Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

Insurance Denial's Gender gap tables hide that bias direction reverses between the biased and mitigated model (same bug class as the already-fixed #681)

Closed Beginner friendly
  • Is something wrong?
#695 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
1/5
Estimated time
Under an hour
Newbie friendliness
90/100
Issue type
Documentation
Clarity
Clearly specified
Activity status
Active
Tech stack
markdown, python
Domain
documentation, machine-learning

Research direction

The issue is in three markdown files: README.md lines 74-75, the Audit 04 detail section, and Insurance Denial/README.md line 30. Run the provided python commands (python3 "Insurance Denial/unfair.py" and python3 "Insurance Denial/fair.py") to verify the current signed gap numbers. Update the tables to show the signed gaps (e.g., "-5.44% → +1.54%") or add a "direction flips" note, following the fix from issue #681. Check the changes by viewing the rendered markdown.

Written by the indexing model from the issue text.

Description

documentation

Where: README.md lines 74-75 (Results at a Glance, row 04) and the Audit 04 detail section; Insurance Denial/README.md line 30.

The gap: All three tables show only the unsigned gap magnitude - "Gender: 5.44% → 1.54%", "72% reduction" - which reads as a straightforward shrinking gap. It actually reverses which group is disadvantaged: in the biased model females are flagged high-cost at a higher rate than males; in the mitigated model males are flagged at a higher rate than females. This is the exact bug class #681 already fixed for Healthcare Readmission's Gender row (which now explicitly shows "-0.02% → +0.04%" / "direction flips" instead of a bogus reduction %) - but Insurance Denial's Gender row was never given the same treatment.

Repro:

$ python3 "Insurance Denial/unfair.py" | grep -A3 "by Gender"
── High-Cost Claim Flag Rate by Gender ───────────────
  female               43.85%
  male                 38.41%

  Fairness Gap (Gender): -5.44%

$ python3 "Insurance Denial/fair.py" | grep -A3 "by Gender"
── High-Cost Claim Flag Rate by Gender ───────────────
  female               48.46%
  male                 50.00%

  Fairness Gap (Gender): 1.54%

Biased model: female (43.85%) > male (38.41%) - females disadvantaged. Mitigated model: male (50.00%) > female (48.46%) - males disadvantaged. The disadvantaged side flips, but README.md's Results-at-a-Glance row 04, the Audit 04 detail section, and Insurance Denial/README.md's own reproducibility table all still present it as an unsigned "72% reduction."

Fix direction: Show Gender signed ("-5.44% → +1.54%") in all three locations, following the precedent #681 already established for Benefits Denial's Age row and Healthcare Readmission's Gender row, or add an explicit "direction flips" note instead of a reduction percentage.

Dominant language
HTML
Stars
47
Forks
45
Avg merge
1h 54m
Merged PRs (30d)
97

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from yakew7/Fair-Code

All issues in yakew7/Fair-Code

Similar issues

More Documentation issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.