S03 · Source School
Build the Countercase
What is the strongest honest reason the measure may be right for its job?
A critique that survives only while the methodology stays closed is not much of a critique. Open the method. Give the statistic its best case. Then see what remains.
01
Steelman the assigned job
State why the population, weighting, adjustment, replacement, or revision exists. Use the publisher's explanation without borrowing its institutional prestige as proof.
02
Name the real tradeoff
A fixed sample becomes stale. A changing sample requires comparability judgments. A current series may sacrifice precision that arrives only with later data. Put both costs on the table.
03
Keep the surviving criticism
After the countercase, narrow the claim to the unresolved mismatch: a different population, access question, distribution, time horizon, or lived quality not established by the current record.
Try to disprove yourself
The strongest honest countercase
Sometimes the method survives and the apparent bend disappears. Publishing that result is part of the method, not a failure to find a story.
A question to carry
Where have you seen this number travel too far?
Choose one criticism of CPI. Write the strongest BLS response first. Keep only the part of the criticism that remains supported afterward.
Transfer test
The hospital score
A hospital's average wait time fell, but the reported population excludes patients who left before treatment.
What moved outside the denominator?
Open the field note
The excluded patients may change the interpretation. Confirm the denominator and missingness rule before claiming the typical patient's wait improved.
Direct sources
Inspect the method at BLS
Sources reviewed August 30, 2026. These links establish the bounded methodological statements used above. They do not make BLS responsible for bent.WTF’s instructional framing.