The Middle Column
No side. Just the record.
Permanent and public. Entries are never deleted; a correction to a correction is a new entry. The first two entries predate launch: they come from the research work these briefs were built from, and we file them here because the standard applies to the work, not just to the website.
What was wrong: An early draft of the Medicare for All explainer stated employers pay “$17,000 or more” per employee toward family coverage, taken from a secondary summary.
What the record shows: KFF’s 2025 Employer Health Benefits Survey puts the average family premium at $26,993 with workers contributing $6,850, which puts the employer share at about $20,143 on average.
What changed: The corrected figure is used in Brief MC-0002. Lesson recorded: secondary summaries do not carry load-bearing numbers; the primary survey does.
What was wrong: An early draft converted average weekday boardings into a share of residents, concluding roughly 1 in 10 city residents use The Rapid daily. Boardings are not unique riders: one person can board several times a day, and no federal source publishes a unique daily rider count.
What the record shows: The FTA counts 6,042,030 unlinked passenger trips in 2023 across a 621,711-person service area, about 9.7 boardings per resident per year. That measures use, not users.
What changed: The conclusion was withdrawn in the original exchange and is not used in Brief MC-0004, which carries the boardings-versus-riders distinction in its glossary. Lesson recorded: a numerator and a denominator must count the same kind of thing before they are divided.