Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Schoolwide Positive Behavior Support and Disciplinary Equity

Abstract. Schoolwide Positive Behavior Support and Disciplinary Equity examines how prevention-oriented behavioral support systems interact with institutional decisions that distribute referrals, exclusions, and educational opportunities. Drawing on foundational PBIS theory, school-randomized effectiveness trials, research on racialized office discipline referrals, an equity-focused randomized intervention, and systematic reviews and observational multilevel research through 2026, the monograph distinguishes overall behavioral effectiveness from equitable disciplinary outcomes. It analyzes functional behavioral assessment, tiered prevention, adult decision thresholds, cultural responsiveness, disability-related access, implementation fidelity, measurement validity, causal identification, and the limits of existing evidence. Across twenty-six scientific chapters, the synthesis finds credible improvements in some behavioral and organizational outcomes, mixed evidence for exclusionary discipline reduction, and a narrower causal research base for the elimination of disciplinary disparities. The concluding research agenda prioritizes preregistered equity estimands, independent school-level replications, reliable measurement of instructional time lost, and sustained accountability for the quality and fairness of institutional decisions.

Keywords: schoolwide positive behavior support; positive behavioral interventions and supports; PBIS; SWPBIS; disciplinary equity; school discipline; office discipline referrals; suspension; racial disproportionality; implicit bias; functional behavioral assessment; Tiered Fidelity Inventory; multi-tiered systems of support; student disabilities; inclusive education

Schoolwide Positive Behavior Support as a Scientific Object

Schoolwide positive behavior support (SWPBS), frequently termed schoolwide positive behavioral interventions and supports (SWPBIS or schoolwide PBIS), is an organizational framework for designing, implementing, and evaluating prevention-oriented behavioral supports across a school. Its unit of intervention is the interaction among institutional routines, adult practices, student opportunities to learn behavioral expectations, and the consequences that follow behavior. Schools establish a limited number of positively expressed expectations, teach their meaning in concrete settings, acknowledge their enactment, respond consistently and instructionally to behavioral difficulties, and use repeated measurements to adjust the system. This is a framework through which schools select and organize empirically supported practices rather than a single manualized treatment. The central empirical questions concern implementation fidelity, changes in adult and student behavior, educational access, and the distribution of institutional responses across students and settings. Sugai & Horner (2002) Horner et al. (2009)

The second term in this dossier, disciplinary equity, describes the just and substantively fair distribution of disciplinary opportunities, decisions, consequences, and access to education. A rigorous evaluation asks whether behavior is interpreted through comparable standards, whether responses are proportionate and instructionally appropriate, whether students can contest inaccurate judgments, and whether outcomes differ systematically by race, disability, language, gender, or intersections among these categories. Equity therefore has procedural, distributive, and participatory dimensions. A school's overall decline in office discipline referrals may coexist with persistent or even widening differences in group-specific referral rates. The research object comprises both the prevention system and the decision system that assigns disciplinary consequences. Skiba et al. (2011) Weeks & Sullivan (2024)

An office discipline referral (ODR) is a recorded action taken by school staff in response to behavior classified as a rule violation. It is an institutional event: some behavior occurs or is perceived, an adult interprets it, a reporting threshold is crossed, and a record is created. The observation can be influenced by supervision arrangements, policy definitions, educator expectations, relationship quality, and documentation practices. Suspension and expulsion are distinct forms of exclusionary discipline that can remove a student from instruction and school participation. Attendance, engagement, student safety, teacher workload, and perceived fairness are further outcomes that provide complementary views of school functioning. Treating these constructs separately prevents a decline in one administrative indicator from being presented as comprehensive proof of behavioral change. Girvan et al. (2017) Gage et al. (2018)

This monograph is an integrative, critical research synthesis of foundational behavior-support theory, school-level effectiveness trials, equity-focused interventions, measurement studies, systematic reviews, and institutional guidance through November 2026. It is organized by scientific problems rather than by administrative implementation steps. The evidence base includes randomized school trials, observational records, implementation studies, and reviews, each interpreted according to its design. Particular attention is given to distinct primary outcomes, the clustering of students within schools, reuse of participants across publications, and the fact that interventions intended to improve behavior and interventions intended to reduce discriminatory decision-making can differ in mechanism. The governing question is which institutional arrangements causally improve both behavioral support and fair access to education. Bradshaw et al. (2010) McIntosh et al. (2021) Scott et al. (2026)

Conceptual History and the Development of the Schoolwide Framework

Behavioral approaches to education developed from experimentally grounded principles concerning the relations among antecedents, behavior, and consequences. Applied educational practice investigated whether environmental arrangements could increase the occurrence of valuable skills and reduce conditions under which disruptive or harmful behavior became more likely. Positive behavior support subsequently combined these functional principles with the practical commitments of prevention, teaching, individualization, and improved quality of life. In their account of the evolution of school discipline, Sugai & Horner (2002) located schoolwide positive behavior support within a historical shift toward designing institutional environments that make appropriate conduct learnable and likely. This history matters because the framework arose as an organizational response to the limitations of purely reactive, event-by-event disciplinary systems.

The schoolwide extension changed the target of analysis. Instead of treating a referral primarily as an individual student's failure, researchers and practitioners examined how supervision schedules, classroom expectations, adult responses, routines for transitions, and team decisions generated predictable patterns of referrals. The institutional system could then be altered through professional learning, consistent instruction, repeated observation, and data-guided problem solving. Horner et al. (2009) tested whether schools receiving technical assistance from ordinary state implementation personnel developed more faithful universal supports. Their randomized wait-list design gave the field evidence about an organizational change process carried out under implementation conditions closer to routine practice than a highly controlled laboratory demonstration.

Research also encountered a methodological problem inherited from discipline data themselves. School records do not constitute standardized observations independent of schools' policies. A recorded decline might result from better prevention, changed threshold for referring, inconsistent record keeping, or several processes together. The Baltimore and Maryland school-randomized research program investigated multiple endpoints, including ODRs, suspensions, school climate, and teacher-rated adjustment, allowing a richer though still imperfect assessment of SWPBIS. The principal studies by Bradshaw et al. (2009), Bradshaw et al. (2010), and Bradshaw et al. (2012) belong to a connected research program, so their reports cannot be added together as if they were independent randomized experiments.

The historical expansion from general behavior prevention to disciplinary equity followed converging evidence that disciplinary systems produced unequal racial and disability-related outcomes. Skiba et al. (2011) documented racial patterns in national school-referral data, and Girvan et al. (2017) examined the contribution of discretionary judgments to those differences. The intellectual implication was profound: a framework organized around positive behavioral expectations also had to examine how its adult decision rules, definitions of misconduct, and culturally situated expectations affected the children it served. Equity-focused PBIS consequently introduced explicit disaggregation of referrals, attention to vulnerability points, and professional development aimed at bias and cultural responsiveness. The randomized evaluation by McIntosh et al. (2021) examined this more specific institutional intervention.

Learning Processes, Ecological Mechanisms, and Institutional Contingencies

The most familiar mechanism in PBIS is instructional: clear expectations can be taught, practiced, prompted, and reinforced in recognizable contexts. Students encounter explicit descriptions of behavior, demonstrations, opportunities for rehearsal, corrective information, and recognition of successful participation. Adult consistency may reduce uncertainty about what is expected, while early acknowledgment makes desirable routines more salient. These are hypotheses concerning behavioral learning within real educational environments. Their operation depends on the quality of teaching, developmental appropriateness, student comprehension, the credibility of adults, and the social meaning of the expectations. A school can formally display its values while providing students with only minimal practice or feedback, illustrating why the presence of posters does not itself establish fidelity. Sugai & Horner (2002) Horner et al. (2009)

The organizational mechanism involves coordinated adult behavior. Teams define referral thresholds, coordinate classroom and common-area procedures, review reports, and provide feedback to one another. Staff may receive coaching and implementation support, and leadership may allocate time for direct teaching of social behavior. This arrangement changes the contingencies affecting teachers as well as students: it clarifies when and how adults should respond, supports cross-classroom consistency, and reduces reliance on improvised reactions under stress. Bradshaw et al. (2009) examined school climate as an outcome of SWPBIS in a group-randomized effectiveness trial. Organizational coherence is a plausible pathway to more predictable classrooms, although measures of perceived climate alone cannot specify which element produced the change.

Behavior is also ecological. A referral rate depends on what activities students are assigned, where adults supervise, the difficulty of instruction, physical accessibility, peer relations, and opportunities for successful participation. The same student may show different behavior in an overcrowded hallway, an inaccessible learning task, a predictable small-group lesson, or an interaction with a trusted adult. Functional analysis therefore asks about antecedents and consequences situated in particular contexts and over time. It examines patterns that make behavior effective for escaping demands, gaining attention, accessing activities, or communicating unmet needs. A function-based interpretation supports intervention designs that adjust the environment and teach alternative skills; it should not be used as a speculative diagnosis of an individual based on a single incident. Sugai & Horner (2002) Sears et al. (2026)

The equity mechanism enters wherever the institution interprets behavior. Consider a classroom in which two students display the same observable conduct but one receives an instructional reminder while the other receives an exclusionary referral. A uniform behavioral curriculum leaves this decision asymmetry untouched unless the referral threshold, educator expectations, supervisory practices, and feedback processes also change. The observed disparities in subjective referrals documented by Girvan et al. (2017) show why disciplinary fairness cannot be reduced to student compliance. Equity-focused PBIS acts on the adult-facing system by making decision points visible, reviewing disaggregated data, and designing responses that better account for context, culture, and the consequences of exclusion. McIntosh et al. (2021)

The Three-Tier Architecture and Functional Behavioral Assessment

PBIS is commonly organized within a multi-tiered system. Tier 1 denotes universal schoolwide supports intended for all students: coherent expectations, explicit teaching, proactive classroom practices, predictable acknowledgment, and a graduated instructional response to behavioral problems. Universal practice has two functions: it offers a shared foundation and creates an environment in which more individualized needs become visible without waiting for repeated failure. The fidelity of Tier 1 is assessed as a property of the implemented school system, not as a label applied to a child. Center on PBIS (2025) distinguishes universal, targeted, and intensive features in the Tiered Fidelity Inventory (TFI) 3.0 and treats each tier as a separately assessable set of organizational practices.

Tier 2 provides targeted supplementary support for students who need greater structure, opportunities for feedback, or explicit instruction beyond universal routines. Such supports may involve scheduled adult check-ins, structured goal setting, small-group skill development, or coordinated family-school communication. Their scientific evaluation requires explicit entry criteria, progress-monitoring variables, an estimate of intervention exposure, and a rule for changing or ending support when results warrant. Unequal access can arise if referral criteria depend on subjective teacher judgments, if students from different groups are selected for support through different thresholds, or if a child is suspended before receiving assistance. Thus, monitoring the distribution of supports across groups is as important as monitoring the distribution of disciplinary sanctions. Center on PBIS (2025) Weeks & Sullivan (2024)

Tier 3 entails individualized, intensive assessment and intervention for a smaller number of students with persistent or complex support needs. Functional behavioral assessment (FBA) ordinarily involves identifying precisely described target behaviors, the circumstances in which they occur, antecedents, consequences, relevant setting events, and contextual demands, then testing a reasoned hypothesis about maintaining relations. A behavior intervention plan translates findings into changes in instruction, environments, alternative communication, reinforcement, and support coordination. Ethical practice requires participation by students and caregivers as appropriate, consideration of disability and accessibility needs, and attention to the least restrictive educational responses available. The strength of an FBA depends on the reliability of observation and whether its recommendations demonstrably improve participation rather than simply remove the student from view.

A tiered framework creates a useful distinction between intensity of need and severity of punishment. The receipt of an intensive service does not signify that a child merits an intensive sanction. Research on students with disabilities makes this distinction especially important. In their review of ten relevant studies, Sears et al. (2026) found a positive association between faithful Tier 1 PBIS implementation and lower exclusionary discipline for students with disabilities, while recognizing that causal evidence remained limited. The finding supports evaluating Tier 1 as a component of inclusive institutional practice, together with separately measured quality of individualized services, disability-related procedural protections, and academic access. Center on PBIS (2025)

Discipline as an Institutional Decision Process

School discipline is a sequence of decisions conducted through a network of classrooms, corridors, administrative offices, policies, and information systems. A complete analytic sequence begins with the availability of educational opportunities and proceeds through exposure to a behavioral demand, the occurrence or perception of an event, interpretation by an adult, choice of response, creation of an administrative record, assignment of a sanction or support, and the subsequent opportunity to return to learning. Each transition has its own selection mechanism. By the time a student appears in a suspension database, several decisions have already filtered the original event. Studying only the final sanction therefore conceals the stages at which disparities first arise. Skiba et al. (2011) Girvan et al. (2017)

Schoolwide PBIS alters some upstream probabilities through teaching and supportive routines, and it can modify downstream thresholds through response protocols. To test which mechanism matters, investigators should distinguish an observed event from a referral, referral from administrative action, and administrative action from actual instructional exclusion. For example, the number of suspensions per enrolled student can decline because fewer concerning situations occur, fewer situations are reported, referral thresholds rise, administrators apply alternatives, or enrollment composition changes. Without corroborating classroom observations, exposure denominators, and decision-specific records, these explanations remain observationally equivalent. Bradshaw et al. (2010) Weeks & Sullivan (2024)

Adult interpretation is particularly consequential when an incident is described as disrespectful, disruptive, defiant, or otherwise dependent on a subjective judgment. Observable aggression and documented prohibited objects also require accurate classification, but their operational definitions are more amenable to behavioral verification. Girvan et al. (2017) compared the contribution of subjectively and objectively defined ODRs to racial disproportionality in records covering more than one million students. Their findings direct attention toward discretionary decision points without implying that ostensibly objective categories are free of contextual interpretation. A complete auditing system records the wording of rules, the initiating observer, the circumstances of the event, and the response actually delivered.

A school-level disciplinary system also creates feedback loops. Students who experience repeated exclusion can miss instruction, have fewer opportunities to build relationships, and return to demands for which they are less prepared. Educators may interpret subsequent difficulties through the lens of prior referrals; data dashboards can inadvertently reinforce this history by presenting risk labels without contextual information. These are plausible processes requiring longitudinal evaluation rather than assumptions about individual students. Equity-oriented prevention breaks the loop by addressing referral thresholds, ensuring access to supports, monitoring repeated exclusion, and observing whether students reengage successfully. Scott et al. (2026) Weeks & Sullivan (2024)

Defining, Measuring, and Interpreting Disciplinary Equity

Disciplinary equity requires the explicit choice of an estimand: the quantity that a study seeks to learn. One estimand is the difference in probability of receiving at least one ODR between groups during a school year; another is a ratio of suspension incidents per 100 enrolled students; a third concerns the probability that an ODR leads to suspension among students with comparable recorded incident characteristics. These quantities answer different questions. The first evaluates exposure to the full referral process, the second reflects accumulated incident counts and repeat occurrences, and the third scrutinizes a downstream administrative decision conditional on having been referred. Conditioning on referral may itself induce selection bias if referral practices differ by group. Skiba et al. (2011) Girvan et al. (2017)

Risk differences, risk ratios, rate ratios, and odds ratios should be specified before comparisons. A risk ratio uses the proportion of students in each group who experience at least one event; a rate ratio may use event counts with person-time or enrollment exposure. Odds ratios are often used in logistic models, but their magnitude differs from a risk ratio when outcomes are common. The number of students with repeated referrals should be reported separately from the number of referrals. A low schoolwide average can hide concentration of repeated exclusion among a small number of students. Group-specific uncertainty intervals and counts matter particularly when some categories contain few students. No single disparity metric provides a complete description of institutional fairness.

Analysts must interpret both absolute and relative change. If a school reduces suspension risk for every group but maintains a substantial relative gap, the intervention has improved absolute access without establishing parity. If a group's referral rate rises because previously ignored peer aggression is now documented, a raw increase cannot automatically be read as a deterioration in discipline practices. Conversely, a school may reduce paperwork while continuing informal exclusion that is absent from databases. A responsible equity dashboard therefore includes attendance and missed instructional minutes, in-school and out-of-school exclusions, event types, repeated incidents, support access, and student reports of procedural fairness. Weeks & Sullivan (2024) Center on PBIS (2023)

Intersectional disaggregation is essential where sample sizes permit. Disability, racialization, economic disadvantage, language background, gender, and grade level can combine in ways obscured by one-variable summaries. Data governance must also safeguard privacy, prohibit stigmatizing public identification, and establish a procedure for verifying coding accuracy and resolving errors. School teams should review patterns in conjunction with local policy, student and family testimony, and opportunities for appeal. The published Center on PBIS guide on discipline disproportionality supplies a practical foundation for data-based team problem solving; its recommendations are implementation guidance rather than experimental evidence of effects. Center on PBIS (2023)

Racialized Referrals, Subjective Categories, and Decision Vulnerability

Racial disproportionality in school discipline is documented in administrative records, although estimates vary by grade, setting, referral type, policy environment, and method. Skiba et al. (2011) analyzed records from hundreds of schools and investigated differences in referrals for Black and Latino students relative to White peers. The scientific significance of this work is its focus on the referral pathway rather than the final suspension count alone. It demonstrates how disaggregated data can locate inequality earlier in institutional processes. Since the study is observational, its statistical associations do not assign the entire difference to a single cause and do not justify assumptions about inherent behavioral dispositions of student groups.

The subjective decision point is a critical mechanism. Girvan et al. (2017) analyzed ODR records for 1,154,686 students across 1,824 schools, distinguishing subjectively and objectively defined misconduct categories. Their research found that discretionary judgments contributed substantially to racial referral disparities. Such findings motivate an intervention on adult response rules, bias-relevant conditions, communication, and classroom practice. But a difference between categories does not prove that one teacher consciously intended to discriminate in a particular incident. Structural routines, accumulated expectations, policy wording, and cognitive shortcuts can yield patterned outcomes even when individual decision makers do not describe themselves as biased.

Institutional vulnerability points are times, places, behaviors, or interactions at which a recurring disparity becomes unusually pronounced. Investigators can map referrals by location, time of day, grade, subject, teacher, type of event, and the presence or absence of previous interventions. A robust design uses disaggregated denominators: a referral hotspot may reflect an area with greater student traffic or more supervision rather than intrinsically higher risk. Corrective strategies should therefore be tied to the hypothesized mechanism. In a corridor with concentrated subjective referrals, explicit expectation teaching and adult coaching may be more appropriate than a generic school assembly; in a classroom with repeated task-avoidance patterns, improving accessibility and instructional fit may be central. Center on PBIS (2023) McIntosh et al. (2021)

Equity-focused analysis also asks who defines appropriate participation. Behavioral expectations can express legitimate safety and educational aims, yet examples of eye contact, tone, movement, deference, or conversational style are culturally and individually variable. Requiring conformity to one communicative style can mistake difference for defiance. A scientifically defensible system translates broad values into observable, teachable practices compatible with disability accommodations and cultural pluralism, and tests whether the system is experienced as understandable and fair. Student and family participation provides information that administrative logs cannot supply. Santiago-Rosario et al. (2022) Scott et al. (2026)

Exclusionary Discipline, Educational Access, and Competing Outcomes

Suspension removes students from at least part of the ordinary educational environment; expulsion represents a more severe form of removal. The consequences of exclusion depend on duration, alternative educational provision, the student's circumstances, and institutional conditions, but lost contact with instruction is a direct outcome that must be counted. Research is often framed as a contest between immediate safety and inclusion, yet the scientific problem is more precise: identify responses that maintain a safe learning environment while minimizing avoidable loss of participation and ensuring consistent application. The appropriate comparison includes prevention, de-escalation, restorative resolution when suitable, individualized supports, and any emergency safety response actually required. Gage et al. (2018) Weeks & Sullivan (2024)

Disciplinary events may have unequal consequences for students already experiencing educational barriers. A student receiving special education services can lose access to planned instruction, accommodations, or supportive relationships when removed; an English learner can lose scarce opportunities for language-rich interaction. The relevant outcome is therefore the duration and quality of educational access, including the effectiveness of services delivered during alternative placements. Sears et al. (2026) reviewed research concerning the association between Tier 1 PBIS fidelity and exclusions for students with disabilities, emphasizing the need for stronger causal designs. This evidence supports careful monitoring rather than an automatic conclusion that schoolwide implementation resolves every disability-related inequity.

An observed reduction in out-of-school suspension may coincide with an increase in in-school suspension, shortened school days, informal requests for families to collect children, or exclusion from extracurricular activities. A single reported rate therefore presents only a part of the disciplinary ecology. Independent observations, attendance systems, educator reports, family contacts, and student perspectives can help triangulate whether a change constitutes genuine inclusion. Another challenge is event severity: an intervention may change the composition of incidents so that administrators handle a different mixture of cases over time. Comparing raw sanction counts without event-level information can then misrepresent the mechanisms of change.

Research designs should separate outcomes that may vary independently. A PBIS program may improve teacher perceptions of safety and still fail to produce a statistically detectable change in ODRs; it may reduce average suspensions while leaving racial disparities intact; or an equity component may lower referrals for the most affected group without changing the overall school climate indicator. These are coherent possible outcome patterns, not mutually exclusive findings. The randomized studies Horner et al. (2009), Bradshaw et al. (2010), and McIntosh et al. (2021), together with the synthesis by Weeks & Sullivan (2024), illustrate why any claim of comprehensive effectiveness requires specifying whose outcome changed, on which measure, during what follow-up period, and relative to what counterfactual.

Foundational School-Randomized Effectiveness Research

The strongest early evidence on universal schoolwide positive behavior support came from group-randomized effectiveness designs in which schools, rather than individual students, received implementation training. Horner et al. (2009) conducted a randomized wait-list controlled evaluation in elementary schools in Hawaii and Illinois over three years. The researchers documented improved implementation of universal schoolwide practices through training and technical assistance by state personnel, alongside favorable changes in perceived safety and the proportion of third graders meeting reading benchmarks. Critically, their abstract explicitly reports that the study lacked experimental control for office discipline referrals and therefore could not establish the intervention's effect on ODRs. This example illustrates why randomized assignment alone does not establish causality for every recorded outcome.

In another research program, Bradshaw et al. (2009) examined school climate during a group-randomized intervention in Maryland elementary schools. Bradshaw et al. (2010) then reported longitudinal effects on implementation fidelity, office referrals, suspensions, and academic outcomes in 37 schools, with findings interpreted through a school-randomized framework. The design addresses some forms of confounding because schools were assigned to conditions, but inference still depends on allocation, attrition, outcome collection, model specification, and clustering of students within schools. Individual student records do not create thousands of independent randomized units: the effective treatment comparison occurs at the school level, with limited numbers of clusters. This point is central when assessing apparent precision or extending a finding to secondary schools or different states.

The Maryland program yielded a later student-level analysis by Bradshaw et al. (2012) in Pediatrics. It used data on 12,344 elementary students in 37 schools over four school years, with repeated teacher ratings and administrative measures. The article reported favorable effects on aggressive and disruptive behavior problems, concentration, social-emotional functioning, prosocial behavior, and ODRs, including an estimated 33% lower likelihood of an ODR for children exposed to the program. This is a specified result in the original publication; it should be interpreted within the study's measures and sample. The 2012 paper is analytically related to the 2010 trial and should not be counted as an independent replication. Bradshaw et al. (2010) Bradshaw et al. (2012)

An instructive discrepancy appears in the What Works Clearinghouse's 2022 review of Bradshaw et al. (2010). What Works Clearinghouse (2022) classified the cluster-randomized design as meeting its standards without reservations, yet reported no statistically significant positive findings for the reviewed academic, behavioral, and disciplinary outcomes under its study review protocol. This does not erase the authors' published findings; it demonstrates that conclusions can change with outcome definitions, statistical adjustments, evidentiary protocols, or decisions about which analyses qualify. A monograph evaluating the literature must disclose this disagreement rather than presenting either the authors' positive findings or the review's null-rated outcomes as an all-purpose verdict. The suitable conclusion is that specific beneficial effects have been reported in strong designs, while the strength of evidence for each endpoint depends on analytic standards.

Systematic Reviews, Meta-Analysis, and Heterogeneous Findings

Evidence syntheses help determine whether early demonstrations generalize across research designs and implementation settings. Gage et al. (2018) systematically reviewed group-based studies of schoolwide PBIS and disciplinary exclusions, identifying four qualifying studies involving 90 schools. In their meta-analysis, the estimated effect on suspension was large and statistically significant, whereas the analysis did not establish an effect on office discipline referrals. The reported suspension estimate, Hedges' g of approximately -0.86, should be read alongside the extremely small number of eligible group studies, differences in contexts and endpoints, and the statistical difficulty of estimating heterogeneity with a sparse set of trials. It does not establish a universal effect of this magnitude in all schools.

A wider analysis by Weeks & Sullivan (2024) assembled 42 articles meeting its inclusion criteria from searches carried out between 2018 and 2020. The review examined associations between schoolwide positive behavior supports, exclusionary discipline, and racial disproportionality. Findings were mixed for overall exclusionary discipline, and few studies offered evidence of reduced inequity in exclusionary outcomes. The methodological value of this review is its distinction between reducing the overall rate of a sanction and changing disparities in who receives it. Its literature search cutoff also defines its temporal coverage: studies published after the search period, including the 2021 equity trial, require separate consideration rather than retrospective attribution to the review.

Reviews differ in the populations and outcomes they are designed to address. Sears et al. (2026) selected ten articles published from 2007 through 2023 on PBIS implementation and exclusionary discipline for students with disabilities. The synthesis found a positive association between Tier 1 fidelity and reduced exclusions in the studies examined, but the authors expressly called for more research establishing a causal relationship. The result does not directly answer questions about racial inequity or a multi-tiered individualized plan. It does demonstrate the importance of defining whose outcomes a review aggregates and how different implementation elements are measured.

The cross-review synthesis is therefore conditional. There is credible evidence that schoolwide PBIS can improve behavioral and organizational outcomes in particular settings, including randomized elementary-school programs. Evidence for suspension reduction is promising in some studies and reviews, while evidence for consistent, generalizable reduction of disciplinary disparities remains limited. A methodological narrative should preserve variation rather than replace it with a single slogan. Review authors may include overlapping primary studies, report outcome domains with different denominators, and analyze implementation fidelity differently. Cumulative confidence requires transparent maps of those overlapping samples and independent replications in schools serving diverse student populations. Gage et al. (2018) Weeks & Sullivan (2024) Sears et al. (2026)

The Equity-Focused PBIS Randomized Trial

The equity-focused PBIS research program tested an intervention specifically organized around reducing racial disparities in discipline decisions. In McIntosh et al. (2021), eight elementary schools with inequitable referral patterns for Black students were matched on existing disparity levels and half were randomly assigned to the intervention. The intervention included analyzing the locations and patterns of disparities, professional development on culturally responsive schoolwide behavior systems, and strategies aimed at the disciplinary decision points identified locally. The study reported significant declines in racial disparities in ODRs and in ODR rates for Black students in the intervention condition, with comparatively little change in control schools. This is direct experimental evidence for an equity-targeted institutional intervention in the sampled schools.

The study's contribution is conceptual as well as empirical. Standard PBIS changes the design of expectations, teaching, and responses; its equity-focused extension adds explicit attention to how adults interpret events, which behaviors become referral-worthy, and where unequal consequences arise. The causal contrast therefore concerns a multicomponent implementation package layered within a PBIS setting. It cannot isolate the independent effect of every component or prove that identical procedures would succeed in all schools. The number of randomized units was eight, which limits precision and the ability to investigate heterogeneous effects across demographic and organizational contexts. A replication in larger, independent school samples remains a scientifically important priority. McIntosh et al. (2021)

A later analysis by Austin et al. (2024) examined 348 teachers from the same underlying randomized research program to explore whether teacher demographic characteristics and attitudes moderated changes in ODR use or intervention acceptability. The analysis investigated individual differences in response to the program and reported findings on its differential effects, but it was not a second independent randomized test of the core intervention. In evidence tables, the 2021 school-level outcome study and the 2024 teacher analysis must share a common trial identifier. This avoids inflating certainty by treating multiple publications from one cohort as separate causal demonstrations.

The Center on PBIS has translated this research direction into a five-point equity approach. Its practice brief emphasizes examining disaggregated data, culturally responsive multilevel systems, effective instruction, equitable policies, and decisions that interrupt patterned bias. Santiago-Rosario et al. (2022) is a useful primary source for the implementation model, while McIntosh et al. (2021) provides the direct experimental evidence relevant to one intervention package. Training manuals and practice briefs can define program procedures, support fidelity assessment, and identify testable mechanisms, but they do not by themselves establish treatment effect sizes. The scientific inference is that intentional modification of adult decision systems can be causally relevant to disciplinary equity, subject to replication and external-validity tests.

Research on Racism-Directed Discipline Interventions

Schoolwide PBIS exists within a larger field of institutional interventions directed at racial discipline disparities. Scott et al. (2026) reviewed 48 school-based studies that directly or indirectly addressed the race-discipline gap in U.S. kindergarten through grade-twelve education. The reviewers concluded that the number of interventions supported by rigorous experiments showing a reduction in racial disparities remained small, despite the size of the general intervention literature. In the review's synthesis, only three experimentally tested interventions reduced the gap, involving strategies such as cultural competence work with educators, coaching on bias and decisions, and data-guided accountability. This result is important because it asks whether interventions affect the disparity itself, rather than whether a program is widely used or positively rated by teachers.

The design of a racial-equity intervention must begin with a causal account of the institutional practice it seeks to alter. Training staff to know that disparities exist may change attitudes without changing day-to-day referral decisions. A generic behavior curriculum can improve predictability without changing the discretion adults exercise. Targeted coaching, feedback at identifiable decision points, and systematic auditing can change the environment in which a consequential decision is made. The evidence reviewed by Scott et al. (2026) supports taking these mechanisms seriously, while also showing that the field still requires better-controlled studies and longer follow-up. An association between an educator's attitudes and student outcomes cannot automatically be treated as proof that shifting self-reported attitudes mediates disciplinary change.

The distinction between racism-directed interventions and the universal PBIS framework clarifies the scope of the research question. PBIS is often used as a delivery infrastructure, offering teams, coaching, data routines, and tiered practices. An equity intervention adds explicitly specified mechanisms targeting racialized patterns of interpretation, policy, and sanction. When the two are combined, attribution becomes more complicated: the outcome can be due to training content, repeated data feedback, administrative leadership, improved classroom supports, or interaction among components. Factorial, dismantling, and stepped implementation designs can help investigate these alternatives, but their empirical findings must not be presumed before corresponding studies exist. McIntosh et al. (2021) Scott et al. (2026)

The larger policy implication is that schools should specify equity outcomes before declaring a discipline initiative successful. A whole-school reduction in suspension may be worthwhile while leaving the disproportionate burden experienced by a particular group unchanged. Conversely, a targeted reduction for the group experiencing the greatest excess risk can improve equity even when the total number of referrals changes only modestly. Evaluations should show the absolute group rates, the disparity measure, the denominator and follow-up time, and the extent to which families and students recognize the disciplinary process as fair. This reporting practice is consistent with the scientific distinction between average effectiveness, subgroup effects, and institutionally meaningful outcomes. Weeks & Sullivan (2024) Scott et al. (2026)

Implementation Fidelity and the Tiered Fidelity Inventory

An implementation framework can be evaluated only when the implemented intervention is described separately from its intended design. Fidelity concerns the degree to which organizational practices correspond to the core features specified in the program. It differs from treatment impact: a school can demonstrate high fidelity while outcomes remain unchanged, and a school's outcomes can improve for reasons other than PBIS. Process indicators include leadership team routines, documented teaching of expectations, use of systematic referral definitions, coaching availability, and delivery of targeted or intensive services. The evidence question is whether the practices actually occurred, how consistently they were used, and whether school conditions allowed them to operate as intended. Horner et al. (2009) Center on PBIS (2025)

The Tiered Fidelity Inventory 3.0, documented by Center on PBIS (2025), separates Tier 1 universal, Tier 2 targeted, and Tier 3 intensive features. The 2025 revision places greater emphasis on mental-health integration, equity, and classroom-level implementation. The instrument is intended to facilitate team assessment, guide improvement, and permit comparisons of implementation trajectories, while its measurement properties should be scrutinized when used for causal inference or cross-district rankings. A fidelity score does not measure a student's experienced safety, cultural belonging, or actual equality of disciplinary decision-making. Researchers should evaluate relationships among fidelity domains, adult behavior, and student-level outcomes instead of treating a total score as a sufficient validity check.

Implementation measures are vulnerable to common-method bias. A school team that delivers the intervention may also score its own implementation, creating incentives and expectations that affect ratings. Independent observations, records of coaching sessions, documentary audits, and student reports provide additional evidence. Repeated assessment can distinguish a short-lived launch effect from sustained practice, while measurement invariance analyses can examine whether an instrument represents comparable constructs across school types. Teams should archive operational definitions because changes in referral coding or fidelity scales can create artificial trends. A version change from TFI 2.1 to 3.0 is an occasion for an explicit measurement bridge rather than a silent numerical comparison. Center on PBIS (2025)

Fidelity should also incorporate equity mechanisms. A school may faithfully teach common expectations while maintaining different informal standards for whose conduct is interpreted as defiant. It can demonstrate functional teams while failing to discuss disproportionality. Thus, measurement of universal practices should be complemented with decision audits, reliable disaggregated data, and evidence that identified disparities actually lead to revised staff practices. Santiago-Rosario et al. (2022) proposes routines intended to integrate equity into PBIS, and McIntosh et al. (2021) provides one trial of an explicitly targeted approach. The conceptual distinction between implementation and fairness prevents fidelity certification from becoming an unwarranted claim of equitable outcomes.

Cultural Responsiveness, Student Voice, and Relational Trust

Behavioral expectations are embedded in language, relationships, power, and local cultural practices. The same response may convey cooperation, disagreement, caution, or distress depending on communicative conventions and prior interactions. Culturally responsive PBIS therefore requires substantive participation in defining expectations, examples, and corrective responses. Students should be able to explain how rules are interpreted in classrooms, hallways, cafeterias, and online school settings, and families should have meaningful channels to challenge policies that create barriers to participation. Such processes improve the information available to teams and can build legitimacy, though their separate causal contribution needs direct testing. Santiago-Rosario et al. (2022) Scott et al. (2026)

Procedural fairness is a relational outcome. The opportunity to be heard, an explanation for a decision, consistent application, access to appeal, and predictable restoration of participation can change how discipline is experienced. A school may reduce ODR counts while students continue to perceive discipline as arbitrary or disrespectful; conversely, improved trust may precede a detectable change in annual administrative totals. Evaluation should therefore combine policy audits with confidential student reports, interviews, and observations of actual adult-student interactions. Qualitative evidence is particularly useful for locating rules that are formally uniform but functionally burdensome for particular students. It also clarifies what a statistical difference means to those affected by it.

Culture-sensitive instruction must avoid substituting stereotyped assumptions about racial or ethnic groups for direct knowledge of a student's needs. Expectations should preserve safety, learning access, and respectful relations while allowing legitimate variation in communication and movement. Teachers need support to distinguish intentional aggression from communication differences, disability-related behaviors, language-learning processes, and ordinary developmental variation. This is one reason why equity training is most appropriately integrated with instructional design, accessible routines, and shared interpretation of behavioral incidents. The schoolwide system becomes more testable when its expectations are concrete and their behavioral examples are negotiated openly. Girvan et al. (2017) Santiago-Rosario et al. (2022)

Partnerships also introduce methodological challenges. Families who attend planning meetings may differ from those who cannot because of employment, transportation, language, or distrust. Student advisory groups may be dominated by students already comfortable with school authority. Research should document how participants are recruited, which groups are missing, and whether recommendations alter policy. Independent and anonymized feedback channels may identify exclusion otherwise invisible to implementation teams. The outcome of participation should be assessed in policy revisions, shifts in referral practices, and continued educational access rather than in meeting attendance alone. Scott et al. (2026)

Disability, Neurodevelopmental Differences, and Inclusive Discipline

Disciplinary equity for students with disabilities requires separating educational support needs from judgments of misconduct. Differences in communication, sensory regulation, attention, motor behavior, emotional expression, and response to demands can influence how school rules are encountered. An accessible behavioral framework anticipates such variation through adjustments to instructions, schedules, physical environments, transition supports, and communication options. Functional behavioral assessment, when indicated and appropriately conducted, can identify patterns relating task difficulty or environmental conditions to distress or avoidance. The purpose is to guide instruction and environmental change. A student's need for more intensive assistance is a reason to increase support, not a reason to presume culpability or to lower expectations of educational inclusion.

The systematic review by Sears et al. (2026) assessed the available evidence on the association of Tier 1 PBIS fidelity with exclusionary discipline outcomes for students with disabilities. Ten studies met its criteria, and the authors described positive associations alongside an explicit need for stronger causal inference. This illustrates a recurrent issue: the descriptive fact that high-fidelity schools have fewer exclusions could be influenced by school leadership, resource allocation, student mix, or prior inclusion policy. Randomized or carefully constructed quasi-experimental comparisons, plus longitudinal measurement of actual supports delivered, are needed to estimate intervention-specific effects. The review provides evidence of promise while delimiting what is established.

Administrative datasets often categorize students by disability eligibility, a variable shaped by access to assessment and classification rules as well as educational needs. Subgroup comparisons require attention to changing eligibility, special education placement, access to inclusive instruction, and age. The same absolute suspension reduction may have different effects depending on how much instruction was already inaccessible. Researchers should examine instructional minutes lost, repeated exclusion, use of informal removals, family-school communication, and continuity of services. Equity analyses should not infer functional needs or diagnoses from referral histories alone. Instead, service-level evidence and documented accommodations can be analyzed alongside behavioral outcomes.

Disability and race also intersect. Averages for students with disabilities can hide differences by racialized identity and gender, and racial averages can hide high-risk combinations. Small strata create unstable estimates, requiring prespecified aggregation rules, privacy safeguards, and uncertainty intervals. Investigators should avoid treating all referrals as equivalent behavioral exposures when some reflect different levels of supervision, disability services, or environmental accessibility. Skiba et al. (2011) Girvan et al. (2017) establish the importance of studying institutional pathways; Sears et al. (2026) directs attention to the quality of evidence for exclusion outcomes among students receiving disability services.

Implementation Systems, Coaching, and Organizational Sustainability

The implementation of PBIS depends on organizational conditions. Leadership establishes formal responsibility, protects time for a multidisciplinary team, coordinates staff training, and ensures that behavior data are collected consistently. Coaching translates abstract rules into adult routines through observation, feedback, and adaptation to local settings. Schools need procedures for integrating new staff, reviewing drift in implementation, and communicating expectations to students and families. The randomized wait-list evaluation by Horner et al. (2009) is notable because it examined technical assistance delivered through existing state implementation personnel, making routine service capacity part of the research question. Implementation science asks whether the system remains effective after specialist support decreases and school leadership changes.

A school can achieve a short-term decline in referrals through intense start-up attention while underlying policies remain unchanged. Sustained improvement requires repeated review of data, a mechanism to revise practices when disparities appear, and institutional memory about why a change was adopted. Coaching content must include not only consistent positive reinforcement and instruction but also the quality of staff judgments at disciplinary decision points. The equity-focused intervention tested by McIntosh et al. (2021) combined examination of local disparities with professional development and concrete changes to systems. These components are plausible candidates for sustained practice, but their long-term independent contributions require follow-up beyond the original trial.

Implementation across districts adds another level. Regional funding, school staffing, class size, local regulation, student mobility, and availability of mental-health or disability-related services can affect the feasible intensity of supports. Cross-site studies should specify adaptation boundaries: which core practices are essential and which implementation details can change without losing the mechanism being tested. A seemingly identical PBIS label may describe interventions with differing quality, dosage, disciplinary alternatives, and data systems. Outcomes should be related to measured implementation characteristics and baseline local conditions, not merely to a binary school's program status. Weeks & Sullivan (2024) Center on PBIS (2025)

Organizational sustainability is also an equity problem. A school may introduce culturally responsive training through a temporary grant and then lose coaching resources, while disciplinary rules remain unchanged. If disparities return, a simple before-after average conceals the time-dependent process. Repeated longitudinal data should measure staff turnover, coaching exposure, policy amendments, fidelity, student composition, and group-specific educational access. This analytic design treats institutional fairness as an ongoing practice maintained by governance and feedback rather than as a one-time certification. Santiago-Rosario et al. (2022) Scott et al. (2026)

Causal Identification and Multilevel Research Design

The fundamental causal question is what disciplinary and educational outcomes would have occurred in the same school, at the same time, had it followed a different support policy. Because one cannot observe both conditions simultaneously, researchers construct comparisons using randomized assignment, rigorous quasi-experimental designs, interrupted time series, or explicitly modeled observational contrasts. Cluster randomization is appropriate when the intervention alters common rules and staff behavior across classrooms. It also introduces statistical dependence among outcomes for students sharing a school. Analysts must account for assignment at the school level, intracluster correlation, baseline differences, cluster attrition, and plausible spillover between schools. The number of students in the dataset is not interchangeable with the number of independently randomized schools. Horner et al. (2009) Bradshaw et al. (2010)

The assessment of equity requires prespecifying both the intervention and the disparity estimand. A school-randomized trial may be powered to detect overall referral reduction yet have limited power to identify an interaction between treatment assignment and racialized student group. When equity is central, researchers must state whether the estimand concerns change in a risk difference, ratio of rates, differential improvement in group-specific outcomes, or distribution of lost instructional time. The interpretation of an intervention-by-group interaction depends on its scale, and a statistical interaction on the odds scale need not be an interaction on the risk scale. Reporting absolute predicted risks is a useful companion to model coefficients. McIntosh et al. (2021) Weeks & Sullivan (2024)

Observational comparisons create additional challenges. Schools that adopt PBIS with high fidelity may already have stronger leadership, greater resources, more reliable administrative records, or a lower baseline propensity toward exclusion. Matching on observable covariates cannot necessarily remove unmeasured selection. Difference-in-differences designs require an assessment of the plausibility of parallel counterfactual trends; interrupted time series require attention to concurrent policy reforms and seasonality. Comparisons of referral counts should include offsets reflecting student exposure and account for changes in enrollment, school closures, and coding policy. Robustness checks should examine how findings shift when severe incidents, repeat referrals, or informal exclusions are included.

Replication involves more than repeating an analysis in the same dataset. Independent trials in urban, suburban, rural, and geographically diverse schools, including middle and high schools, would help establish the transportability of equity-focused PBIS. Multi-year follow-up would reveal whether changes survive staff turnover and implementation decay. Researchers should publicly describe the intervention components, coding rules, attrition, analytic plans, and fidelity assessments to make results interpretable. When administrative records contain sensitive group information, privacy-preserving aggregate reporting can coexist with access for ethical independent auditing. Scott et al. (2026) Center on PBIS (2023)

Construct Validity, Data Quality, and Measurement Triangulation

A disciplinary indicator must be linked to the construct it is claimed to represent. Office discipline referrals measure recorded institutional responses to perceived incidents; suspensions measure formal exclusion events or student exposure to them; self-reports of safety measure perceptions; and direct behavior observations assess specified acts in specified contexts. These are partially related but distinct indicators. Their correlations and differences may contain substantive information about policy and practice. A change in ODR counts alone cannot establish an equivalent change in aggressive behavior, student belonging, or educational equity. Girvan et al. (2017) Bradshaw et al. (2012)

Administrative data systems need explicit audits for missing records, duplicated incidents, ambiguous categories, misclassified sanctions, inconsistent grade or disability codes, and changes in data entry software. The same referral can generate multiple administrative actions, and counting decisions versus incidents changes denominators. The distinction between students suspended at least once and total suspension episodes is crucial for evaluating repeated harm. Fidelity measures also require attention to rater agreement and the extent to which scoring is influenced by the team responsible for implementation. The Center on PBIS (2025) TFI 3.0 offers a structured implementation measure, but independent behavioral, institutional, and experiential measures remain necessary.

Mixed-methods assessment can identify causal hypotheses concealed by an aggregate trend. Repeated classroom observation can test whether adults apply equivalent referral thresholds; student and family interviews can reveal informal exclusion or difficulty contesting errors; administrator records can show the steps between referral and sanction. These sources may disagree, and that disagreement is itself a finding requiring explanation. For example, a school can show fewer formal suspensions but greater use of early dismissals, which would be missed by a dashboard restricted to official suspension fields. Multiple forms of observation make policy substitution harder to mistake for true improvement.

Data should be organized into a longitudinal decision record with clear consent, privacy, access, and retention rules. Useful variables include the learning setting, time, type and severity of incident, adults involved, immediate response, availability of additional supports, actual days or minutes of exclusion, and subsequent student participation. Whenever analyses involve racialized identity or disability status, researchers should justify the purpose of disaggregation and avoid public displays that identify individual children. The ethical value of measurement lies in making school practices answerable to evidence and to those whose education they affect. Center on PBIS (2023)

Complementary Approaches and the Boundaries of PBIS

PBIS does not exhaust the domain of school discipline improvement. Restorative practices emphasize relationships, repair of harm, dialogue, and reintegration; trauma-informed approaches attend to contexts of threat and distress; social-emotional learning supports interpersonal and self-regulatory competencies; and accessible instructional design addresses behavioral difficulties that arise from poorly matched academic demands. These approaches can be combined in a school's prevention system, but the existence of a plausible mechanism is not evidence that every named intervention is effective. The relevant scientific task is to identify which components address which outcomes, for which students, under which implementation conditions. Scott et al. (2026) Weeks & Sullivan (2024)

A restorative process can contribute to fairness when it offers meaningful participation, proportionality, and a practicable path back into the learning community. Its implementation can nevertheless be uneven, with access varying by staff discretion or student identity. An equity evaluation should compare actual referral and exclusion pathways before and after its adoption and assess students' experience of the process. Similarly, classroom-management programs may improve consistency without directly addressing racial disparities in discretionary judgment. Program labels are therefore poor substitutes for evidence about procedures delivered and decisions changed. Scott et al. (2026)

For acute and credible safety concerns, schools still need clear responses that protect students and staff. The evidence question is whether practices are proportionate, documented, time-limited when possible, and accompanied by support and reintegration. A prevention framework cannot eliminate every incident, and the presence of occasional justified emergency action is not by itself evidence of program failure. Conversely, invoking safety in broad or vague terms cannot establish that exclusionary responses were necessary. Valid evaluation separates incident severity and immediate protection needs from longer-term punishment or removal decisions, using transparent and consistent criteria.

The most constructive comparative model treats PBIS as a schoolwide implementation infrastructure whose value depends partly on the practices embedded within it. Explicit teaching, environmental adaptation, equity auditing, family participation, and restorative responses can be operationalized and tested separately or in combination. A factorial experiment could compare combinations of decision coaching and universal instruction; a stepped-wedge design could examine a phased district rollout where appropriate; qualitative process analysis could clarify why an apparent gain was or was not sustained. Such designs are research priorities, not findings already established. McIntosh et al. (2021) Santiago-Rosario et al. (2022)

Research Agenda, Reporting Standards, and Institutional Accountability

Future research should assign equity a primary outcome status when the intervention's public rationale is fair discipline. Study protocols should state the target disparities, the selected metric and scale, the observation period, anticipated missingness, treatment contrast, randomization unit, and power considerations. Authors should describe all concurrently implemented policy changes, the process for handling incidents, and whether staff delivering an intervention also produce its evaluation data. Statistical reports should give group-specific numerator and denominator counts, effect estimates, and uncertainty intervals rather than rely on aggregate statistical significance alone. Independent replications can then compare corresponding quantities across districts. Weeks & Sullivan (2024) Scott et al. (2026)

Mechanistic studies should test specific adult decision pathways. They might ask whether standardized referral definitions reduce subjective-incident disparities, whether coaching changes the decision threshold under ambiguous conditions, whether family participation alters procedural fairness, and whether access to Tier 2 supports mediates subsequent exclusion. Sequential measurement is necessary: a proposed mediator should be observed after the intervention and before the outcome. Analyses must address confounding of mediators and outcomes, and interpretations should avoid mistaking statistical mediation for direct experimental manipulation of a mechanism. Girvan et al. (2017) McIntosh et al. (2021)

Another priority is educational continuity. Researchers often analyze incident counts over a school year, but exclusion can affect attendance, class participation, academic engagement, peer relationships, and family strain over multiple years. Data systems should capture lost instructional time including informal removals, support intensity received, timely reintegration, and student-reported fairness. A school that improves average behavior but fails to protect a high-risk subgroup's instructional access should be evaluated according to both findings. The scientific question is not whether the program can claim a favorable average effect, but how its benefits and burdens are distributed. Sears et al. (2026) Scott et al. (2026)

Sustained progress requires institutional accountability beyond the publication of studies. School boards and administrators can define transparent review cycles, establish a safe route for correcting referral records, make disaggregated aggregate outcomes accessible, and document remedial changes after disparities are identified. Educator preparation can connect behavioral instruction with awareness of discretionary decision points. Community partnerships can help detect unintended effects and test whether rules support participation for students with varied developmental and cultural experiences. These governance measures offer plausible conditions for implementation; their effectiveness remains an empirical question to be assessed with the same rigor as the original behavioral intervention. Center on PBIS (2023) Santiago-Rosario et al. (2022)

Experimental Roots of Functional Assessment and Classroom Management

Function-based behavioral support has an experimental history that clarifies both its power and its limits. Iwata et al. (1994) studied the circumstances of self-injurious behavior in nine individuals with developmental disabilities, using repeated, controlled analogue conditions that varied access to materials, demands, and social attention. The study demonstrated that observed behavior could differ systematically across environmental conditions and that those differences could inform functional hypotheses. It did not constitute a test of universal schoolwide PBIS, a representative classroom sample, or evidence that every school referral can be explained through one simple function. Its significance here is methodological: carefully arranged comparisons can distinguish an environmental hypothesis from a label attached to a person. Translating this logic into schools requires ethically appropriate observation, collaboration, accessibility, and direct checks of whether the intervention helps the student engage in education.

Research on classroom management identifies more concrete practices that can be implemented within the schoolwide framework. Simonsen et al. (2008) conducted a systematic search of classroom-management evidence and described twenty practices judged to have sufficient support for classroom adoption. The practices were organized around classroom structure, explicit teaching and reinforcement of expectations, opportunities for active student response, and instructional responses to challenging behavior. These findings help specify the classroom-level content hidden by the broad label PBIS. They also distinguish a framework's core principle from the observable action through which that principle might influence learning. A classroom receiving PBIS training should be assessed for actual opportunities to respond, clarity of instruction, and observed adult behavior, not only for declared policy adherence.

The distinction between functional explanation and institutional interpretation is scientifically consequential. In a classroom observation, antecedent events and consequences can be coded directly, whereas the term defiance already contains an inference about purpose or intention. Experimental functional analysis can generate and test an environmental hypothesis; disciplinary categorization determines how a school institution responds to the same action. Those processes may be empirically linked, yet they require separate instruments and different ethical judgments. A functional hypothesis should identify the conditions under which behavior varies and the skills or accommodations that might change that relation. An equity audit should identify the circumstances under which an adult chooses referral or exclusion. Both can be studied through repeated observations and structured comparisons. Iwata et al. (1994) Girvan et al. (2017)

The classroom evidence base underscores why universal expectations must be interpreted as educational activities. A rule taught once at an assembly has a different exposure profile from a skill practiced regularly with feedback, opportunities for response, and environmental supports. Variation in instructional dosage can mediate the practical meaning of nominal program participation. Conversely, increasing the consistency of adult rules may reproduce a culturally restrictive practice if the rules are themselves inadequately specified or unfairly applied. Simonsen et al. (2008) helps identify the pedagogy of classroom routines, while Santiago-Rosario et al. (2022) provides a later equity-focused rationale for examining the norms and decisions that surround them.

Functional assessment also raises research-integrity questions. The analyst must guard against inferring a stable behavioral function from a narrow incident sample, ignoring changes in task demands, or treating the observer's preferred explanation as a measured variable. If behavior is uncommon or context-dependent, data requirements can be demanding and the risk of false inference substantial. Collaborative assessment should consider multiple informants and direct observations before a school changes an individualized plan. In schoolwide evaluation, the same caution applies in aggregate: a change in recorded referrals cannot identify a functional mechanism unless the conditions and decisions generating the records are examined. Iwata et al. (1994) Weeks & Sullivan (2024)

Longitudinal Fidelity, School Context, and Differential Benefit

Longitudinal observational research provides information beyond the conditions represented in the major randomized effectiveness trials. Childs et al. (2016) examined data from 1,122 Florida schools, relating scores on the Benchmarks of Quality measure of SWPBIS implementation to administrative outcomes including office discipline referrals, in-school suspensions, and out-of-school suspensions. Higher overall fidelity scores were associated with lower levels of these outcomes in the analyses reported. The study also identified variation across fidelity subscales: classroom-related scores were inversely associated with selected disciplinary outcomes, while the Data Entry Plan subscale showed a positive association with ODRs in some analyses. This pattern warns that more complete data recording can be associated with higher measured counts even when an implementation system is improving.

A plausible mechanism for such a finding is surveillance and documentation. If one school records every minor event reliably and another school documents only selected incidents, an identical count has different substantive meaning. Improved data-entry fidelity may expose more events, change who is referred, or allow teams to distinguish recurrent situations. Consequently, comparisons across schools should account for the quality of measurement and not equate low ODR counts with high educational effectiveness. A longitudinal design can help model differences between schools and changes within schools, but it cannot by itself rule out confounding by leadership, resource availability, or policy change. The empirical contribution of Childs et al. (2016) lies in demonstrating that implementation is multidimensional and that the measurement process is part of the system being assessed.

School context changes which implementation components appear most relevant. Vincent & Tobin (2011) analyzed exclusion data from 77 schools implementing SWPBS and compared associations at schoolwide, classroom, nonclassroom, and individual levels. The researchers reported that classroom implementation was associated with fewer exclusions in elementary schools, whereas nonclassroom practices were associated with fewer exclusions in high schools. At the same time, observed overall declines favored White students, with Black students remaining disproportionately represented, especially in longer exclusions. The study's design and small samples for some comparisons limit generalization. Its significance is the coexistence of apparent overall behavioral benefit and unequal distribution of benefit, as well as the possibility that intervention location matters by school level.

These studies support a contextual interpretation of fidelity. Uniform adherence to a single abstract standard may have different implications for early elementary schools, secondary-school transitions, or highly mobile student populations. Researchers should therefore preserve school type, local practice exposure, data-entry quality, and outcome definitions when synthesizing effects. A study that aggregates classroom and nonclassroom implementation into a total score may obscure the component most closely related to an observed outcome. In addition, neither Childs et al. (2016) nor Vincent & Tobin (2011) uses random assignment to distinguish implementation from the organizational attributes of schools choosing to implement it well. Their associations sharpen hypotheses for future experiments and should be kept separate from causal trial estimates.

The equity implication is especially strong. If White students benefit more from a general decline in exclusion, the rate decrease can be institutionally desirable while leaving the disparity unchanged or enlarged. Research should present changes in absolute group risks and their relative differences, alongside the distribution of intervention exposure. A fidelity analysis that reports only school totals cannot tell whether students from different groups experienced comparable opportunities for instruction, supportive feedback, or appeal. Direct and disaggregated measures of those experiences are essential to evaluating whether a schoolwide intervention realizes its stated universal purpose. Vincent & Tobin (2011) Weeks & Sullivan (2024)

Empirically Identifying Vulnerable Disciplinary Decision Points

The vulnerability-point approach seeks to identify the specific circumstances in which unequal referrals become more probable. Smolkowski et al. (2016) examined 483,686 office referrals from 1,666 elementary schools to assess whether disparity varied across subjective incident categories, classroom context, time of day, and reported severity. The authors reported greater racial and gender disproportionality for subjectively defined behaviors, in classrooms, and for incidents considered more severe, with variation across the school day. This is valuable for program design because it moves beyond a general statement about bias to a map of contexts that schools can measure and potentially modify. The study is observational, so it identifies conditions associated with disparities rather than proving that a particular decision-making intervention will remove them.

A vulnerability point is a patterned feature of an institutional decision environment. It may reflect crowding, educator fatigue, ambiguously worded policy, differences in supervision, the distribution of challenging instruction, or other contextual factors. Researchers should specify which of these are hypothesized and measure them where possible. The same time-of-day effect might have different explanations in a district where students move among classrooms hourly and a primary school that keeps children with one teacher. Intervention design should therefore follow local data and independent observation rather than assume that the same period or behavior category will be vulnerable everywhere. Smolkowski et al. (2016) Center on PBIS (2023)

The relationship between this study and Girvan et al. (2017) provides a useful example of converging but distinct observational evidence. One analysis investigates situations associated with unequal referral patterns, while the other examines the contributions of subjective and objective referral categories to overall racial disproportionality in a large national dataset. Their samples and operational questions should be described separately; even when both point toward discretionary decisions, they do not amount to a randomized experiment. Taken together, they support targeted measurement of adult interpretation as a mechanism worth testing. An equity program should show whether coaching or altered response routines changed referral behavior at the originally identified decision point.

Operationally, an equity team can examine the conditional distribution of referrals by location, offense category, supervising staff role, time of day, and group membership. Raw concentrations should be normalized by student exposure and, where possible, a record of the frequency of similar observed events. Otherwise, an apparent excess in a given location may arise because it is supervised more heavily or serves a larger group of students. Statistical results should be communicated alongside uncertainty, information about missing data, and concrete hypotheses that could be refuted. When results suggest an inequitable rule, transparent changes and repeated follow-up are necessary to demonstrate improvement. Smolkowski et al. (2016) Center on PBIS (2023)

The conceptual advance is the recognition that adult decisions have observable regularities. They can be counted, classified, modeled, challenged, and experimentally altered. This establishes a bridge between social-psychological questions about expectations and institutional questions about rule enforcement. It also creates a stronger standard of accountability: evidence of disparity should trigger investigation of the decision process, while evidence of improvement should refer to an observed change in that process and in the educational opportunities that follow. Girvan et al. (2017) McIntosh et al. (2021)

Comparative Evidence Syntheses on Disproportionality Reduction

A research synthesis must distinguish studies of average discipline rates from studies that test differential outcomes across student groups. Cruz et al. (2021) conducted a best-evidence synthesis of school-based interventions targeting disproportionality in exclusionary discipline. Twenty articles met the review's criteria, and only four provided direct evidence of a reduction in disproportionality using interaction terms. The authors found limited evidence that commonly used interventions reduced disparities and raised the possibility that some programs offered greater protective benefits to White students or girls than to more marginalized groups. Their analysis is particularly relevant because a school intervention may improve its overall disciplinary rate without demonstrating any causal effect on the disparity it was supposed to address.

The broader review by Welsh & Little (2018) investigated racial, gender, and socioeconomic disparities alongside alternatives to exclusionary discipline. Its conclusions emphasized multiple contributors to disparity, including institutional policies and the practices and perspectives of educators and administrators. It found that differences could not be explained solely by socioeconomic conditions or measured misconduct, and it identified important open questions concerning the ability of alternative programs to close the gaps. This contextualizes PBIS within a policy ecosystem involving school governance, school climate, disciplinary categories, and the social distribution of educational opportunity. A behavior program must be assessed against that wider institutional setting rather than interpreted as the sole cause of every observed disparity.

The evidence landscape now contains several syntheses with distinct inclusion criteria. Cruz et al. (2021) focused on empirically evaluated disparity-reduction interventions and sought direct tests of differences across groups; Weeks & Sullivan (2024) systematically considered associations of SWPBS with exclusionary discipline and disproportionality; and Scott et al. (2026) assembled a wider set of racism-directed intervention studies. Their article counts should not be pooled or compared as if they referred to a common population of independent randomized experiments. Each review asks a different question, covers a different period, and may include studies cited in the others. Scientific confidence improves when the unit of synthesis is made explicit and the original study is not counted twice.

The central unresolved problem is the identification of genuinely equity-improving mechanisms across varied schools. The randomized trial in McIntosh et al. (2021) supports one multicomponent PBIS-based approach in eight elementary schools. Yet generalization to secondary schools, differing state policies, and student groups with multiple overlapping disadvantages demands additional trials. Vincent & Tobin (2011) and Childs et al. (2016) demonstrate why implementation setting and fidelity dimensions matter, while Smolkowski et al. (2016) identifies recurrent contexts in which discretionary judgments can be concentrated. Together, these sources offer a cumulative research program: map the disparity, specify its institutional mechanism, test a decision-focused intervention, assess differential effects, and investigate whether improvements remain after changes in staff and leadership.

The mature evaluation standard treats equity as a causal, longitudinal, and participatory outcome. It requires evidence that educational access improved for the students bearing the greatest exclusionary burden, that policy and teacher response mechanisms actually changed, and that data collection was reliable enough to support interpretation. It also calls for careful reporting of incidents, people affected, uncertainty, intersectional differences, and the quality of educational participation after discipline. This synthesis follows the evidentiary direction of Cruz et al. (2021), Welsh & Little (2018), Weeks & Sullivan (2024), and Scott et al. (2026) and identifies stronger independent experimental evidence as the principal requirement for future progress.

Critical Scientific Synthesis of Effectiveness and Equity

The research field contains three distinguishable claims with different evidentiary status. First, schoolwide positive behavioral interventions and supports can be implemented through coherent organizational procedures and can improve selected student and school outcomes. Evidence includes the randomized implementation findings of Horner et al. (2009), the Maryland trial publications Bradshaw et al. (2010) and Bradshaw et al. (2012), and subsequent syntheses. Second, implementation can be associated with lower exclusionary discipline, but findings vary across outcomes and research designs. The group-study synthesis Gage et al. (2018) suggested a suspension benefit while failing to establish an ODR effect; the broader review Weeks & Sullivan (2024) found mixed exclusionary discipline results. Third, reducing racial disparities is a distinct outcome requiring explicitly equity-directed mechanisms. The eight-school trial McIntosh et al. (2021) provides experimental support for that more specific approach, while the much wider racism-intervention review Scott et al. (2026) documents how limited the rigorously tested equity intervention corpus remains.

The first disagreement concerns what constitutes evidence of behavioral effectiveness. Schools can report administrative outcomes, teacher judgments, and student experiences, each of which represents a different process. The author-reported positive results in Bradshaw et al. (2010) and Bradshaw et al. (2012) coexist with the narrower What Works Clearinghouse protocol finding no statistically significant positive results for the outcomes it assessed in Bradshaw et al. (2010). What Works Clearinghouse (2022) This tension should remain visible. It encourages transparent specification of estimands, sensitivity to statistical standards, attention to school-level randomization, and replication with measurements that independently corroborate actual student behavior. A review that makes this disagreement explicit is stronger than one that merges the results into an unqualified efficacy claim.

The second disagreement concerns the location of the problem to be solved. If discipline disparities were attributed only to student-level behavior, a universal instruction program could appear sufficient. Evidence on differential use of discretionary referrals Girvan et al. (2017) and reviews of interventions directed at racial inequity Scott et al. (2026) require a broader model involving adult judgment, policy, accountability, and cultural context. Universal prevention and equitable decision-making can support one another, but their success criteria remain distinct. This analysis also helps explain why an intervention can reduce overall referrals without reliably narrowing group disparities. The scientific response is to measure both outcomes, specify mechanisms for both, and evaluate whether interventions actually alter those mechanisms.

The third unresolved issue is the relation between fidelity and external validity. A highly standardized protocol facilitates interpretation but may be poorly matched to local school conditions; extensive adaptation may improve acceptance while changing the intervention's active ingredients. The research literature has not established a universal optimum between fidelity and adaptation. A practical scientific resolution is to record precisely what was delivered, what changed, why adaptations were made, and how outcomes differed across schools, students, and contexts. Implementation tools such as Center on PBIS (2025) support structured documentation; they are not substitutes for independently measured student participation and fairness. Similarly, guidance such as Santiago-Rosario et al. (2022) specifies promising equity practices whose component effects should remain subject to empirical assessment.

Conclusion

Schoolwide positive behavior support is an organized approach to behavioral prevention that operates through institutional expectations, teaching, adult response routines, tiered assistance, and repeated data-based decisions. Its scientific identity lies in the relationships among these features and their observable effects on school behavior and educational participation. The historical development of PBIS moved school discipline research toward preventive, ecology-sensitive practices, and randomized effectiveness research provided evidence for improved implementation and selected behavioral and organizational outcomes. The strength of these findings depends on study design, the number of randomized schools, the quality of measured outcomes, and the standards used to evaluate estimates. Sugai & Horner (2002) Horner et al. (2009) Bradshaw et al. (2012)

Disciplinary equity is a property of the institution's response to students and the educational opportunities it preserves. It is evaluated through group-specific risk, the consistency and transparency of decision rules, the operation of discretionary referral categories, the availability of support, and the consequences of exclusion for participation. Administrative declines in referrals or suspension rates provide important but incomplete information. Studies of racial disparities and subjective referral categories demonstrate the need to examine adult judgment and policies, while research on students with disabilities illustrates the importance of interpreting exclusion alongside accessibility and the actual supports delivered. Skiba et al. (2011) Girvan et al. (2017) Sears et al. (2026)

The best available synthesis supports two concurrent lines of intervention: universal and individualized behavioral support, and explicit equity-focused revision of disciplinary decision systems. The randomized equity-focused trial demonstrates that a multicomponent intervention can reduce racial referral disparities within the schools studied; systematic reviews indicate that independent causal evidence for broadly generalizable disparity reduction remains scarce. The scientific agenda is to measure institutional mechanisms, distinguish unique trials from repeated publications, conduct adequately powered school-level replications, follow students' real instructional access, and assess the sustained effects of implementation across diverse educational environments. A successful schoolwide system is established through measurable benefits in both behavioral opportunity and fairness of disciplinary decisions. McIntosh et al. (2021) Weeks & Sullivan (2024) Scott et al. (2026)

References

  • Austin, S. C., McIntosh, K., Smolkowski, K., Santiago-Rosario, M. R., Arbuckle, S. L., & Barney, N. E. (2024). Examining differential effects of an equity-focused schoolwide positive behavioral interventions and supports approach on teachers' equity in school discipline. Journal of School Psychology, 104, 101284. https://doi.org/10.1016/j.jsp.2024.101284
  • Bradshaw, C. P., Koth, C. W., Thornton, L. A., & Leaf, P. J. (2009). Altering school climate through school-wide positive behavioral interventions and supports: Findings from a group-randomized effectiveness trial. Prevention Science, 10(2), 100–115. https://doi.org/10.1007/s11121-008-0114-9
  • Bradshaw, C. P., Mitchell, M. M., & Leaf, P. J. (2010). Examining the effects of schoolwide positive behavioral interventions and supports on student outcomes: Results from a randomized controlled effectiveness trial in elementary schools. Journal of Positive Behavior Interventions, 12(3), 133–148. https://doi.org/10.1177/1098300709334798
  • Bradshaw, C. P., Waasdorp, T. E., & Leaf, P. J. (2012). Effects of school-wide positive behavioral interventions and supports on child behavior problems. Pediatrics, 130(5), e1136–e1145. https://doi.org/10.1542/peds.2012-0243
  • Center on PBIS. (2023). Discipline disproportionality problem solving: A data guide for school teams. University of Oregon. https://www.pbis.org/resource/discipline-disproportionality-problem-solving-a-data-guide-for-school-teams
  • Center on PBIS. (2025). Tiered Fidelity Inventory 3.0 (TFI) manual. University of Oregon. https://www.pbis.org/resource/tfi-3
  • Childs, K. E., Kincaid, D., Peshak George, H., & Gage, N. A. (2016). The relationship between school-wide implementation of positive behavior intervention and supports and student discipline outcomes. Journal of Positive Behavior Interventions, 18(2), 89–99. https://doi.org/10.1177/1098300715590398
  • Cruz, R. A., Firestone, A. R., & Rodl, J. E. (2021). Disproportionality reduction in exclusionary school discipline: A best-evidence synthesis. Review of Educational Research, 91(3), 397–431. https://doi.org/10.3102/0034654321995255
  • Gage, N. A., Whitford, D. K., & Katsiyannis, A. (2018). A review of schoolwide positive behavior interventions and supports as a framework for reducing disciplinary exclusions. The Journal of Special Education, 52(3), 142–151. https://doi.org/10.1177/0022466918767847
  • Girvan, E. J., Gion, C., McIntosh, K., & Smolkowski, K. (2017). The relative contribution of subjective office referrals to racial disproportionality in school discipline. School Psychology Quarterly, 32(3), 392–404. https://doi.org/10.1037/spq0000178
  • Horner, R. H., Sugai, G., Smolkowski, K., Eber, L., Nakasato, J., Todd, A. W., & Esperanza, J. (2009). A randomized, wait-list controlled effectiveness trial assessing school-wide positive behavior support in elementary schools. Journal of Positive Behavior Interventions, 11(3), 133–144. https://doi.org/10.1177/1098300709332067
  • Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E., & Richman, G. S. (1994). Toward a functional analysis of self-injury. Journal of Applied Behavior Analysis, 27(2), 197–209. https://doi.org/10.1901/jaba.1994.27-197
  • McIntosh, K., Girvan, E. J., Fairbanks Falcon, S., McDaniel, S. C., Smolkowski, K., Bastable, E., Santiago-Rosario, M. R., Izzard, S., Austin, S. C., Nese, R. N. T., & Baldy, T. S. (2021). Equity-focused PBIS approach reduces racial inequities in school discipline: A randomized controlled trial. School Psychology, 36(6), 433–444. https://doi.org/10.1037/spq0000466
  • Santiago-Rosario, M. R., McIntosh, K., & Payno-Simmons, R. (2022). Centering equity within the PBIS framework: Overview and evidence of effectiveness. Center on PBIS, University of Oregon. https://www.pbis.org/resource/centering-equity-within-the-pbis-framework-overview-and-evidence-of-effectiveness
  • Scott, B. A., Stilwell, S. M., Pearson, Z. V., Zimmerman, M. A., Hsieh, H.-F., & Heinze, J. (2026). Interventions to address racism in disciplinary actions in K–12 schools: A systematic review. Prevention Science, 27(5), 694–723. https://doi.org/10.1007/s11121-025-01855-2
  • Sears, S. L., Xu, X., & Simonsen, B. (2026). Examining the effectiveness of positive behavioral interventions and supports in reducing exclusionary discipline for students with disabilities: A systematic review of the literature. Journal of Positive Behavior Interventions, 28(2), 139–153. https://doi.org/10.1177/10983007251335351
  • Simonsen, B., Fairbanks, S., Briesch, A., Myers, D., & Sugai, G. (2008). Evidence-based practices in classroom management: Considerations for research to practice. Education and Treatment of Children, 31(3), 351–380. https://doi.org/10.1353/etc.0.0007
  • Skiba, R. J., Horner, R. H., Chung, C.-G., Rausch, M. K., May, S. L., & Tobin, T. (2011). Race is not neutral: A national investigation of African American and Latino disproportionality in school discipline. School Psychology Review, 40(1), 85–107. https://doi.org/10.1080/02796015.2011.12087730
  • Smolkowski, K., Girvan, E. J., McIntosh, K., Nese, R. N. T., & Horner, R. H. (2016). Vulnerable decision points for disproportionate office discipline referrals: Comparisons of discipline for African American and White elementary school students. Behavioral Disorders, 41(4), 178–195. https://doi.org/10.17988/bedi-41-04-178-195.1
  • Sugai, G., & Horner, R. (2002). The evolution of discipline practices: School-wide positive behavior supports. Child & Family Behavior Therapy, 24(1–2), 23–50. https://doi.org/10.1300/J019v24n01_03
  • Vincent, C. G., & Tobin, T. J. (2011). The relationship between implementation of school-wide positive behavior support (SWPBS) and disciplinary exclusion of students from various ethnic backgrounds with and without disabilities. Journal of Emotional and Behavioral Disorders, 19(4), 217–232. https://doi.org/10.1177/1063426610377329
  • Weeks, M. R., & Sullivan, A. L. (2024). Systematic review of the associations of SWPBS with exclusionary discipline and disproportionality in U.S. schools. Journal of School Psychology, 106, 101327. https://doi.org/10.1016/j.jsp.2024.101327
  • Welsh, R. O., & Little, S. (2018). The school discipline dilemma: A comprehensive review of disparities and alternative approaches. Review of Educational Research, 88(5), 752–794. https://doi.org/10.3102/0034654318791582
  • What Works Clearinghouse. (2022). Study review: Examining the effects of schoolwide positive behavioral interventions and supports on student outcomes: Results from a randomized controlled effectiveness trial in elementary schools. Institute of Education Sciences, U.S. Department of Education. https://ies.ed.gov/ncee/wwc/study/78551