What about human error? Part 2 of 2

Last week I talked about the concept of "human error" — and especially about why we in the Quality business always insist that "There's no such thing as human error" when it's obvious that there is. I argued that while of course we all know that humans make mistakes, that's never the place to stop in an incident-investigation: focusing hard on human error makes your participants clam up if they are afraid of being blamed, and it shuts off the chance to find systemic improvements that could make future mistakes less likely. Another way to say this is to say that human error is a symptom but not a cause. Somewhere in your system, there is something else that triggered or allowed the human error to happen, and that's the thing you want to find and control.

But if human error is a symptom then we really need to understand what kinds of human error we might encounter, because each different kind of error is probably a symptom of a different cause and therefore has to be treated in a different way. If you go to the doctor because of pain, he'll treat you differently depending whether the pain is in your head or your elbow.

Fortunately this work has already been analyzed and tabulated. The information that I provide below comes from a website page owned and administered by the United Kingdom's Crown Health and Safety Executive, and I gratefully acknowledge permission to use it, as follows: This blog post contains public sector information published by the Health and Safety Executive and licensed under the Open Government License. The full text of the Open Government License for public sector information can be found here.

So, what are the different kinds of human error?

In the first place, a failure is either:

• Inadvertent (an error)
• Deliberate (a violation)

Errors can be either:

• Action errors (where our action is not as planned)
• Thinking errors (where our action is as planned but we planned the wrong thing)

Action errors can be either:

• Slips (where we do something wrong)
• Example: Flip a switch up instead of down.
• Example: Transpose digits during data entry.
• Lapses (where we fail to do something right)
• Example: Forget to signal before turning at an intersection.
• Example: Skip a step in a safety-critical procedure.

Thinking errors can be either:

• Rule-based mistakes (where we misapply a good rule or else apply a bad rule)
• Example: Misjudge passing the car in front of you, because you would have plenty of room if you were in your own car, but you are in your friend's car which has a lot less power.
• Knowledge-based mistakes (where we have no rules and try to figure it out from scratch)
• Example: Rely on an out-of-date map to plan your route through an unfamiliar town.

Finally, violations can be either:

• Routine (where there is no meaningful enforcement, so everyone ignores the rule)
• Example: A lot of cars on the freeway ignore the posted speed limit.
• Situational (where we cut corners in certain cases, because of features specific to those cases)
• Example: You document your design reviews scrupulously for new designs, but never document them for subsequent Engineering Changes because they always look so obvious you can't see taking the time to write them down.
• Exceptional (where we take a calculated risk in breaking the rules in order to address a highly-unusual situation)
• Example: A huge production order in your factory is due by Friday. On Tuesday, one of your machines comes due for preventive maintenance, but you keep it running (rather than shutting it down the way you are supposed to) because that's the only way to meet the production deadline.

That's the list.

Now what do we do about them?

Notice first of all that an approach which works for one kind of failure really won't help with another kind. Requiring someone to fill out a checklist is very likely to prevent lapses, but it will be completely useless in preventing (say) exceptional violations. A checklist, after all, helps the memory, and in case of a lapse the operator just forgot a step. But in the case of an exceptional violation, the operator knows perfectly well what steps he is choosing to skip — and why — so a reminder (in the form of a checklist) will make no difference at all.

The kinds of steps to consider are things like these.

You can download a useful reminder table, which summarizes all this information and more, from my Downloads Page.