District Coverage Coding Scheme

Version 1.0. Issued 21 August 2026. Status: adopted. Consolidates all normalizations established through state file 10.

1. What the codes measure

The coding scheme measures what a local education agency publishes, and nothing else. It does not measure what an agency does. This distinction is the single most important sentence in the scheme and it must appear verbatim in every workbook README: NONE FOUND is a statement about publication, not about service delivery. An agency serving Deaf students well and publishing nothing about it codes NONE FOUND, correctly. The Index measures the public record because the public record is what a family, a researcher, or a regulator can actually see.

2. The six codes

PROGRAM. A named Deaf or hard-of-hearing programme, a self-contained or regional Deaf and hard-of-hearing classroom, a cluster site, or a named teacher of the deaf, Deaf and hard-of-hearing teacher, educational interpreter, or educational audiologist position.

MENTION. Hearing impairment appears only as a special education eligibility category, a Child Find referral or screening area, a generic related service, or an undifferentiated link to an outside agency. Also applies where the agency names its special education delivery mechanism, whether a state school, a regional unit, a resource network, or another district, without any hearing-specific content.

NONE FOUND. Pages retrieved in full, and no Deaf-specific content published.

NOT VERIFIABLE. The site was located but could not be read even in a rendered browser. The reason must be named. Connecticut and Delaware each finished with exactly one. Florida finished with zero.

NOT AUDITED. Not reached at all. Reserved for tooling failure. Colorado, California, Connecticut, Delaware, Florida, and Georgia all finished with zero, and zero is the target.

OUT OF FRAME. A deliberate, documented scope decision under GOV-002, not a failure.

3. Normalizations

These are binding and must be issued to every audit agent before any page is retrieved, rather than applied retrospectively.

Plain nondiscrimination boilerplate is not hearing content and codes NONE FOUND. Federal nutrition programme boilerplate naming American Sign Language among alternative communication formats is the most common false positive in the series and was caught at six Florida agencies. Website accessibility boilerplate referring to blind and hearing impaired users is the second most common.

Naming the special education delivery mechanism without hearing content codes MENTION, with the mechanism recorded in its own column.

An outside agency named solely for gifted education, food service, childcare, adult education, or website work does not lift NONE FOUND. The agency is recorded in the mechanism column regardless.

A named teacher of the deaf or Deaf and hard-of-hearing staff position codes PROGRAM even without a described programme, carrying a weak-evidence flag on the workbook row.

A job description, job classification, salary ladder, or budget cost centre naming a teacher of the deaf, an educational interpreter, or an educational audiologist, published without any programme page, codes PROGRAM with a weak-evidence flag. Six Florida agencies were in this position, two of them publishing four-level interpreter career ladders with pay grades and no programme page at all. This pattern was invisible in every earlier state because human resources pages were not being read. They are now an audit target.

School nurse hearing screening alone codes MENTION as screening, not PROGRAM.

A page reproducing only the state eligibility definition, with no service model, setting, placement, or staffing, codes MENTION and not PROGRAM. Four Florida agencies were held at MENTION on this rule, one of which publishes a page titled "Deaf or Hard of Hearing" whose entire content is the state definition and a link.

Special education obligations under IDEA are not evidence of a programme.

Abbreviation collisions are checked before any hit is counted. Delaware's Delmar School District used the initials of the state Deaf school to mean itself. Board notices for a public hearing produced a false positive in Florida. Every acronym hit is opened and read before it is coded.

No large agency retains a code derived from a retrieval failure. Every failure goes to a rendered-browser closure pass before coding.

4. Weak evidence

A PROGRAM code resting on a staff position, a job classification, or a budget line rather than a described programme carries a weak-evidence flag on its workbook row, and every weak-flagged agency is named individually in the verification trail. The reconciliation script checks that the two lists match.

5. Reporting

Coverage is reported against both the full frame and the readable subset, and the two percentages are given separately. Where a code changed between the first pass and the browser closure pass, the change is recorded in the verification trail with its direction. In Florida twelve of seventy-seven codes changed and every change was upward, which is the strongest available argument that a first-pass-only audit understates provision.