Total suspension rate per 100 students by census tract, combining in-school suspensions (ISS), single out-of-school suspensions (OSS), and multiple OSS. School points are sized by enrollment and colored by level (purple = elementary, blue = middle, green = high).
🟢 Green (under 2%) — low suspension rate
🟡 Yellow (2-7%) — moderate
🔴 Red (above 7%) — high exclusionary discipline
Hover over any tract for suspension rate, enrollment, and school count. Click school clusters to see individual school details. Switch to Data Table to search and sort tracts.
Source: CRDC 2020-21 via educationdata API • Rate: per 100 enrolled students • Includes: ISS + OSS single + OSS multiple
Exclusionary school discipline is one of the most well-documented pathways from school to the justice system. The research is extensive and causal.
Exclusionary discipline and outcomes: Students suspended even once are significantly more likely to drop out, be held back a grade, and experience future justice system contact. The effects are not driven by the behavior that triggered the suspension but by the removal from the learning environment itself — lost instruction time, weakened school attachment, and increased exposure to unsupervised time during peak-risk hours (Fabelo et al., 2011; Balfanz et al., 2015).
Racial disparities: Black students are suspended at 3-4x the rate of white students nationally, even after controlling for behavior severity and socioeconomic status. These disparities reflect differences in school responses to equivalent behavior, not differences in student conduct (Skiba et al., 2011; GAO, 2018).
Alternative approaches work: Schools implementing restorative justice practices have reduced suspensions by 40-60% while maintaining or improving school climate and academic outcomes (Gonzalez, 2012; Augustine et al., 2018). The question is not whether to address misbehavior but whether removal is the most effective response.
Red tracts are where schools rely most heavily on exclusionary discipline — removing students from the learning environment rather than addressing the underlying behavioral or mental health needs. These are the tracts where restorative alternatives and school-based mental health services would have the highest marginal return.
The rate at which schools refer students to law enforcement per 100 enrolled students, aggregated to census tracts. This is the most direct measure of the school-to-court pipeline — the point where a school behavioral incident becomes a criminal justice event.
🟢 Green (0%) — no school-based LE referrals
🟡 Yellow (0.05-0.5%) — some referrals
🔴 Red (above 0.5%) — elevated referral rate
The median tract has zero referrals to law enforcement. Referrals concentrate in specific tracts — this is not a statewide practice but a geographically concentrated one. The tracts with the highest referral rates are the places where school discipline most directly feeds the justice system.
Source: CRDC 2020-21 • Rate: per 100 enrolled students • Measure: students referred to law enforcement by school
School-based referrals to law enforcement transform behavioral incidents into criminal justice events. The research shows this pipeline operates unevenly across geography and demographics.
Referral disparities: Schools with more law enforcement presence refer more students to the justice system, even controlling for school size and incident severity. The presence of school resource officers (SROs) increases the probability that routine school misbehavior results in arrest or citation rather than an in-school consequence (Theriot, 2009; Fisher & Hennessy, 2016).
Geographic concentration: In this dashboard, the median Utah tract has zero referrals while a small number of tracts have rates above 1%. This concentration means that a student’s likelihood of being referred to police for school behavior depends heavily on which tract they live in — a place-based conversion factor shaping justice system contact.
Alternatives reduce referrals: Schools that replace SRO-dependent discipline models with counselor-led behavioral intervention see reductions in both referrals and recidivism without increases in school safety incidents (James & McCallion, 2013).
Chronic absenteeism rate — the percentage of students missing 15 or more school days in the academic year, regardless of reason (excused, unexcused, or suspension-related). This is a measure of educational disengagement, not just truancy.
🟢 Green (under 15%) — low absenteeism
🟡 Yellow (15-35%) — moderate (near national average post-COVID)
🔴 Red (above 35%) — severe disengagement
The national chronic absenteeism rate roughly doubled during COVID and has not fully recovered. Utah’s median tract rate of 29.7% is consistent with national post-COVID patterns. Tracts above 45% represent communities where nearly half of students are missing significant instructional time — a leading indicator of dropout risk and future justice system contact.
Virtual and online schools have been excluded from this dashboard because their enrollment and attendance patterns are not comparable to brick-and-mortar schools. Alternative schools serving already-disengaged populations remain in the data — their high rates reflect their mission, not a discipline failure.
Source: CRDC 2020-21 • Definition: 15+ days absent per year • Excludes: virtual schools (CCD virtual flag = 1)
The average students-per-counselor ratio across schools in each tract. Lower numbers (green) mean more counselor capacity per student. Higher numbers (red) mean fewer counselors relative to enrollment.
🟢 Green (under 250) — meets ASCA recommended standard
🟡 Yellow (250-600) — below standard
🔴 Red (above 600) — severe counselor shortage
Source: CRDC 2020-21 (counselors_fte, law_enforcement_fte) • Standard: ASCA recommends 250:1
How schools respond to student behavior — through care infrastructure (counselors, social workers, psychologists) or control infrastructure (law enforcement, security) — shapes whether behavioral incidents become learning opportunities or justice system entries.
In this data: Schools with both counselors and LE have the highest suspension rates (5.7%), but this reflects selection — larger, higher-need schools invest in both. The more telling comparison is between Care-only schools (4.2% suspension, 33.7% absenteeism) and Neither schools (2.8% suspension, 36.7% absenteeism). Schools without any support infrastructure see lower formal discipline but higher disengagement — students leave rather than act out.
Counselor impact: Research consistently finds that school counselors reduce discipline incidents, improve attendance, and increase graduation rates. Each additional counselor FTE is associated with significant reductions in suspension rates and improvements in academic outcomes, with effects concentrated among the most disadvantaged students (Carrell & Hoekstra, 2014; Reback, 2010).
LE in schools: The presence of School Resource Officers increases the likelihood that routine misbehavior results in arrest or citation rather than in-school consequences, with effects disproportionately affecting students of color and students with disabilities (Theriot, 2009; Fisher & Hennessy, 2016).
This map overlays two independent datasets to identify tracts where the school-to-court pipeline is most active:
22 tracts are flagged as high pipeline risk. 20 of 22 are outside the Wasatch Front urban core — concentrated in Tooele (4), Carbon (3), San Juan (3), Cache (2), and rural counties across the state. Two San Juan County tracts have zero youth organizations within reach.
The school-to-court pipeline in Utah is primarily a rural care desert problem, not an urban policing problem. In these communities, the school is the only institution, and without counselors or community organizations, its default response to behavioral challenges is exclusionary discipline that feeds directly into justice system contact.
Discipline: CRDC 2020-21 • Youth services: NCCS BMF v1.2, 2019-2024 • Threshold: top/bottom quartiles
This dashboard draws from three federal data systems accessed via the Urban Institute’s educationdata R package, which provides a unified API to the National Center for Education Statistics (NCES) and the Office for Civil Rights (OCR).
The CRDC is a biennial survey mandated by the U.S. Department of Education’s Office for Civil Rights. Every public school in the country reports discipline counts (suspensions, expulsions, arrests, referrals to law enforcement), chronic absenteeism, staffing (counselors, social workers, law enforcement FTE), and advanced coursework enrollment. Data are disaggregated by race, sex, disability, LEP status, and homeless status.
The 2020-21 collection covers the first full academic year during COVID-19. Chronic absenteeism rates are elevated relative to pre-pandemic baselines and should be interpreted with this context.
The CCD provides the universe of public schools and districts with enrollment, school level (elementary/middle/high), school type, charter status, and the virtual school flag used to exclude online-only schools from this dashboard.
The National Historical Geographic Information System provides a pre-built crosswalk linking each school (ncessch) to its census tract (GEOID). This eliminates the need for geocoding and ensures consistent geographic assignment across all data sources.
The educationdata API returns special codes for missing data: -1 (missing), -2 (not applicable), -3 (suppressed for privacy). All negative values were converted to NA before analysis.
The CRDC discipline and absenteeism endpoints return data disaggregated by sex (male, female, total) and disability status (IDEA, Section 504, neither, total). Naive summation across these subgroups inflates counts by approximately 3x for discipline and 16x for absenteeism. This dashboard filters to the total row only (sex = 99, disability = 99 for discipline; all subgroup columns = 99 for absenteeism) before computing rates.
Schools where any discipline count exceeded total enrollment were flagged and set to NA (5 schools for ISS, 1 for OSS, 1 for referrals, and 404 schools for absenteeism before total-row filtering). After filtering to total rows, only 1 school exceeded enrollment for absenteeism and was removed.
Schools flagged as virtual (CCD virtual = 1) were excluded from all analyses. Virtual schools have fundamentally different attendance and discipline patterns that are not comparable to brick-and-mortar schools.
Teacher experience (pct_new_teachers), first/second year teacher FTE, and firearm incident data were entirely suppressed for Utah (all values returned as special codes). These variables were dropped from the analysis.
This dashboard provides three complementary geographic views of the same discipline data, each with different tradeoffs between precision and coverage:
| View | Geography | Coverage | Tradeoff |
|---|---|---|---|
| Tract (Schools Only) | Census tract containing each school | ~456 of ~709 tracts | Highest precision. Only tracts with a physical school building are mapped. Tracts where students live but no school is located appear empty. |
| Interpolated (All Tracts) | All census tracts, empty tracts filled via nearest-neighbor | All ~709 tracts | Full geographic coverage. Tracts without a school are assigned values from the nearest tract that has one. Interpolated tracts shown with dashed borders and reduced opacity. |
| District View | Unified school district boundaries | All districts statewide | Complete coverage with no interpolation. Aggregates all schools within each LEA. Smooths over within-district variation but enables district-level policy comparisons. |
Each school is assigned to a census tract via the NHGIS crosswalk (school point location to containing tract). Tract-level rates are computed by summing discipline counts and enrollment across all schools in the tract, then dividing: rate = 100 * count / enrollment. This enrollment-weighted approach ensures that large schools contribute proportionally more than small schools to the tract rate.
District-level rates use the same enrollment-weighted approach, summing across all schools within each LEA. District boundaries are from the U.S. Census Bureau TIGER/Line unified school district files (2021). Single-school charter LEAs appear as their own “districts” and should be interpreted accordingly — the district view is most informative for multi-school traditional districts.
The students-per-counselor ratio is computed as total enrollment divided by total counselor FTE within each tract or district. The ASCA recommended standard is 250:1. Schools reporting zero counselor FTE are classified as having no counselor — these are confirmed zeros in the CRDC, not missing data. The combined care staff metric sums counselor, social worker, and psychologist FTE to create a single measure of behavioral health support capacity per 1,000 students.
School poverty rates use the NCES Model Estimates of Poverty in Schools (MEPS), which provides CEP-corrected poverty estimates. Under the Community Eligibility Provision, entire schools can offer free meals regardless of individual student poverty, making raw FRPL percentages unreliable. MEPS uses statistical modeling to estimate the true share of students in poverty. Tract-level poverty is the enrollment-weighted average across schools in the tract.
Attack and threat incident counts are reported directly by schools to OCR. Three categories are included: attacks with weapons, attacks without weapons, and threats with weapons. Firearm-specific incidents were entirely suppressed for Utah and are excluded. Rates are computed per 1,000 students. The weapon rate combines attacks with weapons and threats with weapons as a more serious incident indicator.
The Pipeline Risk view uses the nonprofit infrastructure dashboard’s tract-level youth reach data as the base layer, ensuring all tracts are mapped regardless of whether they contain a school. Tracts are classified as “pipeline risk” if they fall in both the top quartile (75th percentile) of suspension rate AND the bottom quartile (25th percentile) of youth nonprofit service reach. Tracts without school data are labeled transparently and classified based on youth services alone where available.
School discipline data is collected at the school level. When aggregated to census tracts, only tracts that physically contain a school building receive values. This leaves approximately 250 tracts unmapped — tracts where children live and attend schools in neighboring tracts, but which appear as data gaps on a tract-level choropleth. For policy audiences who need to see discipline patterns across entire communities, these gaps are a barrier to interpretation.
Each empty tract is assigned the discipline values of the nearest tract that contains a school, measured as Euclidean distance between tract centroids. This nearest-neighbor approach assumes that students in a tract without a school most likely attend a school in the closest neighboring tract — a reasonable approximation for urban and suburban areas where school attendance boundaries roughly follow geographic proximity.
On the Interpolated (All Tracts) view, interpolated tracts are displayed with:
Users can toggle the interpolated layer on and off to compare coverage with and without estimation.
| Limitation | Description |
|---|---|
| Attendance boundaries | Students do not necessarily attend the nearest school. Charter schools, magnet programs, and open enrollment policies mean the nearest school may not be the attended school. The interpolation assigns the nearest tract’s values, not the attended school’s values. |
| Rural distances | In rural Utah, the nearest school tract may be 20-50 km away. Interpolated values for distant rural tracts are less reliable than for urban tracts where the nearest school is within 1-2 km. The popup displays the interpolation distance for transparency. |
| Homogeneity assumption | Nearest-neighbor interpolation assumes that neighboring tracts share similar discipline environments. This holds better within school districts than across district boundaries, where policies and practices may differ substantially. |
| Not a modeled estimate | This is spatial assignment, not statistical modeling. No covariates are used to predict discipline rates in empty tracts. The interpolated value is simply copied from the nearest data-containing tract, not estimated from neighborhood characteristics. |
Use the Tract (Schools Only) view for precise, school-located analysis. Use the Interpolated view for geographic pattern recognition across the full landscape. Use the District View for policy comparisons that require complete coverage without estimation assumptions. The three views are designed to be used together — convergence across views increases confidence in the findings.
| Limitation | Description |
|---|---|
| COVID timing | The 2020-21 CRDC covers the first full COVID school year. Absenteeism rates are elevated and discipline patterns may reflect pandemic disruptions rather than typical practice. |
| School-to-tract | Schools are assigned to the tract containing their physical location, not their attendance boundary. Students may live in different tracts than the school’s assigned tract. |
| Reporting variation | Schools vary in how consistently they report discipline incidents to OCR. Differences in rates may partly reflect reporting practices rather than true discipline differences. |
| Alternative schools | Alternative schools serving already-disengaged populations remain in the data. Their elevated rates reflect their student population, not necessarily punitive practices. |
| Suppressed data | Teacher experience and firearm incident data were entirely suppressed for Utah and are not available in this dashboard. |
| Interpolation | The “Interpolated (All Tracts)” view assigns nearest-neighbor values to tracts without schools. These are spatial assignments, not modeled estimates. Interpolated tracts are visually distinguished with dashed borders and reduced opacity. See the Interpolation tab for full methodology. |
| District boundaries | District view uses Census TIGER/Line unified school district boundaries. Charter schools appear as their own single-school “districts” and are not comparable to multi-school traditional districts. District aggregation masks within-district variation that may be substantial in large districts. |
| School poverty (MEPS) | MEPS poverty estimates are modeled, not directly observed. They correct for CEP distortion but introduce model uncertainty. Student-teacher ratios exceeding 100:1 (typically virtual or administrative schools) were set to NA. |
| Safety incidents | Incident counts reflect school-reported data to OCR and may undercount incidents in schools with less systematic reporting. Firearm-specific data was entirely suppressed for Utah. |
Augustine, C. H., Engberg, J., Grimm, G. E., Lee, E., Wang, E. L., Christianson, K., & Joseph, A. A. (2018). Can restorative practices improve school climate and curb suspensions? An evaluation of the impact of restorative practices in a mid-sized urban school district. RAND Corporation. https://doi.org/10.7249/RR2840
Balfanz, R., Byrnes, V., & Fox, J. (2015). Sent home and put off track: The antecedents, disproportionalities, and consequences of being suspended in the 9th grade. In D. Losen (Ed.), Closing the school discipline gap (pp. 17–30). Teachers College Press.
Carrell, S. E., & Hoekstra, M. (2014). Are school counselors an effective education input? Economics Letters, 125(1), 66–69. https://doi.org/10.1016/j.econlet.2014.07.020
Fabelo, T., Thompson, M. D., Plotkin, M., Carmichael, D., Marchbanks, M. P., & Booth, E. A. (2011). Breaking schools’ rules: A statewide study of how school discipline relates to students’ success and juvenile justice involvement. Council of State Governments Justice Center.
Fisher, B. W., & Hennessy, E. A. (2016). School resource officers and exclusionary discipline in U.S. high schools: A systematic review and meta-analysis. Adolescent Research Review, 1(3), 217–233. https://doi.org/10.1007/s40894-015-0006-8
Gonzalez, T. (2012). Keeping kids in schools: Restorative justice, punitive discipline, and the school to prison pipeline. Journal of Law & Education, 41(2), 281–335.
Reback, R. (2010). Schools’ mental health services and young children’s emotions, behavior, and learning. Journal of Policy Analysis and Management, 29(4), 698–725. https://doi.org/10.1002/pam.20528
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.
Theriot, M. T. (2009). School resource officers and the criminalization of student behavior. Journal of Criminal Justice, 37(3), 280–287. https://doi.org/10.1016/j.jcrimjus.2009.04.008
U.S. Government Accountability Office. (2018). K-12 education: Discipline disparities for Black students, boys, and students with disabilities (GAO-18-258). https://www.gao.gov/products/gao-18-258