Enrollment Measurement Framework

Version 1.0. Issued 21 August 2026. Status: adopted.

1. The problem this framework solves

Enrollment figures for Deaf education look simple and are not. A single state will publish a certified district-level count, a special education child count taken on a different date under a different definition, a state school enrollment figure that may or may not include outreach caseload, a federal Section 618 figure aggregated to a category boundary that does not match the state's own, and a screening or early intervention count covering a different age band entirely. These instruments are not interchangeable, and treating them as interchangeable is how most published claims about Deaf student numbers go wrong.

The framework below requires that every figure the Index carries names its instrument, its date, and its source, and that figures from different instruments are never combined or reconciled without an explicit statement that they have been.

2. Required elements for any enrollment figure

Every enrollment figure recorded in the Index carries five elements without exception. The number. The instrument that produced it, named specifically rather than as "state data". The count date or collection period. The population definition, including the age or grade band and the eligibility category. The source, as a full citation with the access date.

A figure missing any of these five is not recorded as a figure. It is recorded as a gap, with a note stating what is known and what is missing.

3. The instruments and what each measures

Certified enrollment count. The state's official pupil count, usually taken on a named date and used for funding. This is the instrument that reconciles across levels of aggregation and is the Index's preferred source for total enrollment at a named institution.

Special education child count. The count of students with disabilities by primary eligibility category. Where a state publishes this by local education agency, it is the single most valuable data object in the state file, because it converts district publication silence from an unmeasurable observation into a measurable one. Florida's is published as a workbook with one worksheet per exceptionality. The Index requires that this file be sought early rather than late, and that its suppression floor, its category label, and the treatment of state schools within it all be established explicitly.

Educational environment or least restrictive environment distribution. The distribution of students across placement settings. Published by disability category in almost no state. This absence has now held in Colorado, California, Connecticut, Delaware, Florida, and Georgia. It is reported as a streak because a streak is the finding.

State school enrollment. The population of a named institution. The standing trap is campus and outreach conflation, which has appeared in every state examined. A state school that reports a single figure combining resident students, day students, and an outreach or early intervention caseload has published a number that cannot be compared with any other state's. The Index separates the components wherever the source permits and states explicitly where it does not.

Federal Section 618 data. Useful as a cross-check, unreliable as a primary source for this field, and frequently inaccessible. Where a state's own publication is better, the state's own publication is preferred and the preference is stated.

Early Hearing Detection and Intervention figures. Screened, referred, diagnosed, and enrolled in early intervention. These cover infants and are not comparable with school-age counts. The diagnosis to early intervention gap is a specific measure the Index captures wherever both figures exist, and state and federal figures for the same year are recorded side by side and are not reconciled if they differ.

National deafblind child count. A distinct instrument from a state's IDEA deafblindness category count, measuring a different population under different criteria. The two are recorded separately and are never reconciled. Florida's diverged by a factor of six.

4. Verification rules

Three rules are mandatory and have each proved decisive in completed state files.

Verify every load-bearing figure yourself from the source document. A figure that will appear in the narrative's most consequential paragraph is re-read directly from the source before it is written. This rule was promoted after a delegated agent reported a category total correctly and inverted its distribution across educational environments.

Make extracted rows sum to the published total. When a table is extracted, the sum is the check. Where the sum falls short of a published total, the residual is examined against the suppression rule and the arithmetic consistency is stated. Florida's district-level counts summed to 3,871 against a published state total of 3,952, and the residual of 81 across 25 suppressed cells was arithmetically consistent with a suppression floor of ten, which confirmed that no row had been misread.

Cross-check instruments against each other and record what you find. Where two instruments covering the same population agree, say that they agree and say that you checked. Where they disagree, record the conflict and do not resolve it silently.

5. Suppression

Every count subject to a suppression rule carries the rule. Suppressed cells are recorded as suppressed rather than as zero, and any total computed over a suppressed table is reported as a floor with the residual stated.

6. Derived measures

The Index computes two derived measures where the underlying data permits, and computes neither where it does not.

Publication against identified population. How many identified Deaf and hard-of-hearing students sit in a local education agency that publishes a programme, as a count and as a share of the state total, with every agency that reports a measurable population while publishing nothing named individually. This measure exists only where the state publishes counts by agency. In Florida it produced 92.0 per cent in a publishing agency and four named systems publishing nothing against populations of eleven to nineteen students.

Separate-school share against state school enrollment. Where a state publishes educational environment data by category, the separate-school and residential shares are compared against the named state schools' combined enrollment. No state has yet supported this measure.

Neither measure is ever estimated in a state that does not publish the underlying data. The absence is reported instead.