How Many Students With Disabilities Represent Students Experiencing Homelessness? The 20.3% Answer

In school year 2020–21, 220,599 students with disabilities served under the Individuals with Disabilities Education Act (IDEA) represented a reported 20.3% of the 1,087,283 students experiencing homelessness in the U.S. Department of Education’s EDFacts subgroup denominator.

Which number answers the question: 20.3%, 220,599, or 1.1 million?

Use 20.3% when the question asks what share of students experiencing homelessness were identified as children with disabilities under IDEA. Use 220,599 when it asks how many such students schools identified. Use 1,099,221 for the broader count of students public schools identified as experiencing homelessness during 2020–21.

Those figures answer different questions. The National Center for Homeless Education’s 2022 Student Homelessness in America report drew them from U.S. Department of Education EDFacts file specification 118, state educational agency level data. Here is the source table in plain English:

| Measure | 2020–21 figure | What the figure counts | | --- | ---: | --- | | Students identified as experiencing homelessness | 1,099,221 | Enrolled students identified across the school year | | Share of public school enrollment | 2.2% | Identified homeless students divided by all enrolled public school students | | Children with disabilities | 220,599 | Homeless students in the IDEA service category | | Reported disability share | 20.3% | IDEA count divided by the subgroup reporting denominator | | Unaccompanied homeless youth | 94,363 | Homeless students outside the physical custody of a parent or guardian | | Students in doubled-up housing | 844,245 | Students sharing housing because of housing loss, economic hardship, or a similar reason |

The denominator deserves attention. Table 5 of NCHE’s 2022 report gives 1,087,283 as the total represented in its subgroup data; 220,599 divided by that figure is 20.29%, which rounds to 20.3%. The report’s main enrollment total is 1,099,221. Its later historical tables use 1,099,269, reflecting revised aggregation, while retaining 20.3%. NCHE explains that historical totals can vary when analysts use state education-unit totals versus aggregated category totals.

Using the 2022 report’s own figures, the subgroup denominator covers 98.9% of the students in the main enrollment count. That is the useful reporting-coverage rate. The remaining 1.1% cannot be coded as students without disabilities; their disability detail is absent from this particular denominator. Coverage belongs beside the percentage whenever a district has the same mismatch.

So a citation should keep the official rate, numerator, subgroup denominator, and school year together. Dividing 220,599 by the broader 1,099,221 total produces 20.1%, a defensible recalculation with a different denominator, though it is not the published 20.3% rate. The 2020–21 state-reported identification rate was 2.2% of public school enrollment; it ranged from 0.7% in Connecticut to 6.3% in Bureau of Indian Education schools in NCHE’s state table.

This is annual enrollment data, accumulated across a school year. It is not a one-night head count, a count of families in shelters, or a survey of every child with a disability.

How should an educator interpret the 20.3% disability share?

Treat 20.3% as a program-participation measure: roughly one in five students identified under McKinney-Vento was also reported in the “Children with disabilities (IDEA)” category. It does not estimate every disability, health condition, accommodation need, or suspected disability among students with unstable housing.

The appropriate national comparison is 15%. NCHE, citing the National Center for Education Statistics, reported that 15% of students overall received IDEA services in 2020–21. The homeless-student share was therefore 5.3 percentage points higher, or about 1.35 times the general rate. “Five percentage points” and “35% higher” describe the same gap in different forms; educators should label which one they use.

The sudden sun across my desk has caught the smallest words in the table: “Children with disabilities (IDEA).” Those parenthetical letters control the meaning. Running a municipal medicine take-back box taught me the same discipline. A gross weight could include unfinished antibiotics, one bottle of liquid morphine, and vitamins by the kilogram. The weight was real. It still could not tell me how many prescriptions had been abandoned.

An IDEA count is equally specific. IDEA eligibility requires an evaluation, a qualifying disability that adversely affects educational performance, and a need for special education and related services. A student with a Section 504 plan alone is outside this numerator. A student awaiting evaluation is outside it too. My father’s medicine cabinet took two hours to clear after his funeral; its leftover stock recorded what had been prescribed and kept. It could never describe all the care he had needed.

Later data suggest the one-in-five pattern persisted. NCHE’s February 2026 data summary reported 301,359 children with disabilities under IDEA among 1,548,191 students experiencing homelessness in 2023–24, or 19.5%. That newer rate updates the trend. It does not replace the 2020–21 answer sought here.

Why can homelessness-disability data understate student needs?

Two identification systems must recognize the same student before the student reaches the numerator. McKinney-Vento identification comes first in the denominator; IDEA evaluation and service eligibility supply the disability flag. A missed identification at either gate makes need disappear from the intersection.

The McKinney-Vento Act uses a broad education definition for a practical reason. It covers children who lack a fixed, regular, and adequate nighttime residence, including those sharing another person’s housing after housing loss or economic hardship; staying in motels because adequate alternatives are unavailable; living in shelters; or sleeping in cars, parks, substandard housing, and similar places.

The doubled-up category shows what a shelter-only count would erase. EDFacts recorded 844,245 doubled-up students in 2020–21, 76.8% of the 1,099,221 identified total. It also recorded 94,363 unaccompanied homeless youth, 8.6% of that year’s total. Many students in both groups can look housed on an enrollment form.

The 20.3% can still understate disability-related need because:

• mobility interrupts evaluations, records transfers, IEP implementation, and the teacher observations that often trigger a referral;
• a child may need accommodations under Section 504 without meeting, or yet having completed, IDEA eligibility;
• trauma, sleep loss, absenteeism, interrupted instruction, and a disability can produce overlapping classroom signals; a percentage cannot sort them for an individual child;
• 2020–21 was an unusually difficult identification year. NCHE reported 14% fewer students experiencing homelessness than in 2019–20, while total public school enrollment fell by about 3%.

The rate can mislead in the other direction as well. If a district identifies homelessness more consistently among students already connected to special education staff than among other students, its intersection rate rises through denominator bias. A high rate may reflect unmet need, strong cross-program identification, selective McKinney-Vento identification, or some mixture. The percentage alone cannot choose among those explanations.

There is another quiet limitation. File specification 118 records primary nighttime residence at the point of identification. It is an annual, unduplicated state count, yet the housing category is a snapshot of the student’s circumstances when identified. A move from a motel to a car two months later will not necessarily change that federal category.

What can staff do when district homelessness data omit disability detail?

Do not estimate the missing intersection by multiplying the local homeless count by 20.3%. That would import a national historical rate into a district with different grades, identification practices, mobility, and missingness. Build the local count from student records under the district’s privacy and access rules.

1. Lock the cohort. Define the school year and include each student who was enrolled while experiencing homelessness. Keep a documented rule for students who changed schools or regained housing.
2. Name both flags. Use the McKinney-Vento identification field and the IDEA-served field. Keep Section 504, referral pending, evaluation pending, and unknown as separate statuses rather than folding them into IDEA.
3. Reconcile the records. Have an authorized data owner match the two program files by student identifier. Count once at district level, even when a student enrolled in multiple district schools.
4. Preserve the residue. Report missing and unknown values beside the numerator and denominator. Suppress small cells according to district and state privacy rules.
5. Check the cases behind the gaps. Review students who arrived without records, transferred during evaluation, changed housing, or are unaccompanied. These are service prompts, not reasons to alter a code without evidence.
6. Publish the equation. Write “IDEA-served students experiencing homelessness ÷ all identified students experiencing homelessness with known IDEA status,” followed by the counts, rate, school year, and extraction date.

If staff cannot perform a lawful record match, the useful output is a documented data gap. Ask the state homeless education coordinator or EDFacts coordinator for the LEA-level FS 118 subgroup extract and its data-quality notes. In the meantime, case-level coordination can continue; a dashboard is never a prerequisite for serving a student.

When should McKinney-Vento and special-education staff coordinate for one student?

Coordination should begin when housing instability could delay enrollment, evaluation, services, placement, transportation, or parent participation. NCHE’s 2024 problem-solving brief recommends ongoing collaboration and annual joint data review. For an individual student, these moments warrant same-day contact between the local liaison and special-education lead:

| Trigger | Immediate joint task | Federal anchor | | --- | --- | --- | | Student arrives without an IEP or records | Enroll the student, request records promptly, and arrange appropriate interim support | McKinney-Vento requires immediate enrollment despite missing records | | Student transfers with an existing IEP | Provide services comparable to the prior IEP while the new district adopts it or develops a new one | IDEA transfer provisions | | Student moves during an initial evaluation | Agree with the parent on a specific completion date and document sufficient progress | IDEA’s transfer exception to the evaluation timeline | | School-of-origin or transportation decision affects services | Put the liaison, IEP team, transportation staff, and family at the same table | McKinney-Vento school stability and IDEA related-service duties | | Unaccompanied youth may need an evaluation | Resolve who can act as parent and whether a temporary surrogate parent is needed | IDEA permits appropriate shelter, transitional-living, or outreach staff to serve temporarily |

IDEA’s federal initial-evaluation period is 60 days after parental consent unless the state has established its own timeframe. A transfer exception is not an open extension: the new district must make sufficient progress, and the parent and district must agree to a specific completion time. NCHE also states that IDEA Child Find includes children who are homeless and may have disabilities.

The meeting should decide concrete ownership. Who obtains records today? Which comparable services start now? Who contacts the parent, caregiver, or youth? Which transportation cost follows housing instability, and which follows the disability? Leaving those questions with “the team” is how a student loses another week.

How can a district track equitable identification over future school years?

Keep one small scorecard whose definitions survive staff turnover. At minimum, show the number of McKinney-Vento-identified students, the number and share served under IDEA, Section 504 alone, referrals opened, evaluations completed within the applicable timeline, comparable services started after transfer, and unknown disability status. Break results out by grade band, primary nighttime residence, unaccompanied status, race or ethnicity, and school when cell sizes permit.

Trend both counts and rates. A flat 20% alongside a rising homeless count means more students require coordinated service. A falling rate can reflect improved identification of homelessness among students without IEPs, stalled evaluations, lost records, or a genuine population change. Read it beside the unknown rate, referral flow, and service-start time.

Use the NCHE FS 118 quality checks as the annual audit spine. Its May 2026 brief says state totals should count a student once, LEA totals may count the student in each LEA attended, subgroup categories can overlap, and counts must be cumulative across the school year. It also tells reviewers to investigate subgroup changes of 15% or more from one year to the next. I would add one local control: reconcile every published percentage to its displayed numerator and denominator before release. The national 20.3% denominator issue is reason enough.

Set a review calendar around the data rather than around compliance deadlines alone: a monthly case review for transfers and incomplete evaluations, a quarterly check of identification and service-start measures, and one certified year-end extract. Freeze the definitions with that extract. If a field changes in the student information system, preserve a crosswalk so a three-year trend does not become a comparison of three different coding rules.

Frequently asked questions

What percentage of students experience homelessness?

In 2020–21, public schools identified 1,099,221 students experiencing homelessness, equal to 2.2% of public school enrollment, according to NCHE and NCES. This is a cumulative school-year identification rate under McKinney-Vento. It differs from a one-night housing count and misses students whom schools did not identify.

What percentage of people experiencing homelessness have a disability?

No single federal percentage is directly comparable to the student figure. HUD measures homelessness through housing-system data and a one-night count; education agencies count students across a school year and classify disability here through IDEA service status. Any broad estimate requires its population, date, disability definition, and denominator before comparison.

How many students with disabilities represent students experiencing homelessness in 2022?

NCHE’s 2022 report identified 220,599 students experiencing homelessness who were children with disabilities served under IDEA, a reported 20.3% in school year 2020–21. “In 2022” refers to the report’s publication year. It is not a calendar-year 2022 count or the later 2021–22 school-year figure.

Why does the McKinney-Vento Act define homelessness broadly?

The definition reaches unstable housing that creates educational barriers even when a child has a temporary roof. In 2020–21, 844,245 identified students were doubled up because of housing loss, economic hardship, or a similar reason. A shelter-only definition would have excluded more than three-quarters of the education count.

Which services are students experiencing homelessness automatically eligible for?

Students identified under McKinney-Vento are automatically eligible for Title I, Part A services, and USDA rules make homeless children categorically eligible for free school meals. They also receive immediate-enrollment, school-of-origin, and related transportation protections. IDEA services still require an individual evaluation and eligibility determination; McKinney-Vento status alone does not create IDEA eligibility.

What does the disability percentage’s denominator include?

The published 20.3% uses NCHE’s 2022 subgroup-table denominator of 1,087,283 enrolled students experiencing homelessness in 2020–21, including ages 3–5 outside kindergarten, kindergarten through grade 13, and ungraded students. The broader enrollment table reports 1,099,221, so any quotation should preserve the subgroup denominator and source table.