sewell et al. 2016). Since encounters with police are frequently unpleasant and potentially frightening, individuals who experience these types of contacts may adopt coping behaviors to avoid future interactions. For example, Weitzer and Brunson (2009) found that individuals who had multiple and repeated negative encounters with police officers engaged in avoidance strategies—evading police for fear of further negative interactions. Such strategies, however, might increase future contacts, since police find avoidance behavior “suspicious” and associate avoidance with attempts to conceal illegal behavior (Geller et al. 2014).

POLICING THE LIVES OF YOUNG PEOPLE

Many young people face a dual exposure to police. Officers both patrol the neighborhoods in which they live, and they are frequently stationed within schools. School Resource Officers (SROs) are law enforcement officers who work within schools. The precipitous rise of police in schools through the use of SROs over the past twenty years has also increased opportunities for contact between law enforcement and children (Bureau of Justice Statistics 2010; James and McCallion 2013). However, their roles are often varied, vague, and, at times, expansive. A report published by the U.S. Department of Justice suggests that SRO duties include: informally mentoring students, educating students, and enforcing the law (Finn and McDevitt 2005). In 1997, the Bureau of Justice Statistics reported that approximately 9,000 police officers were employed as SROs.¹ Ten years later, the number had increased by 3,700 (Bureau of Justice Statistics 2010). In 2017, nearly 65 percent of middle and high schools across the United States reported having at least one SRO or other law enforcement official working within the school (Musu-Gillette et al. 2018).²

The use of police in schools is frequently attributed to a rise in public concern about school violence and federal legislation which provided financial incentives for schools to use officers (French-Marcelin 2017; Kupchik 2010, 2016; Nolan 2011; Price 2009; Simon 2007). In this atmosphere, SROs are seen as a tool for achieving safer schools. Yet, their presence may be detrimental to some students. Police presence in schools and zero tolerance policies can bring small acts of disorder that were formerly dealt with at the school level to the attention of criminal justice (Na and Gottfredson 2013; Theriot 2009) and lead to exclusionary disciplinary measures (Fisher and Hennessy 2016). The use of police in schools may also increase the risk that students will experience police contacts outside of school. Victor Rios (2011) finds that SROs at a public school in Oakland routinely issued court summons for status offenses such a skipping school. Students’ failure to comply with the summons frequently led to warrants or probation status and increased police attention on students outside of schools. Rios (2011:45) refers to this process as “labeling hype” a process by which, “young men were caught in a spiral of punitive responses imposed by institutions which labeled them deviant” (Rios 2011:45).

Students of color and poor students are at an increased risk of experiencing police encounters within schools. Police officers are more likely to be placed in poor schools (Kupchik and Ward 2014)³ and often respond to classroom discipline issues—in which Black and Latinx students are already overrepresented (Eitle and Eitle 2004; Kupchik 2010; Nicholson-Crotty, Birchmeier, and Valentine 2009; Nolan 2011; Rios 2011; Skiba et al. 2002). Further, as Kupchik (2010) and Nolan (2011) demonstrate, police officers stationed in hallways frequently become enforcers of school dress code, question students about tardiness, and stop students for other minor violations. Kupchik (2010:184) argues that students of color are “more vulnerable than Whites students to subjective appraisals of misconduct, but they also are less likely to respond in ways that avoid punishment” (Kupchik 2010:184). While any student attending a school with a police officer may be subjected to police contact, students of color and poor students appear disproportionately burdened by their presence.

LABELING THEORY AND CRIMINAL JUSTICE CONSEQUENCES

While police contact with young people is frequently regarded as a crime deterrent strategy, contacts and arrests may increase the likelihood contacted individuals will participate in illegal behaviors (Liberman et al. 2014; Wiley and Esbensen 2016; Wiley et al. 2013). Much of this research uses a labeling theory perspective. The theory argues that an individual who has experienced an arrest may be perceived as “deviant” by police, peers, teachers, employers, and others. The stigma of a “deviant” label creates subsequent barriers to education, employment, or other important socializing institutions and limits available opportunities (Bernburg and Krohn 2003; Kirk and Sampson 2013; Lopes et al. 2012; Wiley et al. 2013). These barriers can lead to engagement in deviant behavior and development of deviant attitudes/identities (Bernburg and Krohn 2003; Lopes et al. 2012; Wiley and Esbensen 2016).

Labeling research has primarily focused on the role of police contacts and arrests on secondary deviance. However, as noted previously, recent work considers how contact with the criminal justice system may lead to “secondary sanctioning”; Liberman et al. (2014) argue that criminal justice actors and institutions also label individuals with a prior arrest. Police provide an initial “criminal” label and then respond to the label by increasing surveillance and/or creating a lower threshold for re-arrest when an individual has a prior record. Individuals who experience an arrest are more likely than their peers who engage in similar levels of illegal behavior to be re-arrested. Similar work in Seattle and Edenborough find that police encounters, not just arrest, increase the risk of future police contacts (Crutchfield et al. 2009; McAra and McVie 2005). McAra and McVie (2005) argue that young people’s illegal behaviors and their being friends with individuals who have had adversarial police contact increase the risk of experiencing an initial contact with police. However, once young people have had a contact, “they then become part of the permanent suspect population and, as a consequence, any of their friends and associates who have not had past experience of adversarial police contact, become suspect too” (McAra and McVie 2005:26). Police see young people with previous contact as the “usual suspects.” Contact, thus, begets more contact.

These studies provide compelling evidence that police contacts and arrests increase the risk of future contacts and arrests—findings that cannot be explained by engagement in illegal activity alone. However, they do not address whether there are racial differences in the secondary sanctioning process. We build upon this work by investigating whether police contacts have different long-term criminal justice implications for Black and White Americans. We expand on the observations of early labeling theorists and more recent scholarship which examines the relationship between race and policing. Early labeling theorists (Becker 1997; Lemert 1973) emphasized that those with more resources were more likely to withstand the labeling process and its negative consequences. In the United States, being White remains an important and potent resource. Recent work points to the ways in which young people of color are drawn into a web of hyper-surveillance (Jones 2014; Rios 2011; Roberts 1998; Simon 2007), the often negative nature of interactions with police when contact occurs (Brunson 2007; Kupchik 2010; Nolan 2011; Rios 2011), and the consequences of those interactions on trust in the criminal justice system and on mental health (Fagan and Tyler 2005; Gau and Brunson 2015; Hagan et al. 2005; Tyler 2005). In the remainder of the paper, we work to bridge these two literatures: examining the criminal justice consequences for police contact for Black and White individuals.

DATA AND METHOD

Sample

This paper uses data from the Family Connections Study—a stratified random sample of eighth-grade Black and White students in the Seattle Public School District in 2001 or 2002. Parents of eligible eighth-grade students in the Seattle School District received a letter describing the study and were contacted by phone. Eligibility was based on student records indicating that the student was either African American or European American, spoke English as their primary language, and planned to live in the area for at least six months. Families were included if the teen and one or both parents consented to participate. Forty-six percent of those contacted agreed to participate (55 percent of Blacks and 40 percent of Whites). Parents who declined to participate were more likely to be White, married, and to have a higher education on average than those who did consent. The sample was stratified by teen race and gender. A portion of the teens in each racial group self-identified as more than one race (19.6 percent Black youth, 12.5 percent White youth), but were included in these analyses. At the time of the first interview, respondents were enrolled in one of eighteen different schools throughout the Seattle Public School District.

Data collectors went to the families’ homes four times in three years: fall and spring of the eighth grade, spring of ninth grade, and spring of tenth grade. Questionnaires were self-administered to teens and their parents in their homes using laptop computers while the data collector was present, ensuring that parents did not monitor their teens’ responses. Family members received $15 each time they completed questionnaires. The family received additional funds for completing other components of the assessment. At the two young adult assessments (when youth were, on average, ages 20 and 22 years old) respondents who completed the survey received $50 as well as additional funds for completing other components of the assessment.

The original sample consisted of 331 respondents. Complete data are available from 261 respondents (79 percent of the original sample) using four waves of data: baseline (fall eighth-grade parent and youth, tenth-grade youth, and approximately age 20youth surveys). In adolescence, overall attrition of participants from the study was relatively low: five percent at post-test, eight percent in ninth grade, and nine percent in tenth grade. Furthermore, there were no race differences in attrition (X² [1, 331] at post-test=2.56, p=.11; 9^th grade=.082, p=.77; 10^th grade=.50, p=.48), and almost no differences between non-responders and participants on key outcomes at baseline (i.e., index of substance use initiation, violent behavior, initiation of sex, favorable attitudes about substance use, perceived harm of substance abuse). In young adulthood, attrition was nine percent at age 20. There were differences by race (age 20 loss for White 5 percent vs. Blacks 14 percent, X² [1, 331] = 7.66, p<.01, as well as other demographics (see Klima et al. 2014). Non-responders at age 20 were not different from participants on tenth-grade variables related to criminal behavior (i.e., use of illicit drugs, marijuana, heavy episodic drinking, daily smoking). To reduce bias due to attrition, we used a multiple imputation model. Based on current standards (Graham 2012), we imputed 40 data sets using the variables in the hypothesis, testing models for the basis of imputation. This is a particularly efficient way to reduce bias due to missingness, because data on criminal behavior and police contacts in adolescence when very little data are missing are used to impute similar variables in early adulthood.

Measures

Early Police Contact is the primary predictor of interest. Unlike many previous studies, contact includes both encounters that do not result in an arrest and those that do end in arrest. Police contact is measured using three self–reported items in eighth grade. Contact was coded as having occurred if the teen responded affirmatively to any of the following questions: Have you 1) ever been stopped by the police, but not arrested; 2) ever been stopped by the police and arrested; 3) ever been in trouble with the police for something you did? By the eighth grade, 74 (28.4 percent) of the youth had had a police contact. This proportion is similar to another multi-city analysis during a similar time-frame (Wiley et al. 2013), suggesting that Seattle is similar to other locations in rates of police-juvenile encounters. Responses to the police contact questions differed by race, with Black youth being more likely than their White peers to experience a police encounter by the time they were in eighth grade (36.8 percent vs. 21.5 percent, respectively). This difference is statistically significant (t[228.6] = -2.698, p<.01). See Table 1 for descriptive statistics by race.

Table 1.

Means and Standard Errors by Race with Two-Sample t-test with Unequal Variances Using Observed (non-missing) Data

	Black Respondentsn=117	White Respondentsn=144	Difference
Per capita Income	8,254 (860)	21,536 (1,170)	13,282 (1,452)***
Female	.53 (.046)	.49 (.042)	−.04 (-.044)
Age at Last Survey	20.24 (.061)	20.26 (.050)	.02 (.079)
8th-grade Police Contact	.37 (.045)	.22 (.034)	−.15 (.056)**
8th-grade Illegal Behavior	.59 (.046)	.51 (.041)	−.08(.062)
10th-grade Arrest	.11 (.029)	.05 (.018)	−.06 (.034)
Age 20 Illegal Behavior	.32 (.043)	.53 (.042)	.21(.060)***
Age 20 Arrest	.15 (.033)	.06 (.020)	−.08 (.038)*

^*p<.05

^**p<.01

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

Arrest at Age 20 was measured using a count item from the survey. Respondents were asked the number of times they had been arrested in the past year. Responses range from 0 to 10 times. We are primarily interested in involvement in the criminal justice system and not the level of involvement. Therefore, the variable was dichotomized, where 0 indicates the individual did not experience an arrest and 1 indicates the individual had an arrest in the past year. Ten percent (26) of the sample reported having experienced an arrest in the past year. Responses to the question differed by race, with Black respondents being significantly more likely to report an arrest than White respondents (14.5 percent vs. 6.3 percent, respectively) (t[198.23] =-2.15, p<.05) .

Age 20 Illegal Behavior is treated as a mediating variable in the subsequent analyses. We combine measures of property crime, violent crime, and illegal drug use from the age 20 survey. Participants were asked about the number of times that they had engaged in a variety of illegal behaviors in the past year. We included crimes that violated the law and could have been enforced by police. Property crime consists of four items: 1) arson; 2) stealing from a store; 3) breaking into property; and 4) drawing graffiti without permission. Sixty-three (24.1 percent) participants reported engaging in at least one property crime in the past year at age 20. Violent crime consists of three items: 1) using weapon/force/strong arm to get something from someone; 2) hitting someone with the intention of hurting them; 3) throwing an object at a car or person. At age 20, 40 (15.3 percent) participants reported engaging in violent crime in the past year.

Illegal Drug Use includes eight questions about past year frequency of use for the following drugs: LSD/Psychedelics, cocaine/crack, stimulants, party drugs, non-prescribed steroids, non-prescribed Vicodin, non-prescribed OxyContin, and any other illegal drugs. The response categories for each of the drugs were “Never,” “1-2 times,” “3-5 times,” “6-9 times,” “10-19 times,” “20-39 times,” and “40+ times.” 74 (28.4 percent) respondents reported having used at least one illegal drug in the past year. The items were then combined for an overall score of illegal behavior. 115 respondents (44 percent) reported engaging in at least one illegal activity in the past year. While we collected frequency data for these items, the preponderance of 0s and the relative rarity of scores above 1 on any of the three types of crimes produce predictor variables which are highly skewed and introduces a problem with outliers. Dichotomizing eliminates these problems while admittedly introducing some lack of precision. Engagement in illegal behavior differs by race, with White respondents being more likely to report participating in illegal behavior than Black respondents (53 percent vs. 32 percent) (t[253.55]=3.48, p<.001). When crime types are examined separately, White respondents were significantly more likely to report using illegal drugs in the past year (40 percent vs. 14 percent) (t[252.84] =5.12, p<.001). There are no statistically significant differences between violent or property crime offenses by race.

Adolescent Illegal Behavior at the eighth grade is used as a control. The addition of this variable ensures that we are not simply capturing “bad kids.” The argument could be made that youth who engage in property crime, violent crime, or illegal drug use when they are young are more likely to be contacted by police. Therefore, any effect of police contact would simply be capturing misbehaving youth whose crimes persisted into adulthood. To make sure that we are not simply capturing youth who misbehave in childhood, we control for early behavior. Similar to other measures, this variable is based on engagement in illegal behavior in three domains: crimes against property, violence, and illegal drug use. Property crime was measured using three items. Respondents were asked to mark “yes” if they had ever engaged in the following behaviors: 1) arson, 2) theft, or 3) drawn graffiti without permission. Violent crimes consisted of three variables: 1) carrying a handgun to school, 2) hitting someone with the intention of hurting them, and 3) throwing an object at a person or vehicle. In addition to violent and property crime, at the eighth grade, individuals were asked if they had ever used alcohol, tobacco, or other illegal drugs.⁴ These responses were dichotomized, with engagement of any of these behaviors being coded as 1 and a 0 for those who did not report engaging in the above behaviors. Eighth-grade indicators of illegal behavior did not differ by race.

Race was based on parents’ reports of their child’s race on school enrollment forms (0 = White, 1 = Black). Gender was reported by teens on the survey (0 = male, 1 = female) and is used as a control in the current analyses. Household Per Capita Income is also used as a control. Household per capita income was calculated from two items from the parent survey when the child was in eighth grade. The items ask for family income before taxes in the past year. Respondents were presented with 18 categories representing ranges from $10,000 or under to $200,000 or over. The mid-point of the range for each category was divided by the number of people the parent reported were living in the household. Per capita household income has a minimum value of $786 per year/per person and maximum value of $75,000 and mean of $15,582. Consistent with national research on wealth and income inequality (Oliver and Shapiro 2006), parents of Black children reported significantly lower per capita household incomes than parents of White children, with means of $8,254 per person/per year for households of Black youths and a mean of $21,537 per person/per year for households of White youths (t[249.49]=9.15, p<.001). Income data are right skewed and, therefore, a log transformation of the variable was created. Age was based on the calculated age of the respondent at the time of their young adult interview. Respondents’ ages at this data collection point ranged from 18.73 to 22.42 years with a mean of 20.26 years.⁵ Since the opportunity of the individual to experience an arrest would be different for an 18-year-old versus a 22-year-old, age is used as a control in the following analyses.

Analytic Method

Using SAS version 9.4, we perform a series of logistic regression models across 40 imputed datasets and use PROC MIANALYZE (SAS Institute 2002) to summarize and provide appropriate standard errors. We estimate three separate sets of logistic regression models. The first set of models uses a step-in approach, looking at the relationship between early police encounters and young adult arrest with background and demographic characteristics, early illegal behaviors, tenth-grade criminal justice involvement, current illegal behavior, and adding an interaction term for race by police contact. The second set of models examines the relationship between the previously mentioned variables separated out by race—examining the relationship between these variables for Black and White respondents. The final model uses the full sample to examine the effect of police contact by the eighth grade on engagement in young adult illegal behaviors (age 20) and includes a race by contact interaction effect.

RESULTS

Adolescent Encounters with Police and Young Adult Arrest

The first model examines the relationship between early police contact and arrest in young adulthood while controlling for background variables that have been shown to increase the chances of arrest. Table 2 displays the relationship between early police contacts/encounters and arrest while controlling for demographic and background variables (Model 1), eighth-grade illegal behaviors (Model 2), tenth-grade arrest (Model 3), and concurrent (age 20) illegal behaviors (Model 4).

Table 2:

Logistic Regression of Arrest at Age 20 for Full Sample (N=331)

Predictor	Model 1		Model 2		Model 3		Model 4		Model 5
Variable of Interest
8th-grade Police Contact	6.07 (0.40)***		5.95	(0.45)***	5.23	(0.46)***	4.66	(0.49)**	1.23	(0.73)
Demographics
Age	1.36 (0.26)		1.37	(0.26)	1.39	(0.28)	1.34	(0.27)	1.42	(0.30)
Black	1.77 (0.47)		1.79	(0.47)	1.84	(0.48)	2.85	(0.51)	0.85	(0.73)
Female	0.68 (0.40)		0.67	(0.40)	0.67	(0.40)	0.72	(0.42)	0.72	(0.44)
Log Income 8th-grade Illegal Behavior	0.74 (0.23)		0.74	(0.23)	0.80	(0.24)	0.76	(0.26)	0.76	(0.27)
8th-grade Property, Violent, and Drug Use	---		1.05	(0.47)	1.06	(0.48)	0.93	(0.51)	0.89	(0.51)
Early Criminal Justice Involvement
10th-grade Arrest	---		---		2.16	(0.55)	1.67	(0.58)	1.74	(0.62)
Concurrent Illegal Behavior
Age 20 Illegal Behaviors	---		---		---		6.31	(0.47)***	7.05	(0.49)***
Interaction Term
Police Contact x Black Pseudo R2	--- .139		--- .140		--- .153		--- .264		9.55 .290	(0.93)*

Odds ratios reported with standard errors in parentheses

^*p<.05

^**p<.01

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

Net of background variables, individuals who had a police contact by the eighth grade are more likely to experience an arrest in young adulthood. Prior contact increases the odds of experiencing an arrest by 6.07 times. To ensure that we are not simply identifying individuals with a propensity to engage in illegal behavior, we control for early illegal behavior: property crimes, violent crimes, and illegal drug use by the eighth grade. The second model controls for both background variables that might be associated with arrest and early engagement in illegal behaviors. Net of both background variables and eighth-grade illegal behaviors, having a police contact is still a statistically significant predictor of arrest: youth with a police contact have 5.95 greater odds of an arrest in young adulthood than those without a police contact. In model three, we add tenth-grade arrest—a measure of arrest at least two years after the reported police contact. Net of background factors, eighth-grade illegal behaviors, and tenth-grade arrest, having an early police contact is still a statistically significant predictor of arrest. Individuals with a police contact have 5.23 greater odds of experiencing an arrest in young adulthood. None of the other predictors in these models are statistically significant. Model 4 includes concurrent behavior. After controlling for self-reported past-year illegal behavior, the relationship between early police contact and arrest in young adulthood is still statistically significant. In addition, engaging in illegal behavior is, as expected, associated with arrest at age 20. Net of other factors, individuals who have engaged in illegal behavior during the past year have 6.31 times greater odds of experiencing an arrest during the same year compared to those who do not report engaging in illegal behavior. However, the addition of the police contacts by Black interaction term (Model 5) indicates that police contacts have a different relationship with arrest for Black respondents than White respondents. To explore race differences in the relationship between early police contact and arrest at age 20, we examine logistic regression models separately by race.

Race and Adolescent Encounters with Police

Using two sets of regression equations, we estimate the differential impact of police contact on Black and White respondents (See Table 3).

Table 3:

Logistic Regression Predicting Arrest at Age 20 by Race- Examining Delinquency as a Mediator

	Black(N=163)		White(N=168)
Predictor	Model 1	Model 2	Model 1	Model 2
Variable of Interest
8th-grade Police Contact	10.84 (0.65)***	11.10(0.73)**	1.85 (0.82)	1.55 (0.84)
Demographics
Age	1.58 (0.38)**	1.58 (0.40)	1.15(0.29)	1.16(0.38)
Female	0.83(0.53)	1.04(0.58)	0.39 (0.71)	0.40 (0.73)
Log Income 8th-grade Illegal Behavior	0.75 (0.33)	0.476(0.36)	0.83 (0.40)	0.79 (0.43)
Self-reported property crime, violence, and drug use	1.37(0.64)	1.25(0.70)	0.83 (0.76)	0.61 (0.78)
Early Criminal Justice Involvement
10th-grade Arrest	2.17 (0.68)	1.89(0.75)	1.66 (1.18)	1.25 (1.23)
Concurrent Illegal Behavior
Age 20 Illegal Behavior	---	6.85(0.58)***	---	9.72 (1.08)*
Pseudo R2	.279	.431	.037	.106

Odds ratios reported with standard errors in parentheses

^*p<.05

^**p<.01

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

Table 4:

Logistic Regression Predicting Delinquency at Age 20

Predictor	Model 1	Model 2	Model 3	Model 4
Variable of Interest
8th-grade Police Contact	2.17 (.30)*	1.87 (.31)*	7.92 (0.04)*	6.35 (0.06)
Demographics
Age	1.11 (.19)	1.12 (.19)	1.39 (.74)	1.43 (0.72)
Black	0.42 (.30)**	0.42 (.30)**	.09 (.02)*	0.10 (0.02)*
Female	0.75 (.25)	0.75 (.25)	0.31 (0.24)	0.34 (0.27)
Log Income 8th-grade Illegal Behavior	1.04 (.15)	1.09 (.16)	1.36 (0.76)	1.92 (0.52)
8th-grade Property Crime	1.56 (.27)	1.57 (.27)	4.90 (0.11)	5.31 (0.09)
Early Criminal Justice
10th-grade Arrest	---	2.46 (.52)	---	5.70 (0.08)
Interaction Term   Race x Police Contact	---	---	0.68 (0.70)	0.60 (0.61)

Odds ratios reported with standard errors in parentheses

^*p<.05

^**p<.01

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

Model 1 indicates that White youth with a police encounter are no more likely than their non-contacted White peers to experience an arrest in young adulthood after controlling for background characteristics, early illegal behavior, and tenth-grade arrest. However, holding all other factors constant, the odds of Black youth with a police contact being arrested in young adulthood are 10.84 times larger than their non-contacted Black peers. The addition of concurrent behavior in Model 2 fails to explain the relationship between early adolescent police contacts and young adult arrest for Black respondents. Net of all other factors, Black respondents who experience a police contact by the time they reach adolescence have nearly 11 times greater odds of experiencing an arrest when they are young adults than Black respondents without a contact. Concurrent illegal behavior is, not surprisingly, related to young adult arrest for both Black and White individuals. Black individuals who reported having engaged in illegal behavior in the past year (at age 20) have over six times greater odds of being arrested than do their Black peers who do not report illegal behavior. Similarly, White individuals who report illegal behavior have 9.72 times greater odds of being arrested in young adulthood than those who did not report illegal behavior. However, the relationship between eighth-grade police contact and arrest at age 20 is not significant for Whites.

Mediational Relationships: Early Arrest, Tenth-Grade Arrest, and Delinquency

Labeling theory suggests that one of the ways police contacts contribute to future criminal justice involvement is by increasing illegal behaviors, also known as “secondary deviance.” To test the relationship between police contacts and illegal behavior, we use logistic regression using eighth-grade police contacts to predict the mediator variable, age 20 self-reported illegal behaviors. In Model 1, we control for background characteristics, early illegal behaviors. In Model 2 we add tenth-grade arrest. Consistent with prior research, net of all other factors, we find the relationship between police contacts and illegal behavior is statistically significant. Respondents with a police contact have 1.87 times greater odds of reporting engaging in illegal behavior at age 20 than those without a police contact. In Models 3 and 4 we add a race-by-police-contact interaction term. Here, we examine whether the effect of police contacts on later criminal behavior varies by race. There is no statistically significant difference in the effect of police contact on illegal behavior by race.

DISCUSSION

Our findings are consistent with prior research which finds that police contacts lead to future contacts and arrests, and these secondary contacts cannot be fully explained by engagement in illegal activity (Liberman et al. 2014; McAra and McVie 2005). Our findings complicate the assumptions of labeling theory, suggesting that institutional responses to individuals with contact (secondary sanctioning) more than behavioral changes of individuals with contact (secondary deviance) explain future criminal justice system involvement. In our sample, respondents who reported a police contact by eighth grade were significantly more likely to experience an arrest at age 20, after controlling for engagement in illegal behavior. Because our measures of police contact included both arrests and unofficial contacts with police (i.e., police stops that do not lead to arrest), encounters with police and not just official system involvement appear to set in motion events that subsequently lead to arrest.

However, our findings diverge from prior scholarship in one crucial way: only Black respondents in our sample appear vulnerable to the “secondary sanctioning” process. For Black respondents, there appears to be an institutional response to prior contact, while White respondents are not subject to these same responses. Black respondents with police contact by the eighth grade have 11 times greater odds of being arrested in young adulthood. On the other hand, police contact has no identifiable effect on later criminal justice involvement for White respondents. Not only are Black children subject to more stops in their youth than their White peers, but these stops are consequential for future arrests. In the next section, we draw upon the existing literature to help explain our findings. It is important to note that the potential mechanisms that we describe have not been tested in this paper. Rather, we hope that our discussion of potential mechanisms generates avenues for future research.

The Usual, Racialized Suspects

Our findings show that Black individuals with police contact during early adolescence or childhood are more likely than those without contact to experience arrest six years later. Research identifies that Black youth are stopped more frequently than their White counterparts (Crutchfield et al. 2012; Hagan et al. 2005) and these differences cannot be explained by a variety of demographic, behavioral, and contextual factors. One way to interpret these findings is by considering pervasive and enduring cultural scripts within the United States which link Blackness to criminality and drug use (Muhammad 2011; Welch 2007). These scripts, or stereotypes, are at least partially responsible for dictating the spaces that police monitor, the ways that police understand suspiciousness, and determinations about whether to make a stop (Alpert, MacDonald, and Dunham 2005; Beckett et al. 2005; Piliavin and Briar 1964). Prior scholarship suggests that the relationship between being Black and discretionary traffic stops are mediated by prior criminal justice records (Tillyer 2014). Stated another way, the research argues that Black individuals are more likely to be involved in discretionary stops because they are also more likely to have a prior record. The mediation approach assumes that the effect of having a prior record is invariable by race, which we argue is not the case. Treating the relationship between race and prior stops as an interaction has the potential to expose a more nuanced story about how race and criminal justice labels interact to shape future contacts with police—that prior contact might have a different effect on future criminal justice consequences for Black and White young people. Work on stereotype congruence may help us better understand the primary finding within this paper.

Stereotypes are assumptions that distinguishable groups hold specific traits. For example, the stereotype of Black criminality holds that Black individuals are more likely to engage in crime. Stereotype congruence refers to when the behavior of a group member is perceived as consistent with the stereotype. Police, drawing upon common cultural scripts linking Blackness to criminality, may view Black individuals with police contact as stereotype congruent. That is, prior police contact reinforces stereotypes of Black criminality and provides justification for future contact. White individuals, on the other hand, may remain impervious to these same labels. Previous research suggests that White individuals experience a “racial halo effect” that protects them from police suspicion (Weitzer 1999). We suggest that this same type of effect might buffer White youth with police contact from future surveillance and arrest. The image of a “usual suspect” is not race neutral—it is infused with preexisting stereotypes about criminality and race.

Stereotype congruence, however, relies on police having information about prior contacts when determining whether to make a stop or arrest. At the time of data collection in Seattle (2001–2008), the Seattle Police Department used a centralized Records Management System (RMS), which would have allowed police officers in the field to see whether individuals had been previously arrested. However, there was considerable variability when it came to non-official contact. Officers might be able to see if someone was formally associated with a police report. However, police officers during this time were unlikely to use the system to document contacts that did not result in a crime. It was not until the department entered into a Settlement Agreement with the Department of Justice in 2012 that investigatory stops or limited investigative seizures were documented in the RMS and retrievable by officers in the field.⁶ Therefore, the documentation of stops prior to this moment was inconsistent. While this calls into question how much information police might have at the time of the stop, we feel it is important that we not discount potential informal information flows. As Victor Rios (2011) points out in his book, Punished: Policing the Lives of Black and Latino Boys, the informal exchange of information between school resource officers and police outside of schools can lead police officers to initiate contact outside of schools. It is possible that informal networks were activated in a similar way within our sample.

Non-legal Behavior Changes

It is also possible that other potential mechanisms are driving the relationship we found. Non-illegal behavior changes not measured here may attenuate the relationship between early police encounters and later criminal justice involvement. Research has suggested that negative interactions with police officers—particularly those in which individuals feel harassed—can have mental health consequences (Gau and Brunson 2015; Geller et al. 2014; Weitzer and Brunson 2009). Repeated and negative interactions may lead individuals to avoid the police, and this avoidance could have the unintended consequences of piquing police suspicion (Geller et al. 2014; Weitzer and Brunson 2009). Since Black Americans are more likely to report harassment and abuse at the hands of police officers, it is possible that avoidant strategies or other responses to trauma might be more prevalent for Black individuals. Therefore, it would be police officers’ disparate treatment of Black individuals and subsequent coping behaviors of the individuals they contact that set in motion future criminal justice consequences.

Implications for Theory and Policy

Our findings support more recent work which suggests that the “secondary sanctioning” process is critical in understanding the consequences of criminal justice contact. Prior work on labeling theory has focused nearly exclusively on the role of criminal justice contact in future illegal behavior. In doing so, it has overlooked the important process by which institutions respond to individuals with criminal justice contact. Here, we have bridged recent research on labeling theory with a substantial body of work that examines the disparate treatment of Black Americans. We argue that institutional responses to criminal justice contact are not color-blind. As our data indicate, race shapes the frequency of contact and potential responses to contact. These findings have important implications for criminal justice scholarship—pushing the field to address more clearly the complicated relationship between race and contact on a variety of outcomes. While the current paper focuses on the criminal justice system, scholars might also consider how these same processes play out in other institutional settings and contexts. For example, education scholars might consider whether similar processes occur in regard to school disciplinary procedures.

Future research might also examine whether these results hold for other racial/ethnic groups and whether the findings differ by gender or socioeconomic status. Our sample demographics and size restrict us from examining, for example, the consequences of contact for Latinx individuals. Further, we were unable to explore differences by race and gender, or race, gender, and socioeconomic status. Yet we know that race, gender, and class intersect to produce different experiences of policing (Brunson and Miller 2006; Brunson and Weitzer 2008; Chesney-Lind 1989; Morris 2016). We hope that future research will continue to delve into the consequences of police contacts—taking up some of the issues we have not addressed.

Finally, we feel the findings in this article have important policy implications. While some proactive practices have subsided (i.e., stop-and-frisk policing in New York City), many continue. The use of police in schools, for example, has continued to grow in recent years and some argue that police presence within schools protect children from violence. However, police presence in schools also have the potential for increasing encounters between police and youth and may contribute to the type of racial inequalities in later arrest that we have identified within this paper. Pressure from Black Lives Matter movement and associated protests have forced school boards around the country to contend with the often close relationships between schools and police—leading some school districts to dismantle school-police programs. Second, as data systems become more advanced and integrated, there is a risk that access to information about prior contacts or arrests might influence officers’ decisions regarding whether to make contact. Joining with scholars of surveillance, we note that consequences of these systems should be carefully studied. Finally, our findings underscore the deleterious consequence of police contact with Black children—suggesting that even contacts that do not lead to arrest may have implications for racial disparities that appear later in the system. In 2014, Brame and colleagues (2014) reported that by the time Black men had reached the age of 23, nearly half in the United States had experienced an arrest, compared to 38 percent of White men. Official criminal justice involvement is especially troublesome, since it has consequences that extend well into adulthood by negatively affecting employment opportunities, especially for Black men (Pager 2003), decreasing potential earnings (Western, Kling, and Weiman 2001), and leading to the accrual of legal financial obligations (LFOs) and individual debt (Harris 2016; Harris, Evans, and Beckett 2010). Police encounters represent the first step in a complex process that exacerbates racial inequities.