1. Introduction to Work and Organizational Psychology
Gerhard Ohrband
8th lecture
Workload and Task Allocation

2. Course structure Part I Introduction 1 Managing diversity
2 History and context for Work and Organizational Psychology / Roles and methods
Part II
People at work
3 Job Analysis and Design
4 Personal Selection
5 Training

3. Course structure
6 Performance Appraisal: Assessing and Developing Performance and Potential
7 Job Stress and Health
Part III
Human Factors at Work
8 Workload and Task Allocation
9 Work Environments and Performance
10 The Design and Use of Work Technology
11 Safety at Work

4. Course structure Part IV Organizations at Work
Leadership and management
Work motivation
Teams: the challenges of cooperative work
Organizational development (OD)

5. Part III – Human Factors at work 8 Workload and Task Allocation
Outline:
1. What is workload?
2. Why assess workload?
3. Requirements for effective workload assessment
4. Workload assessment techniques
5. The application of workload assessment techniques
6. Automation and task allocation

6. 1. What is workload?
performance problems associated with tasks high in explicit demand (i.e., the overload or cognitive strain scenario)
exploring/modelling errors at lower levels of the task demand spectrum

7. Physical vs. mental workload
Focus on mental workload
Examples: air traffic controllers, pilots, process control operators and medical staff
Primary tasks: cognitive processes which require memory, attention, perception and communication skills rather than extensive physical demands
Definition (Kramer, 1991): Mental workload can be conceptualized as the costs that human operators incur in performing tasks

8. Industries traditionally defined as high risk
Defence
Road transportation
Railways
Aerospace
Process control
Power generation

9. Workload and error
The complexity of the relationship between workload, task performance and task load can be illustrated with reference to the debate in which a number ofinvestigators have aimed to provide an answer to the question “How much workload is too much?” (e.g., de Waard, 1996; Meijman and O’Hanlon, 1984;Teigen, 1994).
To answer this question, investigators have found it useful todivide the Yerkes-Dodson inverted U function into 6 task performance-related regions as shown in Figure 8.

10. Task performance and workload as a function of demand

11. Single resource models

12. Kahneman’s capacity of attention

13. Performance Resource Function (PRF)

14. Multiple resource models
the human information processing system is a multiple channel processor
(e.g., it has multiple structures)
each processor, or group of processors, has its own internal capacity.
In MRT approaches, mental resources are often seen as analogous to fuel that is consumed by various activities, or as a tank of liquid to be divided among several competing tasks (Wickens, 1984).
In stressful conditions, or multitasking situations, the amount of resource may become depleted and give rise to interference effects. One important feature of this view is the idea that the impact ofchanges in task demand on mental resources may not be purely quantitative, but may also be qualitative as well, (i.e., structural).

16. Sources of mental resource relative to information processing stage

17. Diagrammatic representation of Wickens Multiple Resource Theory

18. Composite models
relatively new type of cognitive modelling activity that aims to use elements of the single channel hypothesis and resource theory to account for workload effects in human performance.
These composite models are usually represented in the form of computer programs that are applied to tasks and used to predict various aspects of human performance.
The ability of these models to reproduce workload effects has frequently been mentioned as a justification for their development – although the extent to which they have managed to do this has been a matter of some debate.
Perhaps the most frequently mentioned composite model in the workload literature has been the Executive-Process Interactive Control (EPIC) model developed by Kieras and Meyer, 1997

20. 2. Why assess workload? Solve practical problems in the workplace
Effects of workload on performance, well-being, health and safety
Design complex task environments that do not place disproportionate demands on the human operator
1. Safety critical systems such as air traffic control and aircraft cockpit design
2. Aid evaluation of the effects of automation or the introduction of new technology and other changes in the nature of work on individual well-being and health
3. Assessment of individual operators (selection or training)
Acute vs. chronic effects of workload

21. 3. Requirements for effective workload assessment
Sensitivity
Diagnosticity
Intrusiveness
Validity
Reliability
Acceptability
Applicability
Generality

22. 4. Workload assessment techniques
Subjective/self-report measures:
Cooper-Harper scale (Cooper and Harper, 1969)
NASA task load index (TLX) (Hart and Staveland, 1988)
Subjective workload assessment technique (SWAT) (Reid and Nygren, 1988)
Instantaneous self-assessment technique (ISA)
Situation awareness rating scale (SART) (Taylor, 1989)
Situation awareness global assessment technique (SAGAT) (Endsley, 1995)

24. NASA Task Load Index (TLX)

25. NASA-TLX Scoring example

26. Subjective Workload Assessment Technique - SWAT

27. Time Based Task Loading Models

28. Performance measures Primary task measures Secondary task measures
Primary task: the task or system function whose workload is to be measured
Secondary task: one task that is performed concurrently with a primary task to investigate the workload associated with the latter

29. Physiological Measures
cardiac function
brain function
respiration
eyeblinks
pupil dilation
urine
blood
saliva (hormonal and immunological changes)

30. 5. The application of workload assessment techniques Summary of the capability of different broad categories of workload assessment techniques to satisfy different requirements for use
Intrusiveness
Sensitivity
Diagnosticity
Applicability
Acceptability
Subjective Measures
Post-task measures
Generally not intrusive
Good but may depend on length of task
Generally difficult to use diagnostically
Minimal equipment requirements
Very good
Instantaneous measures
Potentially intrusive
Good
Provides only a global measure
Some equipment required
good
Performance measures
Primary task
Not intrusive
Reasonable but difficulties in interpreting variation
Poor
Depends on task complexity and variability
Should be acceptable to operator
Secondary task
Potential for intrusion
May require training and extra equipment
Additional demands may be distracting

31. Physiological measures
Intrusiveness
Sensitivity
Diagnosticity
Applicability
Acceptability
Physiological measures
EEG measures
Not usually a problem
Good
Varies according to specific measure but reasonably good
Extensive equipment and analysis requirements
Some potential problems
ECG measures
Not intrusive
Not fully established
Not generally found to be problematical

32. 6. Automation and task allocation
advancements in computer technology
Wickens (1992): three reasons for automation
1. Situations which may be hazardous or dangerous to humans or which humans cannot perform (diving operations, handling of toxic materials)
2. Tasks with high levels of workload for the human operator (autopilots)
3. Overcome human limitations, e.g. in memory or attention (radar advance warning systems)

33. Dangers of automation
Thoughtless design which may simply require the operator to perform those functions or tasks that have been unable to be automated
Difficulties:
1. Automation may induce feelings of loss of control and situation awareness when the human is operating “out of the loop”
2. Risk of deskilling the operator in highly automated systems
3. Automation does not always lead to improved performance and levels of operator workload

34. Unscheduled manual interventions (Hockey and Maule, 1995)
operators overriding automated systems in order to assume control of production or other processes at times when the system is scheduled to be under automatic control
Why?
1. desire for control
2. motivation to improve the speed and quality of production
3. low trust in the automated system

35. Selection of the appropriate type and degree of automation
Goal: achieving desirable levels of safety and effectiveness
Question: Which tasks should be allocated to the operator and which to be automated part of the system?
Task allocation/function allocation
Traditional approaches: Fitt’s list (1951, Kantowitz and Sorkin, 1987)
Think about it: Where are machines and where are humans more effective?

36. Answer:
Machines: performing mathematical and computational operations, integrating information and dealing with predictable events reliably
Human operators: making decisions, inductive reasoning, more flexible, particularly when unexpected events occur and possess experience of previous events

37. Dynamic task allocation (DTA)
Def.: flexible allocation of tasks or functions between the operator and the system in human-machine systems; sometimes referred to as adaptive control
some or all of the task elements have the potential to be carried out by either the operator or the system itself
control of task allocation: explicit and implicit allocation
explicit: operator control, implicit: computer control of task allocation

38. Advantages of DTA
1. Workload of operators will be maintained at a relatively constant level
2. Resources of the systems (both human and computer) will be used more fully
3. More acceptances by the operators than static automation
4. Enhancement of situation awareness and prevention of decay of manual control and problem-solving skills which may be required in breakdown or emergency situations
5. Enhancement of the operator’s ability to diagnose failures and errors made by the computer

39. Discussion Points
What are the different workload assessment techniques?
Explain the hazards of automation.
Discuss the objectives of workload assessment.

40. Literature
Hockey, G.R.J. and Maule, A.J. (1995). Unscheduled manual interventions in automated process control. Ergonomics, 38,
Kramer, A.F. (1991). Physiological metrics of mental workload: a review of recent progress. In D.L. Damos (ed.), Multiple-Task Performance. London: Taylor and Francis