Reference

Methodology

From Texas Bond Review Board data as of November 6, 2026, last updated by the Board on November 6, 2026. Covers elections through May 2, 2026. November 2026 results will be added once local officials certify them.

How this record is built: where the data comes from, what each figure means, and which rows each rule affects.

What this record does not show: turnout, why people voted as they did, tax-rate elections, campaign spending, or elections not reported to the Texas Bond Review Board. Methodology

Source

Every row comes from two Texas Bond Review Board datasets on data.texas.gov: Local Debt Bond Election Results (kbmc-qmvg) for propositions and Debt Outstanding By Issuance Local By Year (3qmf-seku) for debt. The snapshot used here was retrieved on November 6, 2026. Snapshots are kept unchanged; rules add flags and derived fields beside the source values.

Coverage

From Texas Bond Review Board data as of November 6, 2026, last updated by the Board on November 6, 2026. Covers elections through May 2, 2026. November 2026 results will be added once local officials certify them.

Since September 1, 2025, state law requires a local government to report each bond election to the Bond Review Board at least 20 days before election day, and its vote counts within 20 days after (Government Code §1231.025). We have not confirmed what was required before then, so some elections may be missing from the Board's data, especially earlier ones.

A coverage check against county election records is planned. Its results will be published here.

Counties

  • County pages list the jurisdictions the Texas Bond Review Board lists under each county, from the Board's own county field. That field can leave out counties a district crosses: for example, Katy ISD's rows name only Harris County, though its November 2023 order calling a bond election also names Fort Bend and Waller.
  • So a county page may not show each district that holds bond elections in the county, and it never claims that a county had none.
  • A district-to-county list from Texas Education Agency map data was considered. Its published list counts any overlap, however small, as a county, so it was not used. A list that weighs how much of each district lies in each county would be needed, and is not built yet.

Definitions

  • Votes on the measure: votes for plus votes against, as reported. Turnout is not shown, because the data has no count of registered voters.
  • Close: the winning side had less than 52% of the votes cast on the measure (strict bound). The inclusive bound (52% or less) is stated beside it.
  • Pass rates are counted in propositions, not elections or jurisdictions.
  • Not decided: a proposition with no decided result in the data. It was cancelled, it is awaiting results, or it was reported without a result.
  • Debt context comes from the latest fiscal year in the debt data for the matched jurisdiction. Debt per student uses average daily attendance, never headcount.
  • School attendance: an attendance figure of 1 or less is a placeholder, so it is not shown and no figure is built on it. When reported attendance changes by a factor of 2 or more from one year to the next within the 5 years up to the year shown, attendance is shown as reported and debt per student is left out. The finance-years download marks these rows in ada_flag.
How each figure is calculated
FormulaDefinition
F-COUNTNumber of propositions (rows) meeting the stated filter. Propositions, not elections or districts.
F-PASS-RATECarried / (Carried + Defeated), counted in propositions. Cancelled, not reported, pending and awaiting rows are left out and counted beside the rate. Rows whose result contradicts the counts are left out until resolved.
F-UNDER-500Decided propositions where |votes for - votes against| < 500, among the rows the margin rule counts: decided, usable counts (no placeholder, no zero counts, no contradiction between result and counts), 2012 or later, and not a tie.
F-CLOSE-STRICTThe winning side had less than 52% of the votes cast on the measure (votes for + votes against), among the rows the margin rule counts (ties left out).
F-CLOSE-INCLUSIVEThe winning side had 52% or less of the votes cast on the measure, among the rows the margin rule counts (ties left out).
F-MEDIAN-VOTESMedian of votes for + votes against, among propositions with usable counts.
F-TWO-THIRDSCarried propositions where votes for >= two-thirds of votes for + votes against, among carried propositions with usable counts.
F-SMALL-ELECTORATEWater-district propositions with usable counts and 25 or fewer votes cast on the measure.
F-SUM-AMOUNTSum of the authorized amount on the ballot (principal), as reported. Not the total cost.
F-JOIN-RATEPropositions whose jurisdiction matched a government in the debt dataset / all propositions, by name, type and county.
F-SHARE-FORVotes for / (votes for + votes against). The share of votes cast on this measure.
F-MARGIN-VOTESVotes for - votes against.
F-MARGIN-POINTS100 x (votes for - votes against) / (votes for + votes against).
F-DEBT-PER-ADATax-backed (GO) principal outstanding / average daily attendance, same fiscal year.
F-DEBT-PER-RESIDENTTax-backed (GO) principal outstanding / population as reported to the BRB, same fiscal year.

Scroll sideways.

Which rows each figure counts

  • Results (passed, failed, pass rates): every proposition with a decided result, except a row whose reported result contradicts its own reported counts. That row is shown as reported, with a note, and named in every total that includes it.
  • Vote statistics (median votes, share for): the result rule, and also no row with placeholder counts (1 to 0 or 0 to 1) or counts of 0 and 0, and no row from before 2012.
  • Margin statistics (fewer than 500 votes, close results, the closest result on a page): the vote rule, and also no tie. A reported tie is rare and is held back until checked against the canvass.
  • Rows flagged as possible copies or repeats are still counted until a person reviews them.
  • Every fact sentence, table, share card and download uses these same rules. The table below gives each published figure, what it left out and the rows.
Rows counted and left out by each figure, snapshot of November 6, 2026
FigurePeriodRows countedRows left outLeft-out rows
School district propositions decided by a margin of fewer than 500 votesJan 1, 2021 to May 2, 20261,6031 row whose reported result contradicts its counts, 2 rows with placeholder counts (1 to 0 or 0 to 1) and 4 tiesfort-stockton-isd/2025-11-04/c, buckholts-isd/2024-05-04/a, graham-isd/2023-05-06/1, bonham-isd/2022-05-07/1, coupland-isd/2022-05-07/1, graford-isd/2023-11-07/a, hubbard-isd-hill/2025-11-04/a
Close school district results (winning side under 52%)Jan 1, 2021 to May 2, 20261,6031 row whose reported result contradicts its counts, 2 rows with placeholder counts (1 to 0 or 0 to 1) and 4 tiesfort-stockton-isd/2025-11-04/c, buckholts-isd/2024-05-04/a, graham-isd/2023-05-06/1, bonham-isd/2022-05-07/1, coupland-isd/2022-05-07/1, graford-isd/2023-11-07/a, hubbard-isd-hill/2025-11-04/a
Close school district results (winning side 52% or less)Jan 1, 2021 to May 2, 20261,6031 row whose reported result contradicts its counts, 2 rows with placeholder counts (1 to 0 or 0 to 1) and 4 tiesfort-stockton-isd/2025-11-04/c, buckholts-isd/2024-05-04/a, graham-isd/2023-05-06/1, bonham-isd/2022-05-07/1, coupland-isd/2022-05-07/1, graford-isd/2023-11-07/a, hubbard-isd-hill/2025-11-04/a
Median votes on the measure, school district propositionsJan 1, 2021 to May 2, 20261,6071 row whose reported result contradicts its counts and 2 rows with placeholder counts (1 to 0 or 0 to 1)fort-stockton-isd/2025-11-04/c, buckholts-isd/2024-05-04/a, graham-isd/2023-05-06/1
Pass rate, school district propositionsJan 1, 2021 to May 2, 20261,6091 row whose reported result contradicts its countsfort-stockton-isd/2025-11-04/c
Pass rate, school district propositionsJan 1, 2016 to May 2, 20262,1991 row whose reported result contradicts its countsfort-stockton-isd/2025-11-04/c
Pass rate, city propositionsJan 1, 2016 to May 2, 2026704no rows
Pass rate, county propositionsJan 1, 2016 to May 2, 202691no rows
Pass rate, college propositionsJan 1, 2016 to May 2, 202622no rows
Pass rate, hospital propositionsJan 1, 2016 to May 2, 202613no rows
Pass rate, other propositionsJan 1, 2016 to May 2, 20263no rows

Scroll sideways.

Possible copies, still counted

These rows from 2016 on have the same counts as a row on another date, as reported, and may be copies. Each is still counted, under the rules above, until a person reviews it. Its row carries the note "Same counts as" with the other date.

Data level

Every row in this release is shown as "As reported to the Texas Bond Review Board". From the November 2026 update, a row compared by a person with the canvass adopted by the body that held the election will say so, with a link to the document and the date. That level is marked by a border, never by color.

Returned propositions

A failed proposition is marked "Returned" when the same jurisdiction's next bond election, within two years, has exactly one proposition with the same authorized amount and the same purpose group, and no other failed proposition matches it. Both rows carry the note. The purpose text is not compared, because the source often words the same request differently. A re-try with a changed amount is not marked.

Cleaning rules

Each rule and the number of rows it affects in this snapshot. Rules never change a source value.

Cleaning rules and row counts, snapshot of November 6, 2026
RuleRows affected
Source rows10,816
Placeholder votes (non-water decided rows with exactly 1-0 or 0-1)1,419
Non-water decided rows (base for the placeholder count)7,268
Small electorate (water-district decided rows with 1-25 votes)2,694
No counts (zero votes on a decided row)48
Inconsistent (result contradicts the counts, including a tie marked Carried)3
Ties (equal counts above zero)38
Possible copy (same jurisdiction, proposition and counts above 20 votes on different dates)10
Possible copy rows20
Possible repeat (same jurisdiction, amount and proposition on May 2, 2026 and a held date)19
Possible repeat rows38
Stale pending (Pending more than 60 days before the snapshot)14
Held results (dated on or after November 3, 2026)297
County text: distinct raw values310
County text: jurisdictions with more than one county (BRB names)83
Real Texas counties (the cleaned list)254
Counties named by at least one jurisdiction250
County text: spellings and case fixed (Miliam -> Milam; Mongtomery -> Montgomery; WISE -> Wise)3
County text: tokens that are not a Texas county (none)0
County text: lists marked incomplete in the source ("+")1
Purpose: distinct raw values128
Names: BRB names merged by the alias table (S&S CISD = S and S CISD)1
Names: BRB disambiguation suffixes removed from display names (county added)21
Status: cancelled92
Status: not reported (NR)19
Status: pending14

Scroll sideways.

Joining elections to debt

Exact BRB name + type + county text; then name + type with a shared cleaned county; then the normalized name (case, "&"/"and", punctuation) + type with a shared cleaned county. Aliases are tried as names. A name that matches more than one debt government is not matched.

School district propositions from 2021 matched to the debt data: 98.5%.

What the data cannot show

  • How many people saw any campaign (reach).
  • Why people voted as they did.
  • Elections never reported to the Bond Review Board.
  • Vote statistics before 2012, when placeholder counts are common.
  • The tax impact on a household.
  • Campaign spending.
  • Tax-rate elections (VATRE and M&O), which are not in the BRB data.
  • Every county a district crosses: county pages follow the Board's county field.

Standards

  • One template for every jurisdiction, whatever the side of any campaign and whether or not anyone sponsors a page.
  • No sponsor input: no sponsor commissions, previews, approves or edits any record page.
  • Errors are fixed through the public corrections route, with the date and the before and after values.

Sponsors, campaigns and districts report errors through the same public corrections route as everyone else, and only that route.

Running a campaign on a local measure?

Campaign pages are paid for by the committee that sponsors them. They are separate from this record, and no sponsor can change it.