Chapter 4 : Descriptive Approaches

1
Chapter 4 Descriptive ApproachesWhat workers
really do

• ? ? ?

2
Contents

1. Purpose
2. Descriptive approaches Current practice
3. Why descriptive approaches are not enough?
4. Existing techniques for getting around the
   task-artifact cycle
5. An alternative way of getting around the
   task-artifact cycle

3
Purpose

• Using Four cases
• To illustrate the important contributions of
  descriptive approaches
• To identify additional dimensions that must be
  considered in work analysis
• Descriptive approach
• One of possible means to investigate intrinsic
  work constraint, rather than an end itself
• Computer-based information systems should not be
  designed based solely on studies of currents. ?
  Formative Approach

4
Scope of descriptive approaches

• Descriptive WA is accomplished by conducting
  field studies that
• document the practical challenges that workers
  actually face on the job
• and the practices that workers have developed to
  cope with those challenges
• Europe VS America
• 30 years tradition in Europe
• Task VS Activity
• Task Official actions that are prescribed to
  workers. (Normative approach)
• Activity Informal actions that workers actually
  perform in practice (Descriptive Approach)
• There are many different perspectives in
  descriptive approaches ? 4 categories
• A case study from situated action
• A case study from naturalistic decision making
• A case study from activity theory
• A case study from distributed cognition

5
A case study from situated action

• Suchman (1987)
• Anthropologist
• ? ?? ????? phototype ?????? ????? photocopying
  machine? ??? ? ??? ???? interact??? ??
• Conversation ??
• ??? ??? ??? ??? ?? ??? ??? ???? ??
• Theoretical Foil
• Artificial intelligence approach to planning
• The same perspective to the instruction based
  approach

6
Suchman(1987) ? ??

• An instruction-based approach is constrained by
  limitation on the designers ability to predict
  any users action
• ????? ?????? ???? ???? ???? ?? ???? ?? ??? ??.
• ??? ????? ??????? ???? ?? ??? ???? ???.
• Suchman said that purposeful actions are
  inevitably situated actions
• Situated actions are responding to local
  interactions contingent on the actors particular
  circumstances
• ?? Actual behavior is far from the rational
  ideal

7
A case study from naturalistic decision making

• Klein(1989)
• Experimental Psychologist
• ????? ??? ??? ????? ???? ?? ???? ??????? ?? ??
• Retrospective naturalistic studies
• ?? 23?? ? ?????? ???? ????? ?? ???? ????? ??
• Theoretical Foil
• Classical decision making To follow a thorough
  and rational approach
• ?? ? ??????, ? ??? ??? ??, ??? ??

8
Klein(1989) ? ??

• ??? ??? ?? ??? ???? ??? ??
• Why experts go with the flow
• ????? ??? ?? ??? ??? ??? ? ?? ??? ?? Too
  laborious mental deliberations.
• Not to find optimal action, but to find
  satisfactory action (feasible, timely and
  cost-effective)
• ????? ????? ???, ?? ?? informative cues ? ??? ?
  ?? ????? ???. Action? ??? ??? ?? ??
• ?? ?? informative cues ? ??? ? ??, ? Cue?? ??
  ????? ????? ????? ??? ? ??. ??? ???? ??? ?? ????.
• ??? ??? ? ?? ???, ??? ??? ??? ??? ??? action?
  ???? ??.
• ??? ???? ??? ?? ??? ??? ??? ??? ?? ??? action?
  ??.

9
Klein(1989) ? ??

• Recognition-Primed Decision-making (RPD)
• ?????? ??? ????, ????? ????? ???? ???
  ??(recognize)? ? ??.
• ??? situation assessment? ?? ???? ?? ???.
• ?? Actual decision making is far from the
  normative approach prescribed by classic decision
  theory

10
A case study from activity theory

• Bødker(1991)
• Computer scientist
• The participatory design of computerbased
  systems
• UTOPIA Project
• ????? ??? ??? ?????? ???? ?? Descriptive task
  analysis? ?????.
• ??????? ?? ?? ?? ??, ????? ?? ? ??
• ? ???? ??? ??
• Theoretical Foil
• Anglo-American approach to HCI

11
Bødker(1991) ? ??
Activity theory-based HCI Anglo-American HCI
Context-based Context-based Little Context-based
Semantic/ Syntactic semantic features of the interface Syntactic and Lexical features of the interface
Usefulness/ Usability Usefulness (functionality) Usability
Perspective Ecological perspective Cognitive perspective

• She wanted to find a systematic framework
• Thus, to adopt activity theory originated from
  the psychology of Soviet union

12
Bødker(1991) ? ??

• Activity theory could be adapted to the practical
  needs of HCI and its concepts provided a good
  descriptive understanding of the idea collected
  during UTOPIA
• Activity theory is the study of goal-directed
  activity
• Instead of focusing on the device, goal-focus
• Goal focus? ??
• ???? semantic? ?? ?? ??? ???
• Goal-oriented activity? ??, ??? context??? ??? ??
  ????.
• Goal-oriented activity? ??? ??? ???? ???? ??,
  ???? ??? ???? ??? ???.
• ?????, context-conditioned variability

13
A case study from distributed cognition

• Hutchins (1995)
• Anthropologist
• Field study of current practice in the domain of
  ship navigation
• On the navigation bridge of a number of US Navy
  ships at sea
• To understand the nature of human cognition in
  naturalistic settings
• Theoretical Foil
• Physical symbol system hypothesis (PSSH)
• Such as digital computer, human cognition
  involves mental information processing driven by
  well-defined rules and representations stored in
  human memory.

14
Hutchins(1995) ? ??

1. Information processing are not confined to the
   brain, but are instead distributed spatially
   across individuals and artifacts and temporally
   as a function of the history of a particular
   culture.
2. Workers frequently accomplished task goals, not
   in isolation through mental information
   processing, but as a functional team through
   mutual coordination of their actions
3. Cognition in the wild is an emergent activity
   that is not completely specified ahead of time.
4. Historical influence of cultures were very
   important. Many useful artifacts and many of the
   practices were adapted products of navigational
   experience. ? Not a strict mental activity
   mechanically performed by an individual.

15
Comparison of four case studies
?? ???? ?? ???? ?? ?? ?? ??
Suchman (1987) Anthropology Photocopying Artificial Intelligence Situated Action
Klein (1989) Experimental Psychology Fire Fighting Classical decision making RPD Model
Bødker (1991) Computer Science Graphic Design Anglo-American HCI Activity theory-based HCI
Hutchins (1995) Anthropology Ship Navigation Physical symbol system Hypothesis Distributed Cognition
???? Simonsen and Kensing (1997), Suchman (1995)
16
Importance of current practice

• 4?? ??? ???
• ?? ??? ?? ????? ?? ??
• ?? ????(??)
• Theoretical foil ? ??
• 4?? ??? ???
• It is the enormous value of conducting
  descriptive studies of work in representative or
  naturalistic settings
• Converging characterization of human work
• Context-conditioned variability? ???
• Human work? strong social components? ???.
• Work is also seldom solely focused on internal
  mental processing because worker create tools
  reducing the burden on scare cognitive resources.
• Current work practices are shaped by historic
  cultural factors.
• Time pressure and other constraint??? ??? ???,
  ??? ?? ????? ??? ????? ???.

17
Implications for work analysis

• Work analysis framework must include at least
  five dimensions of work
• Work domain analysis ??
• ???? ?? ??? ??? ?? information requirements? ????
  ??
• Constraint-based task analysis ??
• ????? flexible, situated manner? ????? ??? ??
  information requirements? ???? ??
• Effective Strategies? ?? ??
• ?? ????? ???? ?? ??? ?? ??? ? ?? ????? ???? ??
• Social and organizational factors? ??? ??
• Computer based information system? ???? ?? ???
  ????? ??? ? ? ????, ?? ??, ????? ?? ?
• ???? ??? Workers competences(??,??) ??
• ??? ??? ????, skill acquisition? ???? ?????? ????
  ??

18
Why descriptive approaches are notenough The
task-artifact cycle

1. Limitations of basing design solely on current
   practice
2. Are these limitations recognized?
3. From descriptive analysis to design implications
   The track record
4. People are adaptive the task-artifact cycle
5. Summary

19
Limitations of basing design solely on current
practice

• Intrinsic Work Constraints
• Constraints on achieving work goals, independent
  of any particular device
• These constraints are an inherent part of work in
  a particular domain
• IWC delimit the actions that are required to get
  the job done
• Current Work Practices
• Current practice is device-dependent

20
Why computer-based information systems should not
be designed to support current practices.
21
Direction of computer-based information systems
design

• Currently unexplored actions may require too much
  time, computational effort, memory demands or
  knowledge with the existing device.
• Then if with the existing information systems,
  workers may simply decides to omit this tasks
• If the requisite computer support were provided,
  these unexplored possibilities could very well
  become a productive part of workers practices

22
Implications

• Computer-based information systems should ideally
  be designed to support intrinsic work
  constraints, not just current work practices
• Two reasons
• We do not want to base our design on the
  workaround activities that are vestiges of poor
  device design
• We want to support currently unexplored
  possibilities.
• Work analysis? ? ????? intrinsic work
  constraints? ???? ???. ??? ??? descriptive
  studies? ???? ??? ?? ?? ???.
• The point is that future designs should go beyond
  current practice by removing unwanted
  inefficiencies and by adding new functional
  possibilities.

23
Are these limitations recognized?

• Limitations of Descriptive Analysis
• Benyon (1992) Embodying current practice in
  future systems is a fundamental error
• Beyer and Holtzblatt(1998), Holmqvist(1991)
• However,
• The analysis of current practices should be
  viewed as one of several possible means to
  investigate work constraints, rather than an end
  in itself

24
From descriptive analysis to design implications
The track record

• Social science
• ? ?? ???? ?????, ??? ?? ??? ???? ???
• Activity theory
• The descriptive nature of activity theory make it
  difficult to develop a novel design
• Francophone ergonomics community
• Human factors community

25
People are adaptive the task-artifact cycle

• Task-artifact cycle (Carroll et al. 1991)
• ?? work practice? ???? requirements? ?????
  artifact? ?????.
• ??? ???? ??? work practice? ?????.

26
People are adaptive the task-artifact cycle

• If we conduct a descriptive work analysis to
  understand workers current tasks, we will
  identify requirements that could be used to
  design a new artifact

27
Existing techniques for getting around the
task-artifact cycle

• Rapid prototyping and iterative user testing
• To create prototypes of new designs and evaluate
  them by having workers use them
• The end goal is to iteratively maximize the
  overlap between the two sets
• The subset of workaround activities becomes
  smaller and smaller
• The subset of currently unexplored possibilities
  becomes smaller and smaller
• Scenario-based design (Carroll et al. 1991)
• Analytical techniques for trying to achieve the
  same objective ? Empirical way
• Scenarios provide a means for analytically
  evaluating Simulated future work

28
Limitations The problems of device-dependence
and incompleteness

• Strong device dependence
• A dog chases its tail continuous iteration
  process
• Incompleteness
• Scenarios representation is incomplete the
  number and range of tasks
• Limitation by the ingenuity and creativity of the
  designer
• EX) Accident data

29
An alternative way of getting around the
task-artifact cycle Modeling intrinsic work
constraints

• Completeness The need for models

Scenario/Prototype Modeling
Inductive, bottom-up activities Top-down basis for generalization
Dont know what factors are included and what factors are missed To make explicit what attribute have been included
Representing exemplars Representing classes
Models provides a more systematic and explicit
basis for work analysis
30
Device-independence focusing on intrinsic work
constraints (1)

• Model what?
• There are many different entities that could be
  modeled
• To find a way escaping from the regress by the
  task-artifact cycle
• To find the set of intrinsic work constraints ?
  Not including workaround activities

31
Device-independence focusing on intrinsic work
constraints (2)

• Work analysis methods should not prespecify the
  follows
• Existing set of sensors
• Contents and structure of the database
• Functionality of the automation
• Allocation of functions between computer and
  workers
• Allocations of job responsibilities
• Appearance and structure of the interface
• Workers competences
• Because each issues is a point of design leverage
  ? inheriting the vestiges of the old

32
Device-independence focusing on intrinsic work
constraints

• The following decisions should be made based on
  the findings obtained from the work analysis
• What information should be gathered
• How it should be organized
• How to automated
• What to automated
• How to organize work
• How to display information
• How to train operators
• Ways of identifying intrinsic work constraints
• Studying the structure of the work domain
  identify efficient tasks and novel strategies
• Analytical model (e.g. Operation research)
• Current practice

33
Summary

• What type of work analysis is appropriate for
  complex sociotechnical systems?
• Normative approaches
• Strong limitation
• Descriptive approaches
• Limitation
• Task-artifact cycle inherit the deficiencies of
  current practice
• To overcome the cycle, prototyping and scenario
  method
• These limitations can be directly addressed by
  explicitly modeling intrinsic work constraints