Why an honest account of an accident can sound like an excuse

Two ways to explain the same event

Picture a hospital nurse who gives a patient the wrong dose. Asked what happened, the nurse could answer in one of two ways.

Account A	Account B
"I made a mistake. I gave the wrong dose."	"It was my third double shift. Two drugs with similar names sat in the same drawer. The label was hard to read, the ward was short two staff, and I was interrupted while drawing it up. I didn't catch it."

Account A sounds honest and humble. Account B sounds, to many ears, like someone pointing at everything except themselves.

Yet Account B is the one a safety investigator wants. It names the conditions that will cause the same error again, with a different nurse, next month. Account A fixes nothing.

Bias one: we blame the person, not the situation

In a classic 1967 experiment, Edward Jones and Victor Harris had people read essays for or against Fidel Castro. Readers were told the writers had been assigned their side. They still judged that each writer believed what they wrote.

The psychologist Lee Ross later named this the fundamental attribution error: when we explain someone else's behaviour, we lean on their character and underweight their circumstances. A 1995 review by Daniel Gilbert and Patrick Malone found this "correspondence bias" across many settings.

So when an account stresses circumstances, it runs against the listener's default. The listener is already looking for a flaw in the person.

Bias two: hindsight makes the outcome look obvious

In 1975 Baruch Fischhoff gave people descriptions of historical events. Some were told how the events ended; others were not. Those told the outcome rated it as more likely, and did not realise that knowing it had changed their judgement.

After an accident, everyone knows how it ended. The warning signs look bright. The listener wonders how anyone missed them, and "I didn't see it" sounds implausible, even when it is true.

Bias three: a bad outcome makes the decision look bad

In 1988 Jonathan Baron and John Hershey showed people the same medical decision, made on the same information. When told it ended badly, people rated the decision itself as worse, and the decision-maker as less competent, than when told it ended well.

The worse the harm, the harder it is for a listener to accept that the choices leading to it were ordinary ones.

Bias four: the wrong emotion reads as guilt

A study published in 2003 showed people a video of a witness giving testimony, played by an actress in several versions. The words stayed the same; only the emotion shown changed. When the witness appeared calm or neutral, viewers found the account less credible than when it came with the emotion they expected. The content of the story mattered less than the display.

People who explain events analytically, in a steady voice, can come across as cold, and cold can read as guilty.

What safety science does instead

Fields where errors kill, like aviation and medicine, learned to work against these biases. The psychologist James Reason, writing in the BMJ in 2000, described two approaches to error:

• The person approach blames individuals "for forgetfulness, inattention, or moral weakness". Reason noted that blaming individuals "is emotionally more satisfying than targeting institutions".
• The system approach starts from the fact that "humans are fallible and errors are to be expected, even in the best organisations". In his words: "We cannot change the human condition, but we can change the conditions under which humans work."

His Swiss cheese model pictures each safeguard as a slice with holes. An accident happens when the holes in several slices line up. Account B above is a list of holes. The safety scientist Sidney Dekker's idea of a "just culture" builds on this: people should be able to report what went wrong without the report itself being treated as a confession.

How to listen to an account of an accident

1. Notice when you are judging the person before you have looked at the situation.
2. Ask what the person knew at the time, not what you know now.
3. Judge the decision on the information available, not on how it turned out.
4. Don't treat calm as coldness. People show distress in different ways, and some not at all.
5. Treat a list of causes as data. Check each one. If they are true, they are the reasons it will happen again.

How to give one

• Start by owning your part in plain words: "I gave the wrong dose."
• Then give the conditions, and say why: "I'm telling you this so it doesn't happen to someone else."
• Offer evidence for each condition where you can: rosters, photos, records.