Cognitive Psychology: Human Error Kevin Paterson

1
Cognitive PsychologyHuman ErrorKevin Paterson
2
Introduction

• Topics
• Why human error is important
• Limits on human behaviour
• Attention, Perception, Memory, Reasoning
• Design issues
• Psychology, health and safety

3
Why human error is important

• Human error is a major cause of accidents in the
  workplace, at home, or on the roads
• Estimated that 90 of all workplace accidents are
  due to human errors.
• Many disasters have been attributed to human
  error Bhopal, Three-mile Island, Challenger
  space shuttle, Chernobyl.

4
Limits on human behaviour

• Attention
• There are limits on our ability to divide
  attention.
• Information overload can result in workplace
  errors.
• Over-learning responses (i.e. automatic
  processing) also can lead to errors.
• e.g. PC trashcan errors.

5
Limits on human behaviour

• Perception
• We need to interpret information that we see.
• Missing some pieces of information can result in
  error.
• One solution is to build redundancy into key
  information - i.e. have multiple cues or signals
• e.g. dual coding for electrical wires.

6
Limits on human behaviour

• Memory
• Limited short-term memory capacity.
• Critical for learning new information,
  procedures.
• Context dependency of learning and recall
• levels of processing - depth of learning is
  important - critical for training.

7
Limits on human behaviour

• Logical reasoning
• People are poor at systematic or logical
  reasoning
• Yet systematic reasoning is often required in
  error checking procedures
• Reasoning errors responsible for major accidents,
  
• e.g. three-mile island nuclear incident.

8
Swiss cheese model of human error
9
Systems design

• Mapping onto users mental model
• Managing information
• Reducing complexity
• Visibility
• Constraining behaviour
• Design for errors
• Standardisation
• User centred design

10
Systems design

• Mapping onto users mental model
• Often there is a gulf between how a system is
  working and how a user thinks a system is
  working.
• E.g. three mile island - lights indicated that a
  system had been instructed to shut-down, user
  thought it indicated actual status of the system.
• Design should ensure that system operation is
  obvious.

11
Systems design

• Managing information
• Avoid relying on memory
• User might forget to perform task, e.g. mechanics
  often forget to replace nuts etc.
• Use systems (e.g. pocket PC with interactive
  checklist) to monitor and provide reminders.

12
Systems design

• Reducing complexity
• Greater task complexity results in errors
• use systems to reduce complexity
• E.g errors in Health care often occur as result
  of prescription errors.
• Using systems that reduce reliance on
  prescribers knowledge can reduced errors.

13
Systems design

• Visibility
• User should be able to perceive what actions are
  possible and which might be best.
• Feedback is important to demonstrate that action
  occurred and whether it was successful.

14
Systems design

• Constraining behaviour
• Reduce errors by designing the interaction in
  ways that minimise their opportunity.
• E.g. cash machines used to give out money before
  card
• people used to forget to collect card
• e.g. cash machines used to give out money before
  card
• people used to forget to collect card
• now they give card before money

15
Systems design

• Design for errors
• Errors are inevitable so design should anticipate
  them.
• E.g. On PCs you have to click an OK box to delete
  contents of trashcan.
• However, often use OK boxes to confirm action.
• Should use non-standard response box for critical
  decisions to prevent accidental selection.

16
Systems design

• Standardisation
• Reduces learning time and confusion.
• E.g. ER units use different procedures in
  different hospitals
• However, considerable swapping of staff
• Increases likelihood of confusion, re-learning
  and errors.

17
Systems design

• User centred design
• Occurs when systems design reflects user needs
  rather than designer needs.
• Sounds obvious, but designers often have
  different knowledge from user and do not
  appreciate their needs.
• Importance of usability assessment.
• Often empirical tests using sample of potential
  users.

18
Summary

• Demonstrated that human error is important
  because it is widespread and can be costly.
• Considered some causes of human error Attention,
  Perception, Memory, Reasoning
• Considered methods of reducing and addressing
  human error in systems design.