Aglet

Verify Weighted Average Report Measures

Weighted-average verification checks arithmetic and meaning together. Use a fixture where weighted and unweighted calculations diverge, then apply one filter that removes a high-weight row. Compare source inputs, intermediate contributions, displayed results, and export labels so a correct equation still carries the right explanation.

Check whether the outcome improved

  1. Define arithmetic expectations

    Write expected numerator, denominator, excluded rows, precision, and displayed result for the uneven fixture. State how missing, zero, and negative weights are handled or marked unsupported. A pass requires the metric name to reveal whether the result is weighted or unweighted.

  2. Exercise population changes

    Run the same measure for the full fixture, a filtered subset, and a group-level view. Compare weighted and unweighted results, denominator changes, and row membership. Inspect a chart, card, table, and export for consistent values and labels.

  3. Review interpretability

    Ask a reviewer to identify the observation unit, weight source, and denominator from the report or export. Preserve the fixture, calculation trace, filter state, and revision. Record a bounded result when a surface omits the information needed to explain its average.

What to carry forward

Verification passes when the chosen average matches its documented arithmetic and a reviewer can identify its unit and denominator. Report a bounded pass for unsupported weight cases. Keep verification inconclusive if a filter changes the weight rule without changing the metric label.

Technical background: PostgreSQL documentation.

Keep the decision with the work.

Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.

Create an account See the product workflow