# F4-dev-01 — Headline figures and one assertion, Statistics Canada LFS September 2026

## Question
From the provided text of Statistics Canada's *The Daily* release "Labour Force Survey, September 2026" (released 2026-10-09): (a) extract the national September 2026 unemployment rate (seasonally adjusted, %) and the monthly change in employment (seasonally adjusted, persons), each with a section locator; (b) decide whether the following assertion is supported: "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%."

## Data and exclusions
- `data/sources/statcan-daily-lfs-2026-09.txt` (164 lines; SHA-256 matches the value in `data/README.md`), and `data/README.md` (9 lines). These were the only evidence used. I did not use remembered LFS values.
- Charts, infographics, maps and Tables 1–12 are not in the extract; only their headings remain. All values below come from narrative sentences. Nothing was excluded beyond that.

## Method
`out/analysis.py` parses the extract line by line and records the most recent `##`/`###` heading for each line. Regular expressions then pull out the figures exactly as printed. The script prints every matched source line, so each value can be traced to the text:
- Unemployment rate: Highlights, line 17, "The unemployment rate increased 0.1 percentage points to 6.5%." Line 43, under "Unemployment rate increases to 6.5%", gives the same value ("rose 0.1 percentage points to 6.5% in September").
- Employment change: Highlights, line 17, "Employment declined by 68,000 (-0.3%) in September". "declined" is coded as a negative sign. Line 24, under "Employment falls for second consecutive month", gives the same value.
- Assertion: I split the assertion into six parts (direction, size of change and level, for women and for men). Each part was checked against lines 74–75 under "Unemployment rate up for core-aged women".

## Results
| item | value | locator |
|---|---|---|
| Unemployment rate, Sept 2026 | 6.5 % | statcan-daily-lfs-2026-09 \| section "Highlights" |
| Employment change, Sept 2026 | −68,000 persons | statcan-daily-lfs-2026-09 \| section "Highlights" |
| Assertion verdict | **not_supported** | statcan-daily-lfs-2026-09 \| section "Unemployment rate up for core-aged women" |

The release supports the women's part of the assertion: "the unemployment rate rose 0.3 percentage points to 5.3%". For men, the size of the change (0.2 points) and the level (5.8%) match, but the direction does not. The release says "the unemployment rate for core-aged men **fell** 0.2 percentage points to 5.8%", and it explains why: employment held steady while the number of men searching for work edged down. The assertion says the rate "also rose", which the release directly contradicts, so the verdict is `not_supported`.

## Uncertainty
These are point estimates copied exactly as printed; I did no estimation and no re-computation. The release notes that it treats differences as statistically significant at the 68% confidence level. The extract does not give sampling errors for these figures, so I cannot report any. The verdict rests on explicit wording ("fell" against "rose"), and the extract leaves no doubt about which is meant.

## Limitations and open issues
- The data tables are not in the extract, so the figures could not be checked against Table 1 or Table 3. The only checks possible were between sections of the text, and those agree.
- The "Sustainable Development Goals" section ends with "the following goals:" and the list itself is missing from the extract. This does not affect any of the results.
- Both headline figures appear in two sections. The locators name "Highlights", but the thematic sections given in the Method give identical values.