Same acronym, different institution
In 2014, the CDC tracked a measles outbreak in Ohio’s Amish community — 383 cases. Epidemiologists arrived within seventy-two hours, sequenced the virus, traced it to the Philippines, and ended the outbreak in four months. Competent, transparent, life-saving. In 2020, the same CDC told the country masks were not effective, reversed within weeks, and spent the next year producing guidance that contradicted itself on timelines too short for any reasonable person to track.
Same institution. Same letterhead. Different questions, different incentive structures, completely different reliability.
The category error both sides make
The trust defaulters treat both episodes as confirming the same conclusion. We consider that reasoning error roughly as dangerous as the institutional failures it responds to. The person who refuses a safe vaccine because the approving agency also approved a dangerous drug is making the same error as the person who takes a dangerous drug because the agency also approved safe vaccines. Both collapse a complex landscape into a single binary.
Our framework is boring. Evaluate each institution on each topic. The CDC’s track record on outbreak investigation is excellent. On politically charged communication, poor. The 5:1 media lean predicts bias on contested political questions. It predicts nothing about hurricane tracking.
The rational skeptics present a numerator without a denominator. The FDA also approved effective vaccines and antiretrovirals. The repair advocates are doing the work of building what we describe — institutions with transparent, domain-specific trust records.
Where we concede ground: Our framework is cognitively expensive — a luxury unavailable to most people making real decisions.
What would change our mind: Disaggregated evaluation producing no better outcomes than simple trust-or-distrust heuristics.
Read the full synthesis: When is distrusting institutions the rational move?