What an admission opens

If a pilot breaks an aviation rule by accident and reports it to NASA within ten days, the Federal Aviation Administration can still find a violation but, as a rule, will neither fine the pilot nor suspend their certificate.¹ The agency gives its reason in plain regulatory prose: it treats the report as "indicative of a constructive attitude", an attitude that "will tend to prevent future violations."¹ NASA, which runs the reporting system, uses the reports to find deficiencies in the aviation system and to send alerts to the people in a position to correct them.²

That is an old maxim, that humility earns honour, written as policy and given a reason. On that reading, humility is honoured less as a display of virtue than because it opens a channel: once you have said where you went wrong, other people can tell you what they see, and you are committed to looking. The respect is for what the channel produces, and it arrives later than the admission. The rule also draws a boundary, in the same place the psychology does. The admission pays for the mistakes that correction can reach, and not for failures of intent.

Asking for help and admitting errors threaten face, and people at work are often reluctant to disclose their errors even when doing so would help their team, as Amy Edmondson summarised the research in 1999.³ In an earlier study of hospital patient-care teams she found teams whose members openly acknowledged medication errors and discussed how to avoid them, and teams whose members kept a drug error to themselves.³ A nurse in one team said, "you're never afraid to tell the Nurse Manager"; a nurse in another team in the same hospital said that people there "get blamed for mistakes".³

Edmondson argued that team members watch the leader in particular, and that a leader who coaches rather than punishes, and who visibly engages in learning behaviour, shows that the risk is safe to take.³ The first admission that makes the others cheap is usually the leader's.

The best units looked worst

The hospital study also produced a result that reads like a scandal. The units with stronger nurse-manager coaching, better relationships and higher perceived performance had significantly higher rates of medication errors.⁴ Edmondson's explanation was that a detected error rate depends both on the errors made and on how many of them are reported.⁴ In her later study of 51 work teams at a manufacturing company, teams whose members felt safe engaged in more learning behaviour, and that behaviour in turn affected how well they performed.³

This is the first cost of humility, and it is paid in reputation. A unit that admits its errors looks, on paper, like the worst unit in the building until the corrections show. Anyone who ranks teams by reported mistakes punishes the behaviour that makes improvement possible.

Where the line falls

The FAA's waiver has conditions. The violation must be inadvertent and not deliberate; it cannot involve a crime, an accident, or an act that discloses a lack of qualification or competency; and the person must not have been found in violation in the previous five years.¹ The waiver is for the slips of people who are qualified, not for discovering that someone is not.

Experiments on trust draw the same line. Peter Kim and colleagues found that trust was repaired more successfully when a person apologised for a violation that concerned competence but denied culpability for one that concerned integrity.⁵ An apology also worked better when evidence of guilt emerged afterwards, and a denial when evidence of innocence did.⁵

Admitting a competence error says: here is something I can learn, tell me when you see it. Admitting a failure of integrity says something about what you wanted, and no one else's report can fix that. Self-abasement fails for a related reason. A general confession of unworthiness points at nothing specific, so it gives no one anything to report.

The late payment

When the payment comes it can be large. Since 2001 the University of Michigan Health System has disclosed medical errors to patients and offered compensation.⁶ After the programme was fully in place, new claims fell from 7.03 to 4.52 a month per 100,000 patient encounters, lawsuits fell from 2.13 to 0.75, and liability costs fell as well.⁶ The study cannot establish that disclosure caused the fall: malpractice claims were declining across Michigan, and the system's closed staff and its own insurer make it unusual.⁶ At the least, telling patients about errors and offering to pay did not increase the health system's claims or costs.⁶

The pattern across a cockpit, a ward and a hospital's legal office is the same. An admission is cheap to make only where someone has decided it should be, and it is honoured only after it has had time to work. That makes humility less a trait than an arrangement between the person who admits and the people who could punish them. The honour is what both sides get when the arrangement holds.