Lecture 2 'The Problem with Human Error'

Question 1

What is the definition of ‘human error’?

Answer 1

inappropriate human behaviour that lowers levels of system effectiveness or safety

Question 2

What are the four classifications of human error? Describe the intentionality of each one

Answer 2

1. Slips - unintended incorrect act
2. Lapses - unintended failure to act
3. Mistakes - intended, knowledge/rule-based
4. Violations - intended non-compliance (benign/malign)

Question 3

What kind of factors are associated with accidents?

Answer 3

1 - employee related
2 - job-related
3 - equipment/tool relared
4 - physical environment
5 - socia;/psychological environment

Question 4

When designing warning labels, what three things do you need to consider?

Answer 4

People must be able to NOTICE/SEE the warning
People must be able to READ the warning
People must COMPLY with the warning

Question 5

According to a study by Nevo, healthcare practitioners were generally worse at remembering to wash their hands coming into or leaving an examination room? What was the best method for compliance?

Answer 5

Leaving (because it was out of sight)

Having a warning label = best. However, long term effect may be eliminated through habituation

Question 6

What are two problems with defining human error as being caused by ‘inappropriate human behaviour’?

Answer 6

1. “inappropriate” human behaviour often RAISES levels of system effectiveness or safety
2. “appropriate” human behaviour can cause accidents by lowering system effectiveness or safety

Question 7

Explain the Swiss Cheese model

Answer 7

Accidents are caused by multiple factors which have ‘gone through’ layers of defence put in place by an organisation. Despite multiple layers, holes may still line up enabling error trajectory to pass through the holes

Question 8

In the Swiss Cheese model, error trajectory passes through the holes, what is this called?

Answer 8

latent conditions

Question 9

What are six limitations of the Swiss Cheese model?

Answer 9

• sometimes only in hindsight
• cannot predict future failures
• metaphor only (what are “holes”?)
• focus on defence = more defence not necessarily safer
• suggests all accidents are preventable
• current thinking: accidents caused by processes being too complex to be fully predictable and therefore fully preventable

Question 10

Aircraft carriers are considered to be systems full of holes, but yet it works. What are two reasons for this?

Answer 10

flexibility

adaptability

Question 11

Resilience engineering focuses on what to combat what?

Answer 11

how good human adaptability and variability is - to combat current thinking of “blame person/system > fix person/system”

Question 12

What is the difference between Safety I and Safety II?

Answer 12

Safety I - assumption that safety could be established/maintained by keeping human performance within boundaries or norms. “Error” is something that can be counted

Safety II - people can actually be positive to safety by being adaptive / may address gaps in system design / unplanned situations. “Error” isn’t as important as needed a theory of action including an account of performance variability

Question 13

What are the 10 misguided questions about error? HARD BONUS QUESTION

Answer 13

1. Was it mechanical failure or human error?
2. Why do safe systems fail?
3. Why are doctors more dangerous than gun owners?
4. Don’t errors exist?
5. If you lose situation awareness, what replaces it?
6. Why do operators become complacent?
7. Why don’t they follow procedures?
8. Can we automate human error out of the system?
9. Will the system be safe?
10. Should people be held accountable for mistakes?

Question 14

Compare the old vs. new view of whether accidents occur because of mechanical failure or human error?

Answer 14

Old = If no mechanical failure, must be human error (error is satisfactory to explain accident)

New = Error as a starting point for investigation (not the cause)

Question 15

What is the formula for human error?

Answer 15

human error = f (1-mechanical error)

Question 16

2. What does Dekker (2005) suggest is a reason why a safe system may fail?

Answer 16

Efficiency often takes precedence over thoroughness

Question 17

3. Why does Dekker (2005) believe it is misguided to ask why doctors are more dangerous than gun owners?

Answer 17

Because counting errors / looking at statistics will not help to understand system safety

Question 18

4. Why does Dekker (2005) believe it is misguided to ask if errors exist?

Answer 18

An ‘error’ is only considered so by the outside. Humans attribute the cause to an ‘error’ of some sort to make sense of the world

Question 19

5. Why does Dekker (2005) believe it is misguided to ask what replaces situation awareness if it becomes lost?

Answer 19

Shared situation awareness is folk concept: There is no ideal level of “situation awareness”

Question 20

6. Why does Dekker (2005) believe it is misguided to ask why operators become complacent?

Answer 20

Automation complacency is a folk concept:

Question 21

7. Why does Dekker (2005) beleive it is misguided to ask why operators do not follow procedures?

Answer 21

Sometimes procedures do not work (e.g. power plant licencing test failed when plant used own operations; but when they followed procedure a major procedural loop was discovered)

Question 22

8. Why is it misguided to ask if we can automate human error out of the system? / Is it possible to automate human error out of the system? (Q. 12)

Answer 22

No because automation increases human workload (have to check automation constantly)

Question 23

Why does Dekker (2005) believe it is misguided to ask if the system will be safe?

Answer 23

Never a guarantee, system is always changing/adapting

Question 24

Why does Dekker (2005) believe it is misguided to ask if we should hold people accountable for their mistakes?

Answer 24

> Cannot punish people and have organisational learning at the same time

> Second victim issue: (e.g. nurse who committed suicide after mistake)

Question 25

(Q. 10) Why is “human error” not an accurate or useful concept for making complex sociotechnical
systems safer, according to many current human factors theorists?

Answer 25

“Human error” does not identify the aspects of a situation that make the activity that led to an
accident more likely

Question 26

(Q. 16) According to DEKKER (2006) “the operator lost situation awareness” is what kind of statement?

Answer 26

Only a description, using a theoretical construct, of how accidents sometimes happen

Question 27

(Q. 11) Selecting the wrong procedure to solve a problem in an industrial plant, but without knowing the
procedure is wrong, is an example of which kind of error?

A. Slip
B. Lapse
C. Mistake
D. Violation

Answer 27

(incorrect act + not intended) = Slip

Question 28

(Q. 8) Which is the best description of the message underlying Reason’s (1990) Swiss Cheese Model?

A. More “slices” of defence create more opportunities for errors-producing conditions to be
missed.
B. The model shows that the best way to reduce accidents is to prevent workers from making
active failures.
C. More “slices” of system defence make it more likely that errors will be averted.
D. A model of system defences can help analysts predict the build-up of latent failures.

Answer 28

C. More “slices” of system defence make it more likely that errors will be averted.

Question 29

(Q.9) Which of the following is a reasonable criticism of Reason’s (1990) Swiss Cheese Model?

a. Latent failures are difficult to identify in hindsight.
b. The model suggests that all accidents are preventable.
c. The model cannot illustrate the principle of “defence in depth”.
d. The model cannot distinguish between plausible “active triggers” for an accident.

Answer 29

b. The model suggests that all accidents are preventable (are they?)