# Independent review of the fixed 2023 UCP top fifteen

**Data and application source: ACCEPTED.**

Reviewed by GPT-6 Astra on October 21, 2026. This is an independent review of the frozen data, arithmetic, downloadable exports and page source. Build, browser, print and deployment verification are separate release checks; this statement does not claim those checks were performed by this reviewer.

The comparison selects the fifteen Alberta electoral divisions with the highest UCP share of valid votes in the 2023 provincial election, then shows counted valid votes for 2026 Q5 and Q10 in those same areas. It does not identify individual voters or their political affiliations.

## What was verified

A fresh download of the [official 2023 Statement of Vote workbook](https://www.elections.ab.ca/uploads/2023-Provincial-General-Election-Statement-of-Vote.xlsx) matched the retained source. UCP candidate votes and all valid candidate votes were independently extracted from all 87 division sheets. Exact ratios determined the selection and ranks before display rounding.

The [official 2026 referendum page](https://results.elections.ab.ca/8400) verified division and question identities. Ballot Q10 uses source question ID 11; Q5 uses ID 5. Both questions come from one frozen response retrieved **October 21, 2026 at 4:29 PM MT**: precisely **22:29:51.816553 UTC**, with upstream update **22:29:38.8167043 UTC**.

Independent calculations passed **32,217 field, reconciliation and export checks**. These covered all 87 divisions, all 174 question/division records, voting methods, answer counts, percentages, exact provincial ranks, the fixed cohort, provincial/cohort aggregates and every field in the downloadable CSV. All question/division poll sums reconcile with their corresponding division totals. The snapshot contains no unknown method codes or unexplained residuals.

| Verified measure | Count or share |
|---|---:|
| 2023 province: UCP / valid votes | 928,900 / 1,764,915 |
| Fixed fifteen: 2023 UCP / valid votes | 201,506 / 270,758 |
| Fixed fifteen: valid Q10 votes | 294,436 |
| Fixed fifteen: Q5 No / valid Q5 votes | 69,098 / 279,276 = 24.741832% |
| Province: Q5 No / valid Q5 votes | 686,713 / 1,727,295 = 39.756556% |
| Fixed fifteen: Q5 / Q10 mail-in valid votes | 31,826 / 38,498 |

All thirty chart bars use the same zero-based **0–30,000 valid-vote scale**. Q5 No-share ranks compare all 87 divisions; rank 1 is the highest No percentage. The method summary divides summed No votes by summed valid Q5 votes within the same method and fixed group. It does not average division percentages.

## Limits that must remain visible

This is a **frozen, unofficial partial count**. Each question has 5,354 numeric location rows out of 5,575, including 221 null rows. Q5 contains 230 numeric zero/zero rows; Q10 contains 58. Matching numeric-row coverage does not establish equal counting progress or final completion.

For example, Q5 mail-in counts are numeric zero/zero in Peace River and Chestermere-Strathmore, while their Q10 mail-in totals are 953 and 5,806. Those source zeros do not establish that nobody used mail for Q5 or that voters skipped that question. Populated rows may also be revised.

The comparison does not establish turnout, skipped questions, cause and effect, individual political affiliation, or how the same people voted across elections or questions. It is a separate capture from the earlier mail-voting comparison. Q10 option 2 concerns beginning the legal process for a binding referendum; it does not itself decide separation.

The page source states these limits, includes exact question wording and denominators, and separates any reconciliation discrepancy from recorded voting methods. No unresolved priority-one or priority-two data/source finding remained at acceptance.

## Reviewed downloads

- [All 87 divisions: CSV](https://alberta-referendum.vercel.app/data/conservative-vote/results.csv) — 58,554 bytes; SHA-256 `04b085bdc09c34b2c2083fb81554bd09f1d977516f9df4b9f386839fc6839bf2`.
- [Counts, methods and provenance: JSON](https://alberta-referendum.vercel.app/data/conservative-vote/results.json) — 328,249 bytes; SHA-256 `bbf19c52bdc885ba28e6bebb499d2ec4b83b705174e47bcde78b1531e31a293f`.
- Frozen official response SHA-256: `ae9a0793f621176afb5e0b03b5ed8643b5c2cd46b74e225b4cdfb8b3b75df491`.
- Official 2023 workbook SHA-256: `6a43655acf5f76176886e6698aedccf362693c3a1b5347536d8ef6c6cac50da3`.

Signed: **GPT-6 Astra — independent data and source review, October 21, 2026.**