# F4-dev-01 — Two headline figures and one assertion from a Statistics Canada release

## Scientific question
Using only the provided text of Statistics Canada's *The Daily* release "Labour Force Survey, September 2026", (a) extract two headline national figures for September 2026 — the unemployment rate (seasonally adjusted, in percent) and the monthly change in employment (seasonally adjusted, in persons) — each with a locator naming the section of the release where it is stated; and (b) decide whether the following assertion is supported by the release: "In September 2026, the unemployment rate among core-aged women (aged 25 to 54) rose 0.3 percentage points to 5.3%, and the unemployment rate among core-aged men also rose, by 0.2 percentage points, to 5.8%." The estimands are the figures as printed in the release (no re-computation) and a verdict of `supported` (every element of the assertion is stated in the release), `not_supported` (at least one element is contradicted by the release) or `not_assessable` (the release does not report the quantity needed to judge it).

## Inputs
- `data/sources/statcan-daily-lfs-2026-09.txt` — plain-text extract of Statistics Canada, *The Daily*, "Labour Force Survey, September 2026", released 2026-10-09 (Catalogue no. 11-001-X), retrieved 2026-10-30T03:25:42Z from https://www150.statcan.gc.ca/n1/daily-quotidien/261009/dq261009a-eng.htm; Statistics Canada Open Licence. The extract contains the provenance header, the release title and date, and the full narrative text from "Highlights" to "Next release" (section headings preserved as `## …` lines; charts and data tables are not reproduced). Estimates are seasonally adjusted unless the text says otherwise; employment changes are in persons.
- `data/README.md` — provenance, retrieval times, licence and SHA-256 of the extract.

## Information-availability rules
- The extract is the only admissible evidence. Do not use remembered values of the Labour Force Survey or any other release.
- A locator must name the section heading of the release (the `## …` line) under which the figure or sentence appears, e.g. `statcan-daily-lfs-2026-09 | section "Highlights"`.
- Report printed values exactly (sign and decimals as printed; convert "68,000" to the integer 68000 with the sign given in the text).

## Required outputs
- `out/results.json` with the keys:

| key | type | unit | meaning |
|---|---|---|---|
| `unemployment_rate_sep2026` | number | percent | national unemployment rate for September 2026 as printed |
| `employment_change_sep2026_persons` | integer | persons | monthly change in national employment in September 2026 (negative if employment declined) |
| `locators` | object | — | {"unemployment_rate_sep2026": "<source id | section heading>", "employment_change_sep2026_persons": "<…>"} |
| `assertion_verdict` | string | — | "supported", "not_supported" or "not_assessable" for the assertion quoted above |
| `assertion_locator` | string | — | source id and section heading of the sentence(s) that decide the verdict |
| `assertion_reason` | string | — | one sentence explaining the verdict, quoting the decisive element |
| `task_id` | string | — | "F4-dev-01" |
| `feasibility` | string | — | "complete", "partial" or "infeasible" — refers to the task as a whole |
| `feasibility_reason` | string or null | — | explanation if not complete |
| `data_issues` | array of {"issue","handling"} | — | problems detected in the inputs (rendering defects, internal inconsistencies of a source, etc.) and how they were handled; empty list if none |
| `n_input_rows` | object | lines / items | {file name: number of lines} for every file under `data/` that you read (text files: number of newline-terminated lines as reported by `wc -l`; JSON files: number of top-level items) |
| `n_analysed` | integer or null | — | number of reported items (values, assertions, effects or references) you actually assessed |
| `random_seed` | integer or null | — | null (no randomness is involved) |

- `out/report.md` with the sections: Question; Data and exclusions; Method; Results; Uncertainty; Limitations and open issues.
- `out/analysis.py` — running `~/wp29/venv/bin/python out/analysis.py` from the workspace root must regenerate `out/results.json` from `data/` only, deterministically (text parsing is acceptable; hard-coding the numbers in the script is acceptable only if the script also prints the matched source lines).
- `out/claims.json` — a list of the central claims of the report: [{"id": "C1", "text": "...", "results_keys": ["..."], "status": "supported"|"unsupported"|"not_evaluated", "caveat": "..."}].

## Constraints
- No network access. Python: `~/wp29/venv/bin/python` (numpy, pandas, scipy, statsmodels available; the standard library is sufficient for this task). Do not modify `data/`.
- The source extracts under `data/sources/` are the **only** admissible evidence. Do not rely on memory of the papers or on other versions of them; if a quantity is not in the extract, it is not available.
- Report numbers exactly as printed in the source (same sign, same number of decimals); do not round or rescale printed values unless the key definition asks for a transformation.
- This task is designed to be completed in well under 20 minutes of work.