Disproportionate Discipline Referrals for Students of Color & Students With Disabilities

Rethinking Changes in Policy and Restorative Justice Programs and Practices

By Howie Knoff, Ph.D.

For decades, school discipline, classroom management, and student behavior has been a dominant national concern for educators—a concern similarly expressed by students and parents, community organizations and leaders, and state departments of education and the U.S. Department of Education.

And while positive school and classroom climates and students’ effective social, emotional, and behavioral interactions strongly correlate with academic engagement and achievement (and fewer school or classroom discipline problems), many educators still haven’t fully embraced the fact that:

  • Prevention is the key, but multi-tiered services, supports, and interventions must still be available for non-responsive students with social, emotional, and behavioral challenges;
  • Students need to be directly and explicitly taught (like academic skills) the interpersonal, social problem-solving, conflict prevention and resolution, and emotional control, communication, and coping (i.e., social, emotional, and behavioral self-management) skills—from preschool through high school—that they need to be successful; and
  • Schools are not effectively teaching these skills to all students in an integrated, scaffolded way and, instead, many are using policies and unvalidated programs (not scientifically-proven, field-tested practices) to address school discipline and student behavior.

Compounding the school discipline landscape is the fact that, for decades, students of color and/or with disabilities are disproportionately referred to their school principal’s office for disciplinary infractions, and they are similarly disproportionately suspended, referred by and to law enforcement, put in alternative school programs, and expelled. Critically, many of these students’ original behavioral offenses are minor disruptions like “disrespect,” “defiance,” “talking back,” and “refusal to comply”—disruptions that result in student-teacher conferences for White students, but office discipline referrals for students of color and/or with disabilities.

To address this disproportionality, many states and districts have made policy changes, while others have advocated school-wide programs. These “fixes,” however, have often been implemented without field-testing them across different student ages, grade levels, geographic settings, and economic and racial demographics (Gordon, 2018; Knoff, 2018; U.S. Government Accountability Office, 2018; U.S. Commission on Civil Rights, 2019).

Changes in Policy do not Change the Presence of Disproportionality

In the first area of disproportionality “remediation,” many states (e.g., California, Maryland), urban districts (e.g., Minneapolis, Dallas), and individual schools have passed policies whereby students at specific age levels cannot be suspended for minor disruptions. While this has decreased office discipline referrals overall, this has not decreased disproportionality, and it has increased teachers’ dissatisfaction with their district/school discipline systems (Knoff, 2018, 2019a).

More compelling is the fact that these policy changes have inherently discouraged schools from identifying the root causes of the disproportionate office referrals that persist. Hence, many students are not receiving the services, supports, strategies, or interventions they need to eliminate their challenging behavior when it continues. Moreover, teachers and other school personnel are not receiving the training, classroom management consultation, or coaching they need when their disproportionate office referrals of students of color and/or with disabilities are due to implicit bias, lack of multicultural- and disability-sensitive skills, or outright prejudice.

Critically, relative to the students, there are many different reasons why they present with social, emotional, or behaviorally challenges in a school or classroom. To decrease, eliminate, or treat these challenges, different reasons must be connected to the right services, supports, or interventions. Some of the primary reasons include that some students:

  • Have biological, physiological, biochemical, neurological, or other physically- or medically-related conditions or factors that are unknown, undiagnosed, untreated, or unaccounted for.
  • Do not have positive relationships with teachers and/or peers in the school, and/or the school or classroom climate is so negative (or negative for them) that it is toxic.
  • Are either academically frustrated (thus, for example, they emotionally act out, become anxious, or withdraw) or academically unsuccessful (thus, they are behaviorally motivated to escape further failure and frustration).
  • Have not learned how to demonstrate and apply effective interpersonal, social problem-solving, conflict prevention and resolution, and/or emotional control, communication, or coping skills to specific (school-based or home-based) situations in their lives.
  • Do not have the skills or motivation to work with peers—for example, in the cooperative or project-based learning groups that are more prevalent in today’s classrooms.
  • Do not have access to consistent, meaningful incentives (to motivate appropriate behavior) or consequences (to discourage future inappropriate behavior).
  • Are not held accountable for appropriate behavior by, for example, requiring them (a) to apologize for and correct the results of their inappropriate behavior; and (b) role play, practice, or demonstrate the appropriate behavior that they should have done originally.
  • Are responding to inconsistency—across people, settings, situations, or other circumstances. For example, when teachers’ classroom management is inconsistent, some students will manipulate different situations to see how much they can “get away with.” Or, when peers reinforce inappropriate student behavior while the adults are reinforcing appropriate behavior, students often will behave inappropriately because they value the peer attention more than the teacher/adult attention.
  • Are experiencing extenuating, traumatic, or crisis-related circumstances outside of school, and they need emotional support (sometimes including mental health) to cope with these situations and be more successful at school.

Once again, it’s not all about the students. Some inappropriate office referrals occur, for example, when teachers (a) do not have effective classroom management skills; (b) at the same grade or instructional levels do not use consistent classroom management approaches; (c) are demonstrating implicit bias, insensitivity, or outright prejudice; or (d) are not held accountable by their principals or supervisors for appropriate professional and pedagogical behavior.

Restorative Justice Programs and Practices: The Research Just Doesn’t Add Up

In the second area of disproportionality remediation, many districts and schools have adopted an assortment of school-wide discipline programs—often advocated by the U.S. Department of Education and/or Justice (and their federally-funded National Technical Assistance Centers), state departments of education, or as reported in the popular education press. None of them (e.g., PBIS or SEL) has had any significant, objective, sustained, or demonstrable effect on disproportionality. This includes Restorative Justice, which uses inclusive, equity-responsive messaging to encourage its adoption, even as it has also engendered a cottage industry of “knock-off” companies, consultants, and capitalizing entrepreneurs.

Unfortunately, many districts have invested significant amounts of time, staff, resources, and money to implement Restorative Justice programs, and yet they often do not review their research and sustained outcomes and, hence, their (low) probability of long-term success. And well-intended in their desire to decrease disproportionate discipline referrals and actions, the implementation of ineffective approaches simply wastes resources, delays more effective strategies, and gives staff the impression that the task at-hand may be out-of-reach.

Below are three areas that summarize the current Restorative Justice “state of affairs.” The “spoiler alert” is that Restorative Justice programs have not objectively and empirically proven their utility, and they should not be considered—much less adopted—by districts and schools across the country.

There is No Consensus on What a Restorative Justice Program Is or Consists Of

The WestEd Justice & Prevention Research Center has published an ongoing comprehensive review of the Restorative Justice research. In the Overview of its most-recent version, its authors (Fronius, et al., 2019) state:

Restorative justice (RJ) is a broad term that encompasses a growing social movement to institutionalize non-punitive, relationship-centered approaches for avoiding and addressing harm, responding to violations of legal and human rights, and collaboratively solving problems. RJ has been used extensively both as a means to divert people from traditional justice systems and as a program for convicted offenders already supervised by the adult or juvenile justice system.

In the school setting, RJ often serves as an alternative to traditional discipline, particularly exclusionary disciplinary actions such as suspension or expulsion. RJ proponents often turn to restorative practices out of concern that exclusionary disciplinary actions may be associated with harmful consequences for children (e.g., Losen, 2014). More recently, it has also been embraced as a preventative intervention for building an interconnected school community and healthy school climate in which punishable transgressions are less common (e.g., Brown, 2017).

Within school settings, RJ encompasses many different program types. An RJ program can involve the whole school, including universal training of staff and students in RJ principles, or it can be used as an add-on to existing discipline approaches and philosophies. It also has been combined with other non-punitive discipline approaches, such as Social and Emotional Learning and Positive Behavioral Interventions and Supports.

Given such mixed implementation approaches, it is not easy to define exactly what constitutes RJ in schools. Sellman, Cremin, and McCluskey (2014) argue that from “a theoretical perspective, RJ is essentially a contested concept” and “it is unlikely that there will ever be one agreed definition.

This concern is echoed by a commissioner who served on the U.S. Commission on Civil Rights which published a significant report on July 27, 2019, Beyond Suspensions: Examining School Discipline Policies and Connections to the School-to-Prison Pipeline for Students of Color and with Disabilities. In this report, Commissioner and lawyer Peter N. Krisanow stated:

Restorative practices, as they are typically called in a school or community setting, include many specific program types and do not have one specific definition in the literature; they are broadly seen as a nonpunitive approach to handling conflict (Fronius et al., 2016). When we talk about ‘restorative justice,’ as we see here, we are not talking about a clearly defined set of practices. We are talking about fuzzy muffles. Telling children to reflect upon the harm they have caused to others may be effective for children who are predisposed to empathize with others or to care about disappointing their teachers. But not all children care about the effects of their actions on others, or indeed, may be pleased that the harm they caused had its intended effect.

The overall implication here is that, even if an effective Restorative Justice program existed (and it does not—see below), schools would still need to know exactly how the program they chose defined its terms, established its goals, evaluated its outcomes, and impacted a specific well-described population of students. But this is not the case right now, so schools need to seriously question whether any Restorative Justice Program is worth any time or consideration.

Research Does Not Definitively Support Any Restorative Justice Program

As noted above, there is virtually no objective, methodologically-sound, comparative research that has demonstrated the sustained and causal impact of a Restorative Justice program (or collective restorative practices) on changing student behavior or decreasing disproportionate office referrals for students of color and/or with disabilities. This is the clear conclusion from the WestEd Justice & Prevention Research Center’s comprehensive review of the Restorative Justice research through 2018 (Fronius, et al., 2019).

Indeed, after a second review of hundreds of articles, chapters, theses, and dissertations published from 2014 through 2018 (extending their initial review of articles that dated back to 1999—published in 2016), they found only the 31 articles that used any kind of quantitative approach to evaluate the K-12 implementation of Restorative Justice in U.S. schools.

After objectively analyzing the methodological integrity of these 31 articles, they concluded:

  • Despite the popularity of RJ in the United States, most programs are still at the infancy stage (Guckenburg et al., 2015). As such, there are a limited number of evaluations and other studies. One trend in the available literature is that RJ qualitative reviews and descriptive reports are much more prevalent than RJ evaluation studies.
  • Many of the studies we located are descriptive or use a pre/post (before/after) evaluation design. Critically, many of the reports attempt to control for student-level and school-level factors by using multivariate regression, and others use time-series modeling to attempt to isolate the impact of the introduction of RJ on students and schools. Nonetheless, these designs lack a control (comparison) group and thus may suffer from a range of statistical biases that render them a poor fit for ascribing any observed changes to RJ specifically (e.g., Weisburd, Petrosino, & Fronius, 2014). This limitation does not mean that these studies have no value. The promising results reported across these studies help contextualize and echo the findings of the single, published, rigorous experimental test of RJ (Augustine et al., 2018) and continue to serve as foundational groundwork of other rigorous studies currently underway.
  • A . . . critical limitation (of the research reviewed) is that the internal validity of these studies is generally low. Much of this research would not meet the standards of evidence for evidence-based registries in education or justice (e.g., the U.S. Department of Education’s What Works Clearinghouse, the U.S. Department of Justice’s Crime Solutions). The methods employed in many studies make offering any conclusive recommendations a challenge. For example, the most common evaluation design reported in the literature is based on pre- and post-tests. By nature, such pre/post designs only study those individuals exposed to the program (i.e., a single-group design) with no counterfactual (control or comparison condition), so the studies are considered low in internal validity (Weisburd et al., 2014).

As noted above, the WestEd review found only one “published, rigorous experimental test of Restorative Justice” (i.e., Augustine et al., 2018), and it characterized the results of this study as “promising.” However, a “deep reading” of the actual Augustine et al. (2018) report—published by the Rand Corporation—suggests some troubling Restorative Justice outcomes at the molecular level.

The Augustine et al. (2018) study was conducted and published as an independent evaluation of a Restorative Justice program implemented by a separate organization. More specifically, the Rand Corporation study evaluated the efficacy of the International Institute for Restorative Practices’ SaferSanerSchools Whole-School Change program in 22 randomly-selected Pittsburgh Public School District (PA) schools as compared with 22 randomly-selected non-participating Control schools during the 2015–16 and 2016–17 school years.

Based on this highly sophisticated randomized controlled study, the general results found that, in the Restorative Practice schools: (a) alternative school placements decreased; (b) students were less likely to be suspended multiple times; (c) disparities in suspension rates between African-American (vs. Caucasian), and low-income (vs. higher-income) students, respectively, decreased; and (d) suspension rates for female students declined.

But the “deeper,” more troubling results revealed that:

  • While suspension rates in the Restorative Practices schools declined by 36% during the two-year study, suspension rates in the Control schools also declined 18% during the same time period.
  • Fewer suspensions were not found overall for male students or students with disabilities.
  • There were no reductions in student arrests, or for incidents of violence or weapons violations.
  • The overall suspension results were driven by lower rates in the Restorative Practices elementary schools.
  • That is, fewer suspensions were not found in the Restorative Practices Middle schools (Grades 6 to 8). Moreover, in the Restorative Practices Middle schools, academic outcomes actually worsened when compared with the Control schools.
  • Finally, survey results from staff in the Restorative Practices schools indicated that they did not think the IIRP program was affecting student behavior.  They did, however, report that their relationships with students had improved because of program involvement.

Thus, while the generalized results of this well-designed study look “promising,” the deeper results are troubling and, perhaps, more telling. And these troubling results are compounded by additional concerns, voiced in the Rand Corporation Report, regarding the (low) probability that the 22 Restorative Justice schools would be able to sustain their activities given the loss of the outside grant that paid for (a) the professional development time needed to implement the program, and (b) the direct coaching from the program developers at the International Institute for Restorative Practices.

Critically, even if the Rand Corporation Report provided a glowing recommendation of the Restorative Practices program in the Pittsburgh Schools, it would still only be one study. But the analysis above also provides a cautionary tale. It suggests that educators should not act and implement comprehensive, school-wide programs based on the general results often reported in the popular education press. Instead, they need to read, in this case, the original Restorative Justice program research themselves, as there may be deeper, troubling results beneath the more positive, general results being publicized and touted (Knoff, 2018, 2019a, 2019b).

Another example of this caution is evident in the 2019 U.S. Commission on Civil Rights referenced earlier. Here, the Commission Report discussed a situation in the Washington, D.C. Public Schools where the Washington Post reported that some previously-published “successes” attributed to that district’s restorative justice practices were grossly overstated.

More specifically, the Washington Post investigation found that:

at least seven of the city’s 18 high schools had removed disruptive students from schools, but not recorded it as a suspension. In several cases, students who had been barred from entering the school were marked as present, others were marked as attending an “in-school activity,” or absent without an excuse. Dunbar High School had the most underreported suspensions compared to any other high school in January 2017, according to Washington Post investigators. In data obtained by reporters, only 7 percent of the days that students were kept out of school for misbehavior were actually correctly reported as suspensions.

In summary, the collective research does not support the implementation of any Restorative Justice program in any school across the country. Moreover, educators need to conduct objective, in-depth, and empirically-based analyses of any school-wide program they are considering, relative to decreasing disproportionate disciplinary referrals of students of color and/or with disabilities, before proceeding into actual implementation.

Behavioral Science and Common Sense Show the Limitations of Restorative Practices

Restorative Justice or Practice programs rarely discuss, based on behavioral science, the kinds of inappropriate behaviors they are best-equipped to address (and not address), and the scientific (not philosophical or altruistic) reasons why they believe they can be effective.

Without over-simplifying, a foundational premise of Restorative Practices is that—when students commit a rude, crude, or anti-social act—they should be given the opportunity to restore, repair, remediate, re-pay, and/or apologize for their transgression with the offended party(ies) and in the settings involved. In the context of disproportionality, it is further encouraged that these restorative acts take the place of an administrative or disciplinary response such as a detention or suspension.

Embedded in this premise is the expectation (hope) that, from a behavioral science perspective, a restorative practice will act consequentially, interpersonally, and motivationally to change a student’s inappropriate behavior in the future.

The consequential premise is that the offending student views the (required) act of restoring, repairing, remediating, repaying, and/or apologizing for their transgression as so significant (aversive) that they never want to be in that position again (and, hence, will not re-commit the offending act).

The interpersonal premise rests with the belief that, if the offending student interacts personally to apologize to the offended party or victim, that that interaction will establish a (positive) relationship and level of both affiliation and accountability such that no future offenses will occur.

And, the motivational premise integrates the two premises above—expecting that the combination of (a) negative or aversive consequences for inappropriate acts, and/or (b) positive incentives or outcomes after developing personal, empathetic relationships with people will result, once again, in future prosocial student interactions.

But there is an important footnote in this scientific thinking; a contextual limitation that is embedded in the entire conceptualization of restorative practices.

If an offending student has not learned and mastered the interpersonal, social problem-solving, conflict prevention and resolution, or emotional control, communication, or coping skills and abilities that underlie his or her inappropriate behavior, then the restorative practice and experience will not result in changed behavior.

That is, regardless of a student’s remediation or apology, or even his or her motivation to change, the same offense is likely to reoccur in the future, because the student does not have the social, emotional, or behavioral ability to respond differently.

Said a different way: You can’t motivate a student out of a skill deficit.

This same scientific principle occurs when students have not learned specific academic skills. For example, if a student has not mastered long division and gets a 40% on her Long Division Unit Exam. . . the motivation to do better on the next test (due to the low grade) will be inconsequential unless she learns and masters the long division skills that were the root cause of her 40% grade.

Thus, motivation will not change her next test score; instruction and learning will.

Similarly, if students are committing anti-social acts because they have not learned and mastered, for example, self-control and communication skills. . . a restorative practice will not change their skill deficits. Change will only occur with instruction, learning, and mastery.

Once again, Restorative Justice and Practice programs simply do not account for this basic behavioral science. Moreover, they oversimplify the complexities embedded in student motivation.

Specifically, most students “live” in a “motivational triangle”—between themselves, their dominant peer group, and their teachers and other adults—when they are in school. If there are different “motivational pulls” relative to an individual student’s behavior, then she or he will demonstrate the behavior that is most strongly reinforced or consequated. That is, even when some students have prosocial skills, if the positive attention or negative rejection of their peer group (clique, or gang) more powerfully motivates inappropriate behavior over a teacher or administrator’s incentives or consequences focused on motivating appropriate behavior, then the inappropriate behavior will most likely occur.

In a restorative practice context: Students will commit inappropriate behaviors—knowing that an adult-directed restorative practice apology, consequence, and/or remediation will follow—when the adoration and potential rejection of their peer group is more compelling (think “Mean Girls”). Moreover, their inappropriate acts will likely continue until something in the dynamics within their triangle, between their peer group and the adults, changes.

Significantly, this motivational situation—involving peer group dynamics, especially in an urban setting—may be one reason why the Rand Corporation study found (a) fewer suspensions in the Restorative Practice middle schools; and (b) no reductions in student arrests, or for incidents of violence or weapons violations.

As noted by Commissioner Kirsanow relative to the Restorative Practice initiative in the Pittsburgh Public Schools (as well as similar programs in the Baltimore and Minneapolis City School Districts):

The best that can be said for ‘restorative practices’ and reducing suspensions is that in some school districts, students who would otherwise have been suspended are in school for more days. This is a paltry return for the price of increased classroom violence and disrupting the education of students who are there to learn.


This discussion focused on the ongoing reality that students of color and/or with disabilities continue to be disproportionately referred to the principal’s office for discipline—many times for minor offenses that are culturally- or disability-related, and that these referrals often result in disproportionate administrative actions related to suspensions, referrals to law enforcement, placements in alternative school programs, and expulsions. This disproportionality has not been impacted by changes in state or district policy, or any number of school-wide change programs or efforts.

We also outlined the primary reasons for students’ inappropriate behavior, while emphasizing how teachers and administrators often are enmeshed in these behaviors, as well as how they perceive and respond to these behaviors. Here, we noted the importance of determining the root causes of students’ inappropriate behavior so that these underlying reasons can be linked to high probability of success services, supports, or interventions that, as needed, change the behavior, its triggers, or its dynamics.

Finally, we addressed the too-common use of frameworks or programs that have not been objectively and causally proven to impact student behavior or disproportionality in a sustained and time- and cost-effective way using methodologically-sound research studies with matched comparison or control groups. Here, we specifically focused on Restorative Justice and Practice programs, reviewing the current research-to-scientific-practice, and demonstrating the importance of deeply reading the original research so that the complexity of the results is understood.

Relative to Restorative Justice and Practice programs, there is virtually no research that would warrant their consideration by any district or school. And yet, unfortunately, many districts have invested significant amounts of time, staff, resources, and money implementing Restorative Justice programs resulting in a waste of resources, delays in more effective strategies, and a host of unsuccessful and—sometimes—unintended, negative outcomes.

Summarizing all of this—with additional discipline data from New York City, Seattle, Philadelphia, and Las Vegas that show increases in school discipline problems and continued disproportionality—Eden (2019) has noted that, “Policymakers have tied teachers’ hands and undermined their authority in the classroom by mandating ‘restorative justice’ practices.”

But he also summarized the effects of “school discipline reforms” like restorative justice practices on students, staff, and schools saying:

School and system leaders should, of course, take their bearings not only from academic studies but also from the perspective of teachers. Nationwide, a majority of educators express sympathy for the idea of discipline reform. But, as I’ve documented, teachers in school districts that implemented discipline reform under pressure from federal investigations do not believe it works.

In Denver, only 23 percent of teachers say the new approach to discipline improves behavior. In Charleston, South Carolina, just 14 percent of teachers believe it is an improvement over previous discipline policies, and in Madison, Wisconsin, only 13 percent of teachers think that it has a positive effect on behavior. In Oklahoma City, 11 percent of teachers said a greater implementation of “positive behavior interventions and supports” would help them be effective, compared with two-thirds who said greater implementation of traditional discipline would help.

School discipline and the existence of disproportionality represent complex social, psychological, and educational issues. We are not going to solve these challenges through simplistic policy fixes and by using unproven frameworks or programs—no matter who is advocating or promoting them.

Instead, we need to follow the science. But we also need to recognize that there is no one answer, no single solution, and no magical one-size-fits-all program.



Augustine, C. H., Engberg, J., Grimm, G. E., Lee, E., Wang, E. L., Christianson, K., & Joseph, A. A. (2018, December 27). Can Restorative Practices improve school climate and curb suspensions? An evaluation of the impact of Restorative Practices in a mid-sized urban school district. Santa Monica, CA: Rand Corporation.

Eden, M. (2019, June 10). Studies and teachers nationwide say school discipline reform is harming students’ academic achievement and safety, the74million. Retrieved from:

Fronius, T., Darling-Hammond, S, Sutherland, H., Guckenburg, S., Hurley, N., & Petrosino, A. (2019, March). Restorative Justice in U.S. schools: An updated research review. San Francisco, CA: WestEd Justice & Prevention Research Center.

Gordon, N. (2018, January 18). Disproportionality in student discipline: Connecting policy to research. Brookings Institute. Retrieved from:

Knoff, H. M. (2019a, December 21).  Schools’ pursuit of effective school discipline, classroom management, and student self-management strategies: Prevention, disproportionality, trauma, and seclusions and restraints. Project ACHIEVE Educational Solutions. Retrieved from:

Knoff, H. M. (2019b, January 26).  New Rand Corporation study finds Restorative Practices produce mixed and underwhelming results:  But some publications are “spinning” the outcomes and twisting these results. Project ACHIEVE Educational Solutions. Retrieved from:

Knoff, H. M. (2018, April 15).  New federal government report finds that disproportionate school discipline actions persist with black, male, and special education students: Manipulating policy, buying programs, and following federally-funded technical assistance centers do not work. Project ACHIEVE Educational Solutions. Retrieved from:

United States Government Accountability Office. (2018, March). K-12 Education: Discipline disparities for black students, boys, and students with disabilities. Washington, D.C.: Author.

United States Commission on Civil Rights. (2019, July). Beyond suspensions: Examining school discipline policies and connections to the school-to-prison pipeline for students of color with disabilities: Briefing Report. Washington, D.C.: Author.

Howie Knoff, PhD, NCSP, is an international consultant, speaker, and author specializing in school improvement and strategic planning, social-emotional learning and social skills training, multi-tiered systems of support, and interventions with behaviorally challenging students. He is a practitioner who has also been a university professor (22 years), and State Department of Education federal grant director (13 years). The author of 24 books and 100+ articles/book chapters, he was the 21st President of the National Association of School Psychologists. Learn more at

Share this page with your friends and colleagues:


Interested in writing for us?

Send us your insight so we can share it! We welcome article submissions from educators, advocates, thought leaders and companies who are working to make education more equitable, accessible and inclusive.
Let's keep in touch! Click to join our email list & subscribe to our newsletter.

Skip to content