Traditional EdChoice Scholarship

Eligibility is a property of the building, not of the family. A student qualifies by being enrolled in a building operated by the resident district that was ranked in the lowest twenty per cent of all buildings statewide by performance index score. verified R.C. 3310.03(A)(1) That is the design the corpus’s central caution is about: the program is triggered by a performance measure this site elsewhere shows to be substantially a composition proxy, so the buildings it designates are disproportionately the poorest ones. inference

This is the program whose original mechanism generated the districts’ complaint. Under the deduct-era design, a scholarship was funded by reducing the resident district’s foundation payment — and if the student had never attended the district, the district lost money for a child it had never educated. verified That is no longer the mechanism; see mechanism.

The 2025 report does not add up, in two ways, and both are recorded rather than resolved. Its executive summary says the five programs served “more than 175,000” students; the five program counts it publishes sum to 166,587, a gap of 8,413. And its stated average awards disagree with expenditure divided by participation for this program, the expansion, and Cleveland — while matching exactly for Autism. verified by arithmetic on the report’s own figures The natural reading is a denominator: an average over students holding a full-year award rather than over everyone who used the program at any point. Nothing in the report says so, and the reading is not adopted here on the source’s silence. open

What this repository computed # ↑ Contents

Not what Ohio publishes. Figures derived here from committed fixtures, each one citing the test that reproduces it.

A third way the published average does not sit right, and this one is arithmetic # ↑ Contents

The two above are the report’s own internal disagreements. This one is a constraint the award schedule puts on the figure. Every award under R.C. 3317.022(A)(10) is the lesser of a tuition and a per-grade ceiling of $5,500 for kindergarten through eight or $7,500 for nine through twelve, and in FY2025 those are the operative amounts without indexing — the clause raising them tracks the statewide average base cost per pupil, which H.B. 33 froze at its FY2024 level. verified crates/project

So a published average of $6,808.75 requires a grade mix of at most 0.345625 in kindergarten through eight — and that is before any tuition binds, since every student paying less than the ceiling forces the share lower still. verified crates/project Grades K-8 are nine of thirteen grade levels. A program triggered by a building’s performance index could plausibly skew towards high schools, but a bound near a third is a long way from an even spread, and the report’s own implied average only moves it to 42.7%.

Either this program is overwhelmingly a high-school one, or the published average is on a basis this arithmetic does not fit. The first limb is priced below; the second is the one left standing.

What the high-school limb would cost, now that the eligible population can be counted # ↑ Contents

The designated list carries the grades each building serves and the report card carries what each enrolls, so the eligible population splits on the same boundary the award schedule does. Hand every building that spans the boundary to the high-school side — the assignment least favorable to an elementary reading — and 64.96% of the pupils a designated building makes eligible are in kindergarten through eight. verified crates/xcheck The three published editions read 64.96%, 65.45% and 65.80%, so the shape belongs to the designated set rather than to one year’s list.

Read the eligible population as the pupils still enrolled in a designated building plus the program’s own 42,607 recipients — the two groups that can be counted, which makes it a floor, since a pupil residing in a designated zone who never enrolled there is eligible and is in neither — and the ceiling on the recipient mix forces a take-up rate on each side. It requires 28.02% of eligible high-school pupils to be holding a traditional EdChoice scholarship against 9.98% of the others: better than one in four, against one in ten, in a population that is two thirds elementary. verified crates/dispersion That 28.02% is the floor of the claim and not the middle of it, because the spanning buildings were handed to the high-school side; assign them the other way and it is 34.91%. Reading the report’s implied average instead relaxes the mix to 42.7% and does not relax this — still better than a fifth against under an eighth.

So the first limb is not “this program skews towards high schools”, which is plausible. It is a specific and large claim about who takes it up, and eligibility cannot settle it, because a designated building measures who may claim. What would settle it is a grade-band split of the 42,607: the Scholarship Annual Report publishes participation for each program as one number and no departmental document this project can fetch breaks it by grade, so the split would have to come from the per-student route that report cites and deduction records as entitlement-gated — not from anything committed. open

Who may claim it is now a per-building fact, and the department publishes the criteria # ↑ Contents

The eligibility this node described from statute is published as a list. The department’s Designated List 2026-2027 carries every building in the state that a district operates — 2,877 of them — with the designation and every input to it. 513 buildings are designated, and they sit in 78 districts, one in eight. verified crates/dispersion

This is eligibility and not participation, so it does not close the per-district question the deduction stub records as blocked. It is the public substitute for one of the two portal routes that stub records as unreached — a 404 there was once read as a withdrawal, and is not one — answering a different question at the same grain.

Reproducing the determination finds two things the file does not say. The designation is exactly Option A or Option B, but Option B is not the two criteria it names: it is those two and the building being open, which holds across all 2,877 rows and fails on four without it. verified crates/dispersion And the Title I criterion is an average, not a floor in each year — R.C. 3310.03(A)(1)(b) requires “an average of twenty per cent or more” over three years, which is what the column implements and not what its name says. Reading the name literally misclassifies 164 buildings. verified crates/dispersion

The designation rebuilds from the raw counts, and the ranking sheet hides the population # ↑ Contents

The paragraph above checks the list against itself: it takes the department’s bottom-20% flags as given and asks whether the designation follows. The workbook’s six other sheets are the inputs to those flags — three years of building Performance Index rankings and three of district Title I formula counts — and with them the flags are reproducible rather than given. They reproduce exactly: every yearly flag, the two-of-three column, the Title I threshold column, Option B and the designation itself, in all 2,877 buildings, with no disagreement anywhere. verified crates/dispersion Only the academic-distress route is read rather than derived, because it is a determination of the state board’s that this workbook records and does not compute.

The one column that is read is also the one that can go stale, and it has # ↑ Contents

Every other column is an arithmetic on the sheets beside it, so it is exactly as current as its inputs. Academic Distress School is not an arithmetic at all: R.C. 3310.03(C) makes a student eligible “if the student’s resident district is subject to section 3302.10”, so the column records whether a commission exists on some date. That is (E)(2)‘s test rather than (C)‘s: R.C. 3302.103(E)(1) takes a district implementing its improvement plan out of R.C. 3302.10, and both flagged districts were on their plans, so read literally (C) was off for them while (E)(2) waited for the commission to end. inference Every plan-year edition read follows (E)(2), from the 2022-23 list, which flags all three districts in their first plan year, to this one. verified EdChoice Designated List — which buildings' students may claim a scholarship The department’s rule, OAC 3301-11-03(A), lists buildings meeting (C) and removes them only on division (E), which is why. The workbook’s own docProps gives that date — created and last modified 13 November 2025, thirty-four minutes apart. verified crates/dispersion

The column is maintained rather than carried forward, and Lorain is the proof. Of the three districts ever under the current R.C. 3302.10, Lorain City is flagged no across all fifteen of its buildings, its commission having been dissolved by H.B. 33 on 4 July 2023. verified crates/dispersion

East Cleveland’s flag went stale six weeks after the file was written. In late December 2025 the director released the district from commission oversight on a majority of its plan benchmarks, and R.C. 3302.103(F)(2) dissolves the commission on exactly that — so R.C. 3310.03(E)(2)‘s instruction to “cease awarding first-time scholarships pursuant to division (C)” had already bitten before the 2026-2027 school year this list governs opened, (E)(3) leaving prior-year holders eligible. The file at the department’s URL is still byte-identical to the November edition. verified Academic Distress Commission — the department's commission page

And the exposure is nil, which is the part worth measuring rather than asserting. The stale flag stands on 5 East Cleveland buildings and every one of them satisfies Option B as well — bottom fifth in two of three rankings, Title I average 48.9% — so they are designated on the derived route whatever the read one says. Honoring Option A only for the commission that still exists leaves all 513 designations standing. The whole load-bearing surface of the unreproduced column is 2 buildings, both Youngstown’s Rayen Early College schools, in the one district still under a commission. verified crates/dispersion

Reproducing R.C. 3302.10(A)(1)‘s trigger would not close this, and East Cleveland is why: it was released on an overall rating of three stars, well above the trigger, and was flagged yes all the same. The trigger predicts establishment; the column reports existence, and divisions (N) and (O) and the General Assembly separate the two. inference It is also not computable from what is held — the trigger needs three consecutive years of the overall district rating, and the corpus holds one year of the Achievement component star and no overall rating in any cached report card download. open

The rule is one line and three of its readings are decided nowhere in the file. The cut rounds up, not down — rounding down misses one flagged building in each of the two earlier years. A building carrying no index leaves the population rather than sorting to the bottom of it. And the population is the one R.C. 3310.03(A)(1)(a) names, not the one the sheet prints.

That last reading is the finding. The 2025 ranking sheet holds 359 buildings the statute excludes from the ranking — 258 community schools, 8 STEM schools and 93 Cleveland buildings, against none at all in the 2023 and 2024 sheets — and a community school holds rank 4. Cutting the bottom fifth of the sheet as published flags 646 buildings and misses 163 of the 576 the department flags. verified crates/dispersion The department’s own flag is right: it is a fifth of 2,877, the eligible population. So the arithmetic is the statute’s and the sheet published beside it is not the one the arithmetic was done on — the same shape of defect as the Option B column above, a rule applied more narrowly than the file states.

The Title I criterion is a federal poverty estimate with an Ohio garnish # ↑ Contents

R.C. 3310.03(A)(1)(b) turns on how many of a district’s children “were qualified to be included in the formula to distribute funds under Title I”. The Title 1 sheets decompose that count for 611 districts over three years, and 96.7% of it is census poverty — the Census Bureau’s estimate, which each sheet dates in its own column heading to two years before the school year it decides. verified crates/dispersion The oldest of the three editions rests on a 2021 estimate, five years before the scholarship it helps award.

The four categories Ohio collects each October come to 3.3% between them, and two of them are barely there: the delinquent column is never once a positive number in three editions, and TANF falls from 930 children statewide to none at all between the first edition and the second, with no notice in the file. A reader averaging the three shares is averaging across that change.

The 3.3% is nonetheless decisive at the margin. The four can only add to a count, so they can only move a district over the twenty per cent bar and never under it — and they move 16 of 611 districts over it, six of which hold designated buildings. 31 designations rest on them. verified crates/dispersion

The sheets carry 611 districts against the list’s 606, and the five extra are absent for two different reasons. Cleveland Municipal is ranked in 2025 and report-carded like any district, and is off the list because R.C. 3310.03 puts it off — it stands sixth of 611 on the very criterion it is excluded from. The other four are in no ranking sheet and on no report card at all: three Lake Erie island districts and College Corner Local, whose children attend school in Indiana. A Title I formula count is a count of resident children and needs no building, which is why North Bass Local has a row and a 5-to-17 population of zero. verified crates/dispersion

Three editions now, and what moves is the route the department derives # ↑ Contents

Everything above reads one file. The department serves one edition at a time and deletes the last, so eligibility has been a snapshot here — and the two editions before the current one survive in the Internet Archive, which makes it a series. 460 buildings were designated for 2024-2025, 494 for 2025-2026, and the 513 above for 2026-2027. The designated set grew in both transitions while the population it is cut from shrank in both: 2,937 listed buildings, then 2,906, then 2,877, across 62 districts, then 68, then 78. verified crates/dispersion

Every one of the 80 buildings that entered the designated set entered through Option B. 42 entered for 2025-2026 and 38 for 2026-2027, every one of them meeting the performance-index and Title I route and 0 of them the academic-distress one. Departures run the same way: 27 buildings left across the two transitions — 7 undesignated and 1 delisted, then 18 and 1, which are different events and are counted apart — and 1 of the 27 held Option A. verified crates/dispersion

So the one input this repository reads rather than derives is also the one that moves nobody. The academic-distress column’s load-bearing surface in the current edition is the 2 buildings measured above; across two transitions it accounts for no entry at all and a single departure. Whatever moves eligibility moves inside Option B — a performance-index window and a Title I average, both derived and both reproducible from the sheets beside them. That narrows this node’s central caution rather than confirming it: the margin is not moving on the route a commission decides, and which of Option B’s two criteria moved each entrant is not decomposed here. open

FY2026 has an in-year count, and it sits below the enacted estimate # ↑ Contents

No actual for FY2026 is published yet: the Scholarship Annual Report reaches 2024-25, and LSC’s catalog of budget line items for 2026 has not appeared. What does exist is an in-year figure. LSC’s January 2026 Budget Footnotes counts 42,000 traditional EdChoice recipients as of December 2025, a 1.5% fall from FY2025, and estimates their cost at $280 million. verified crates/project The fall closes on the report: 42,607 less 1.5% is 41,968, which rounds to the stated figure.

The enacted greenbook’s Table 3 had put the same year at 45,260 students and $316.5 million. The in-year figures sit below both, by 3,260 students and $36.5 million: a gap of 7.2% in heads and 11.5% in dollars. verified crates/project

This is a mid-year estimate, not an actual, and it bounds the year without measuring it. Every figure in the paragraph is hedged: approximately, as of December, estimated, roughly. The program takes applications all year, and LSC says the figures “may change throughout the school year”. So the gap says the enacted estimate was running high in December. It does not say the year closed below it. open