As a result of these and other zero tolerance or “crackdown” laws and policies (Hirschfield, 2018), public schools have become sites of great risk for children—not so much from crime perpetrated by other students, which has been decreasing or flat for decades (depending on the crime), but from the punishments that schools mete out when children transgress (IES, 2016). The pipeline is characterized both by overall scale, creating risk for all students, and disproportionality, creating more risk for some groups of students than others. And research suggests that the pipeline phenomenon has grown worse over time (Losen et al., 2015; R. J. Skiba, Chung, et al., 2014).
A “pipeline” is a useful metaphor for this cluster of school policies. The metaphor evokes a series of pathways and junctions, clarifying the ways in which consequences of individual choice and actions are channeled on the aggregate, like drops in a flow of water. We cannot predict that any given child will end up in prison later in life. Instead, zero-tolerance discipline, overpoliced schools, dysfunctional juvenile justice interventions, and other institutional factors create pathways of probability, where arrival at each new point of punishment increases the likelihood of arriving at the next.
In a developmental approach to student discipline, a moment of institutional response to poor behavior would be designed to decrease likelihood of future behavior, and/or decrease the severity of the behavior over time. The primary purpose of such approaches is changing behavior, not increasing severity of punishment for behavior. Zero-tolerance has given us instead a model where each moment of institutional response makes things more risky for the student—without addressing the root causes of the student’s behavior. Such approaches can directly and indirectly lead to the same or worse behavior in the future, which further increases future risk of punishment. In mathematics this is known as a Markov chain, which predicts the probability of certain outcomes of seemingly random sequences of events. While the pipeline metaphor can be constraining, it nevertheless offers us a powerful policy map for understanding, and reversing, the disastrous relationship between public schools and prisons.
The pipeline has many components, or junctions. The following (not strictly in order) are key:
Junction 1: School Failure. The first junction is, simply, the risk that accrues to kids who fail school (Lochner and Moretti, 2004; Wald and Losen, 2003). In an era of mass incarceration, the risks associated with dropping out of school are scaled up for all youth, but have been disproportionately concentrated on the children who attend poorly resourced and operated schools. Such schools correlate with poverty and racial segregation—the latter of which has been increasing steadily since the mid-1980s (Frankenberg and Orfield, 2012; Orfield, 2001). The risk is so concentrated, for example, that some age cohorts of African American men without high school diplomas face a lifetime cumulative risk of incarceration of up to 70 percent (Pettit and Western, 2004; Western and Wildeman, 2009). Rather than meliorate the problem, the high-stakes testing regime forced on states by the federal No Child Left Behind Act actually incentivizes the push-out of struggling students (The Advancement Project, 2011.) Moreover, there is a link between this junction, academic achievement, and the discipline-related junctions that follow (Gregory, Skiba, and Noguera, 2010). High suspension rates correlate with negative school climate and lower academic performance—not just for the students suspended, but for their peers as well (Skiba, Chung, et al., 2014; Skiba, Arredondo, et al., 2014).
Junction 2: Office Referrals. School discipline practices are a significant junction in the pipeline, and have been restructured since the 1970s to reduce teacher autonomy through systems of “referral” to central administration, and also by the actions of police officers in schools (e.g., Kafka, 2011). Teachers remain instrumental, however, in the initial decision to identify, interpret, and refer student behavior during class. Several studies show that the initial decision to refer a student to the office is a driver of scale and racial disproportionality. Although zero tolerance originated in school safety imperatives regarding school violence and (later) drugs on campus, most out-of-school suspensions have been for minor infractions—most often defiance of school authority (Mendez and Knoff, 2003). In their seminal study The Color of Discipline, Skiba et al. (2002) synthesized and replicated previous findings that black students were more likely than white ones to be referred to the office, but for offenses that were more subjective and less serious (“disrespect, excessive noise, threat, and loitering,” for example). Boys, too, were more likely to be referred than girls for all offenses except truancy, although the difference appeared to be accounted for by differences in boys’ rates of offense. Studies by Mendez and Knoff (2003), Kinsler (2011) and Losen et al. (2015) have similar findings: initial office referrals were the site of serious racial disproportionality. There is also a growing body of scholarship on classrooms as sites of risk for girls of color, and black girls in particular, who are five times more likely to be suspended than white girls—a statistic attributed by many scholars to stereotypes and nonconformity with the norms of white female deference to authority (Epstein, Blake, and Gonzalez, 2012; Onyeka-Crawford, Patrick, and Chaudhry, 2017).
Junction 3: Special Education. Children with disabilities have special legal protections from punishment for behaviors associated with their disability. Nevertheless, researchers have consistently found evidence of disproportional office referral and suspension for students with diagnosed disabilities, including disabilities that are associated with the offensive behavior (Advancement Project, 2005; Losen, 2015). We see this statistic reflected at a later junction point, the juvenile justice system, where students with disabilities are overrepresented, and yet often have disabilities with related social and emotional needs that could be more effectively served within schools or by home-based services (Tulman and Weck, 2009).
Junction 4: School Climate. A school day is full of interactions, formal and informal, that go well beyond the formal interactions during class. Researchers refer to school “climate” as the quality of lived experiences in school—the quality of interpersonal relations, explicit and implicit norms, goals, and structures that shape how people feel about themselves and each other. As school climate becomes more negative, suspension rates increase, putting students at greater risk (Bickel, Qualls, and O’Neill, 1980; Davis and Jordan, 1994; Kinsler, 2011). This rise can be understood in terms of more punitive adults (teachers and administrators), or more transgressive students, or both. Indirectly, negative school climates exact psychological costs from students, who rely on positive adult and peer relationships to develop pro-social dispositions and behaviors, and trust in authority figures and the law (Tyler and Trinkner, 2017; Wilson, 2004).
Junction 5: Suspension and Expulsion. School exclusion due to suspension and expulsion is a powerful driver of the school-to-prison pipeline for all students. Exclusion erodes trust in authorities; it puts students behind academically; and it puts them on track for academic failure, dropout, contact with juvenile justice, and incarceration. Suspensions also increase the likelihood of future suspensions (Brown, 2007; P. J. Hirschfield, 2008; Mendez and Knoff, 2003). In addition to the direct institutional consequences, suspension and expulsion have profound social and emotional consequences for children that can make them more likely to engage in later criminal behavior (P. J. Hirschfield, 2018; P. J. Hirschfield and Celinska, 2011). Rather than provide a developmental mechanism for meliorating negative behavior, school exclusion puts children on a path toward future negative behavior.
Severe consequences are not distributed equally, and this has been a consistent research finding for nearly fifty years. The best demographic indicators for whether a student will be suspended are gender, race, special education status, school attendance, and previous suspensions. For example, despite the fact that black students are referred for less serious and more subjective infractions, they receive more serious punishments than white students, including more suspensions, expulsions, and corporal punishment (R. J. Skiba, Arredondo, et al., 2014; Russell J. Skiba et al., 2002). During the 2011–12 school year, 10 percent of America’s high school students were suspended from school at least once. For black students that rate was 23 percent, while for students with disabilities the rate was 18 percent (Losen et al., 2015, p. 1). Importantly, there is great variation in school discipline rates from school to school, but less variation within any particular school (Kinsler, 2011; Losen et al., 2015).
Junction 6: Contact with Police in School. Between 1975 and 2007, the rate of schools reporting the presence of assigned police officers exploded from 1 percent to 40. Despite this revolution in the relationship between public schooling and criminal justice, there has not been coordination or oversight of this phenomenon, nor is there a research base to support its effectiveness (James and McCallion, 2013; Na and Gottfredson, 2013). Research—both through fine-grained studies of particular schools and districts and through national data sets—suggests that the presence of school resource officers (SROs) puts children at greater risk for committing minor infractions that are reported as crimes, while decreasing teacher and administrator discretion (Kupchik, 2010; Na and Gottfredson, 2013). Without good training and clear agreements between law enforcement agencies and school districts, SROs can also negatively affect school climate, alienate students, faculty, and communities, and escalate conflict (Dycus, 2008; Mukherjee, 2007; Peake, 2015). A study of police behavior in New York City schools with metal detectors, for example, found that 77 percent of reported incidents with students did not involve criminal activity, but instead reflected police overreach into the ordinary disciplinary issues of the schools (Mukherjee, 2007). There is a direct correlation between the percentage of students of color in a given American high school and the likelihood that the school will have sworn police officers placed there (Onyeka-Crawford et al., 2017). As a general matter, Hirschfield (2009) found that being arrested as a ninth or tenth grader greatly increases the likelihood that a child will drop out of school.
Junction 7: Juvenile Justice. In addition to being punished within the school framework, students subjected to zero tolerance are also more rigorously referred to alternative school placements, including into the juvenile justice system (and even, in extreme cases, the criminal justice system). Juvenile justice facilities hold children for temporary detention between arraignment and trial, or for a long-term period of commitment in lieu of prison. In theory, the juvenile justice system exists to provide children with an alternative to the adult criminal justice system. As with other junctions in the pipeline, however, problems of scale, disproportionality, and future risk plague the juvenile justice system. And as with schools, geography plays a role. Feld (1991) found that urban juvenile justice intervention is more formal, bureaucratized, and severe in both pre-trial detention and sentencing than suburban and rural interventions. Many referrals to juvenile justice are often not for serious crimes, and in some cases not for crimes at all. In 2004, for example, 35 percent of juvenile justice referrals were for truancy (Curtis, 2013, p. 1260). For decades, scholars have found racial and disability disproportionality within the juvenile justice system (Gottesman and Schwartz, 2011; Pope, Lovell, and Hsia, 2002). Once referred to the system, non-white youth are more likely to have their case formally processed than white youth (Maume, Toth, and Spears, 2006). While juvenile justice can have positive outcomes for youth (Thomas, 2014), serving time in a residential facility in the system is very risky for future criminal justice contact, and often harmful (Gottesman and Schwartz, 2011).
It is an open question whether zero-tolerance policies, as a package of reforms, have had a significant effect on reducing violent crime. Crime has been declining across society over the same period that zero-tolerance has been scaling up, but serious crime has remained steady. It is certainly the case that attending school has become vastly more risky. The scale of the problem affects students of all races, and both sexes, at unacceptably high rates, but the disproportionate effects on youth of color, and black boys in particular, have been devastating. Whether or not students end up at the terminus of the pipeline—prison—many scholars have noted the significant harm that each junction poses to their healthy social and emotional development (American Psychological Association Zero Tolerance TaskForce, 2008; Council on School Health, 2013; Hagan, 1997).
It should not be surprising that scholars working within the pipeline paradigm point to school-based institutional solutions to disrupt the scale and disproportionality of the pipeline at each junction. To meliorate academic failure, reformers look to teacher training in culturally relevant pedagogy and conflict de-escalation for youth of color, so that children can be appreciated and find opportunities for success, and also so that teachers do not misread and punish behaviors that are less serious than imagined (e.g., Sartain, Allensworth, and Porter, 2015). Critics of dysfunctional, school-based policing call for scaling back officer involvement in the daily operations of schools, and more training and clearer guidelines for SROs (e.g., Dycus, 2008; Mukherjee, 2007). Critics of juvenile justice call for better and broader services and community-centered programming (e.g., Dahlberg, 2008; Gottesman and Schwartz, 2011). And the most consistent call for stopping the school-to-prison pipeline is for restorative justice practices after students engage in bad behavior (e.g., Karp and Breslin, 2001; Varnham, 2005). What each of these changes requires is training—and lots of it—as each of these reforms, in its own way, represents a break from a current institutional practice that is incentivized to push out children who most rely on school as a place to learn.
The History of Zero Tolerance in American Public Schooling by 
Why children follow rules : legal socialization and the development of legitimacy by 