Week 10 - Human Error Flashcards
(14 cards)
Human error
Actions or decisions made by users that result in unintended or undesired outcomes.
Systems need to be designed to reduce potential for human error.
Generic error modelling system (GEMS)
A framework developed by James Reason to explain human error.
It categorises human performance into 3 main categories that can lead to errors in certain conditions.
3 main categories are skill-based, rules-based and knowledge-based performance.
Skill-based performance (GEMS actions)
Routine, automatic actions requiring little cognitive effort.
e.g., typing on a keyboard without looking.
Rules-based performance (GEMS actions)
Following established rules or procedures.
e.g., running a script to back up a database.
Knowledge-based performance (GEMS actions)
Using knowledge to address unfamiliar situations.
e.g., debugging a novel software issue without prior guidance.
GEMS error categories
GEMS classifies errors into 3 main categories based on the type of performance theyre associated
Slips (GEMS errors)
- “execution failure” where the user’s intentions are correct, but the actions are not carried out properly.
e.g., clicking the wrong button in a user interface despite knowing the correct one. - Related to skills-based performance.
Lapses (GEMS errors)
- “execution failure” often involving forgetting to perform an action.
e.g., forgetting to save a file after editing. - Related to skills-based performance
Mistakes (GEMS errors)
- “Planning failure”; an inappropriate set of actions is executed due to incorrect reasoning or rule selection.
e.g., using the wrong configuration settings when deploying a server. - Related to rule-based and knowledge based performance.
Advantages and disadvantages of GEMS
- Advantages: provides a structured framework for designing systems that minimise, detect, correct and tolerate human error.
- Disadvantages: Primary focuses on non-deliberate errors and overlooks deliberate actions like attempting to take shortcuts.
It is a high level approach that neglects the importance of contextual factors.
How should systems address human error
Design systems that minimise the potential for human error.
Create systems that detect and correct human error.
Develop systems that tolerate human error and mitigate its impact.
Challenges of addressing human error
- Humans are infallible - errors are unavoidable.
- Human behaviour varies significantly.
- General approach emphasises error-tolerance over error-avoidance.
Planning for human error
- Increase System Visibility: Avoid hiding complexity, make system behaviour transparent.
- Include Errors in Training: Train operators using error scenarios
- Design User-Centred Interfaces: Align interfaces with human behaviour.
- Design for Errors: Assume errors will occur, make actions reversible and recovery easy.
Cons of using automation to reduce errors
- Automation has limited scope - handles skill and rule based tasks but leaves complex knowledge based tasks to humans.
- Automation can reduce system visibility making errors harder to understand and troubleshoot.
- Errors are shifted from user/operator mistakes to design flaws which can be harder to detect and fix.