Don’t Walk Back Needed Discipline Reform

In 2014, the Obama administration’s Departments of Education and Justice took an important step to respond to the excessive and racially disparate removal of students from schools across America. They jointly issued guidance to schools in the form of a “Dear Colleague” letter, describing “how to administer school discipline without discriminating on the basis of race, color or national origin.” It included a section on how to identify and avoid policies that may have an unjustifiable disparate impact.

The disparate impact guidance prompted districts to review the justification for policies that produced sharp racial differences. This was an important step, given the immediate harm from the loss of instruction and research that had established potentially long-lasting effects on students’ academic success and well being. Most importantly, the guidance reminded schools and districts that, pursuant to the Civil Rights Act of 1964, (and the corresponding disparate impact regulations established in that era), they have an obligation to eliminate unjustifiable policies associated with excessive and disparate discipline.

The letter makes clear that frequent suspensions for minor behaviors (not fighting) are most likely to trigger concerns, stating:

Examples of policies that can raise disparate impact concerns include policies that impose mandatory suspension. . . upon any student who commits a specified offense—such as being tardy to class, being in possession of a cellular phone, being found insubordinate, acting out, or not wearing the proper school uniform; corporal punishment policies that allow schools to physically punish students; . . . . suspensions or expulsions for truancy also raise concerns because a school would likely have difficulty demonstrating that excluding a student from attending school in response to the student’s efforts to avoid school was necessary to meet an important educational goal.

Long-Term Harmful Effects

Eliminating harm to children from unjustified discipline policies is at the heart of the guidance. While discipline measures like suspensions are intended to enforce school rules and improve student behavior, the negative outcomes of such policies should give us pause. Perhaps the strongest findings on the harm from suspension come from a rigorous 2018 study published in the peer-reviewed journal Youth & Society. The study’s author, Janet Rosenbaum, compared otherwise similar suspended and non-suspended children and found that, after 12 years had passed, students who were suspended were less likely to have graduated from high school or college and more likely to have been arrested or on probation. The study controlled for 60 variables that might have otherwise explained the subsequent negative outcomes, such as delinquency or socioeconomic status. Its findings bolster the research-based consensus that suspensions predict lower achievement, higher drop out rates, and greater risk of involvement in the juvenile justice system.

It is worth noting that oft-cited studies that document harm to peers from exposure to disruptive youth were conducted in high-suspending schools and districts, not in schools or districts that had reduced suspensions in response to reform efforts. Even conservative writer Max Eden, in his widely criticized and polemical study of New York City’s reform efforts, admitted finding zero correlation between suspension reduction and degradation of school climate at the school level, as well as no degradation when former Mayor Michael Bloomberg’s reforms reduced suspension lengths and eliminated out-of-school suspension for first-time minor offenses. Similarly, when Philadelphia initiated reforms, negative peer effects only occurred in the schools that continued to suspend students for minor offenses. Meanwhile, there was no harm to peers associated with full reform implementation.

Unfortunately, in his contribution to this forum, Michael Petrilli cites studies that distort the harms while overlooking positive findings about alternatives to suspension. Among the overlooked is a randomized controlled study of a teacher-training program found that a yearlong training designed to improve student achievement through greater teacher engagement of students caused a reduction in exclusionary discipline and near-elimination of racial disparities in disciplinary office referrals.

Other studies that controlled for prior behavior and poverty have demonstrated that schools run by principals who were committed to using suspensions sparingly had lower suspensions, lower racial disparity, and higher achievement. Schools run by principals who believed a more punitive approach was necessary for an orderly environment show the opposite patterns. These are just some of the reasons that reform proponents argue that our schools cannot ignore the school discipline gap if they aim to close our nation’s racial achievement gap.

Tremendous Loss of Instruction

Opponents of discipline reform often ignore the magnitude of the problem and its impact on educational opportunity. When my organization, the Center for Civil Rights Remedies, studied the distribution of high and low suspension rates for every district in the nation, we found high schools in thousands of districts that suspended 25 percent or more of the children of color. At Oklahoma City High School, for example, school leaders suspended 65 percent of all black students and 33 percent of all white students. However, we found that most high schools had much lower suspension rates for all subgroups, indicating that less-discriminatory and reasonable discipline policies are common.

A wealth of research demonstrates that in most high-suspending schools and districts, the majority of the offenses and largest share of racial disparities are punishments for minor nonviolent violations of school codes of conduct, not unlawful or dangerous behavior. Many lower-suspending schools never added harsh policies such as suspending kids for being tardy. Together with the aforementioned research, these data tell us that policies and practices often drive the harmful disparate impact of suspensions on educational opportunity.

What is truly alarming are the racially disparate impacts on instructional loss caused by discriminatory discipline policies and practices. Students with disabilities are more likely than their peers to be suspended overall; but again, we find major differences in suspension rates by students’ race. Recently, I shared distressing new findings from a soon-to-be-published Civil Rights Project report on federal school-discipline data, which show the impact on black students with disabilities, who lose the most instructional time due to disciplinary removal of any student group (see Figure 1). Based on federal data from 2015-16 (two years after the guidance was disseminated), the Civil Rights Project estimates conservatively that due to disciplinary removals, black students with disabilities lost 113.3 days of instruction per every 100 black students with disabilities enrolled nationwide. By comparison, white students with disabilities lost an estimated 39.5 days of instruction per 100 students. Many states had much larger differences. Even more pronounced differences in lost instruction due to discipline are found at the district and school levels.

Claims of Chaos Contradicted

In a series of essays, Michael Petrilli has raised concerns that applying disparate impact theory to school-discipline reform is an unsound overreach by the federal government and will cause chaos in the classroom. His belief that administrators across the country are so afraid of suspending students of color that they have resorted to quotas and made schools dangerous is totally unsubstantiated. No districts have been found to have adopted disciplinary racial quotas in response to the guidance. National school-discipline data from 2013-14 show a slight decline in suspension rates when compared to 2011-12, but the rates today are far higher than they were in the 1970s and 1980s. There is no evidence that a slight decline of a percentage point or two has caused safety problems.

In fact, students attended safer schools in 2015 than in 2013, federal data show. The Institute of Educational Sciences tracked more than 20 indicators of school climate and safety and by nearly every measure, safety and order either improved or stayed the same. Several rates of dangerous or troublesome behaviors declined, including bullying; fighting; victimization of students; students reporting being injured or threatened; and student reports of disorder or the presence of gangs.

Thankfully, there is now a bi-partisan understanding that struggling students typically need more attention and support from adults in school, not less. Both Republicans and Democrats share concerns about unnecessary loss of instructional time, and many policymakers realize that a high percentage of suspended students have experienced trauma or are dealing with mental health problems. For these students, suspension is a non-intervention, because when students are suspended, their ties to a wide range of school-based supports and services are also cut. This is why in Texas, the Republican-controlled state legislature has prohibited suspensions for minor behaviors for all children in grades K-2, for example. And in Ohio, Republican governor John Kasich has agreed to prohibit suspensions for truancy, and a state senator has introduced a proposal similar to the Texas ban on suspending young children.

Ironically, in his recent writings, Petrilli argues that the disparate impact guidance would be wrongfully applied to limit suspensions for minor violations like tardiness. He also suggests that the disparate impact guidance would limit responses to fighting even though the examples in the guidance are policies calling for suspending students for minor non-violent behaviors. Moreover, Petrilli erroneously asserts that disparate impact theory assumes that there is an intent to discriminate from data disparities, alone. He argues that suspending black students more often for truancy is not discriminatory if the disparity was the result in behavioral differences, and not intentional different treatment, and supplies evidence that black students self-report being truant nearly twice as often as white students.

However, the guidance clarifies that disparate impact analysis makes no assumption about the intent to discriminate. Petrilli’s misunderstanding of the law likely explains why his analysis overlooks the central question that disparate impact analysis seeks to answer. It is not whether bias caused educators to label black students as tardy more often, but whether it is educationally sound and justifiable to suspend children for tardiness. An even clearer example is the use of corporal punishment. Schools are allowed to hit children in 19 states. Given that doing so is regarded as child abuse in most states, there are clearly better alternatives. Where legal, the use of corporal punishment is often decided at the school level. The disparate impact guidance doesn’t ban corporal punishment but warns school officials that if they hit some groups of children much more than others, their policy might violate civil rights law.

Critical Protection From Bad Policies and Practices

To keep the distinction between different treatment and disparate impact clear, the guidance has two flowcharts with a series of questions. The disparate impact flowchart asks first whether a policy has an adverse impact. If yes, a reviewer should proceed to the second question, which reads “Is the discipline policy necessary to meet an important educational goal?” If the answer is no, the policy is discriminatory. If it is yes, the flowchart points to a third question, “Are there comparably effective alternative policies that would meet the stated educational goal” with less of an adverse impact on the group most harmed under the current policy?

In other words, only unjustified policies that produce disparities can be deemed discriminatory pursuant to the disparate impact regulations. The core question for disparate impact is whether the policy is justifiable in light of its harmful impact. In contrast, only the different treatment flowchart asks whether similar students were treated differently. By providing the step-by-step analysis regarding how a policy might be deemed discriminatory, the guidance hinders administrators and advocates alike from jumping to conclusions that intentional discrimination is at play based on data disparities alone.

Here’s an example of how the disparate impact analysis might actually play out. Imagine a district where school principals have autonomy over discipline for minor behaviors. A review reveals that in schools with majority black populations, principals adopted suspensions for truancy and tardiness as a matter of policy, while principals at schools with fewer black students do not suspend for any attendance issues as a matter of policy. A review establishes that the higher suspension rates for black students in the district were primarily due to the harsher attendance rules at the schools with majority black enrollment. At the other schools, the review finds that instead of suspension, leaders respond to tardiness and truancy by investing in parental outreach, and counting attendance toward grades and for sports eligibility. They also appear to have reduced truancy and tardiness.

Under this scenario, the district’s inconsistent responses to truancy and tardiness could be challenged on the basis of their racially disparate impact. Hopefully, prompting from the guidance to look at the variation in disciplinary responses by schools within the district would encourage the local school board to voluntarily eliminate suspensions for attendance violations and adopt (district-wide) those responses already working well in its less racially isolated schools.

Ideally, the disparate impact guidance prompts districts to eliminate unsound or unjustifiable discipline policies. Ultimately, the disparate impact guidance both protects certain groups from harmful policies that impact them more than others, and inspires systemic changes that benefit all children.

Trumped-up anecdotes and flimsy research should not drive the evaluation of this important guidance. Therefore, I urge the small group of vocal conservatives sounding the alarms to revisit the hard evidence and to stop calling on U.S. Secretary of Education Betsy DeVos to strip schoolchildren of their civil rights protections.

Dan Losen is director of the Center for Civil Rights Remedies, an initiative at the Civil Rights Project/Proyecto Derechos Civiles at the University of California-Los Angeles.

This is part of a forum on school discipline reform. For an alternate take, see “A Supposed Discipline Fix Threatens School Cultures,” by Michael Petrilli.

Last Updated


Notify Me When Education Next

Posts a Big Story

Business + Editorial Office

Program on Education Policy and Governance
Harvard Kennedy School
79 JFK Street, Cambridge, MA 02138
Phone (617) 496-5488
Fax (617) 496-4428

For subscription service to the printed journal
Phone (617) 496-5488

Copyright © 2024 President & Fellows of Harvard College