Refresh stale README create_mock_data() example output

[DougManuel]

Follow-up from the PR #48 (#38 seed-contract) multi-agent review (doc-staleness agent).

The README.md create_mock_data(..., seed = 123) example against the packaged minimal-example CSVs hardcodes #> output showing 3 columns (age, smoking, interview_date). Running it now yields 6 columns (age, smoking, BMI, height, weight, interview_date) with different values and a proportion-normalization warning. Two causes:

1. Pre-existing: the minimal-example metadata grew extra variables (predates [v0.5] ADR: seed and RNG-stream contract across generation stages #38).
2. [v0.5] ADR: seed and RNG-stream contract across generation stages #38: seeding switched Mersenne-Twister -> L'Ecuyer-CMRG, so the literal digits change regardless.

README.md is plain Markdown (no Rmd/qmd source — output is hand-pasted), so it needs a manual refresh. The two create_con_var() examples above it were verified still-accurate (those helpers remain on plain set.seed()).

Fix: rerun the example against current minimal-example on a v0.5 build, paste the real output, and either show the columns as-is or trim the example. Low priority, docs-only. Consider a lightweight render check so README code output doesn't silently drift again.