Display Design and HCI

1
Display Design and HCI

• IE 5511 Human Factors

2
Projects

• I want you to
• Identify several performance parameters that help
  you to tell if you have met your goals.
• Identify a method by which you will measure these
  performance parameters. What activities,
  procedures would you follow, specifically?
• Plan an evaluation experiment to compare before
  and after, or to compare design alternatives.
• To accomplish this assignment may be useful to
  look ahead to assignment 3 start to consider
  alternative designs. How would you measure those
  alternative designs?

3
Examples of performance parameters

• Example 1 Intelligent traffic gap display
• Performance parameters
• average wait time at intersection,
• perceived mental workload for drivers,
• number of (simulated) crashes,
• estimated fatalities and injuries.
• How can you measure these things? You may want
  to use inexpensive, low fidelity methods at the
  start of the work, and more time consuming or
  costly, high-fidelity methods at the end.

4
Examples of performance parameters

• Example 2 Change the check-in process for
  patients at a clinic to improve patient thru-put,
  improve patient and staff experience.
• What are relevant performance parameters?
• How would you measure them? (formative and
  summative)?

5
Projects

• Formative evaluation inexpensive, low fidelity,
  early feedback.
• Summative evaluation more expensive, higher
  fidelity, later and final feedback.

6
Projects

• Goals and performance parameters should be
  directly related performance parameters are
  used to measure progress on goals.
• The challenge is that some goals are
• Hard to quantify (quality),
• Difficult or unethical to measure directly
  (accident and death rate),
• Prohibitively expensive to set up a test
  situation (what is material handling time in my
  new factory layout?)

7
Examples

• Goal
• Reduce time
• Increase quality
• Increase safety

• Performance Parameters
• Time to perform task
• Number of rejects
• Customer satisfaction
• Lifting index (LI)
• Number of near misses (runway incursions, etc.)
• Count number of accidents
• Count number of patient deaths.

8
Projects

• Hard data showing the difference between the old
  and the new goes much further than opinion in
• Selling your project,
• Making a convincing argument that your approach
  is a good one.

9
Projects

• Make sure your performance parameters measure
  things relevant to your most important goals.
• Think about what is the best way to measure your
  performance parameters in your re-designed
  situation
• Will you build the real thing?
• Build a paper, or cardboard prototype?
• A computer mock-up?
• Use a simulation?

10
Why are Display design and HCI important?Example
1 Power Plant Emergency Scenario

• A power plant operator is monitoring the plant
  from the control room
• An alarm sounds
• Many warning tiles are lit up, but no pattern
  is apparent.
• Gauges (below warning tiles) show no pattern.
  Some are out of range. But not clear how gauges
  connect to warning tiles.
• Manual shows clear flow diagram of plant
• Page on warning lights in separate place at back
  of manual,
• Emergency procedures are found on yet another
  page.
• Operator gives up and shuts down whole plant!!
  Very costly.

11
Challenges

• Operator must try to integrate 5 separate sources
  of information warning tiles, gauges, 3
  different parts of the manual.
• The displays did not help the operator to
  interpret or integrate complex information.
• How might you change the design or the system,
  equipment, process or environment to improve
  safety and reduce the probability of error?

12
Displays

• Artifacts designed to support
• Perception of relevant system variables
  (situation awareness SA)
• Facilitate further processing of information
• Facilitate formation of task and context
  appropriate abstractions and concepts
• Facilitate integration of information

13
Types of Displays

• A display does not have to be associated with a
  computer (example speedometer on car, knob on
  gas stove position shows current setting).
• A display can be static e.g. road signs
• Displays may be visual, auditory, haptic (touch),
  etc. Examples
• Visual Flashing light on ambulance,
• Auditory siren on ambulance,
• Haptic cell phone in vibrate mode.
• Smell Odor-rama in movie theater with scratch
  and sniff card.

14
Designing Displays

• First steps are
• Identify task
• Identify goals
• Perform a detailed information analysis

15
Principles of Display Design (grounded in
strengths and weaknesses of human information
processing)

• Perception
• Make displays legible/audible
• Avoid absolute judgment limits
• Facilitate top-down processing by providing
  context
• Use redundancy to reduce errors display in
  several forms.
• Discriminability make options easy to distinguish

• Mental Models
• Pictorial realism displays should look like the
  variables they represent.
• Principle of moving parts elements of display
  should move in a way compatible to users mental
  model

16
Principles of Display Design

• Attention
• Minimize information access cost, particularly
  when information must be integrated,
• Proximity compatibility two sources of info that
  must be integrated must be made close and/or
  compatible.
• Principle of multiple resources can be helpful
  to use multiple sensory channels

• Memory
• Use external (usually visual) memory aids to
  reduce load on STM.
• Aid people in predicting future events, so they
  may be proactive, rather then reactive.
• Consistency.

17
Why are Display design and HCI important?Example
2 Radiation Therapy (true story)

• Goal deliver two short burst of low power
  radiation to patient.
• Nurse uses radiation therapy machine from remote
  location.
• Video, auditory monitoring broken
• Nurse makes mistake in entering command, x sets
  machine to high-power (but it is OK no
  radiation delivered yet.)
• Nurse edits command. Changes to e low power.
  Software shows change. However, due to software
  error, settings on machine are not changed.
• Nurse delivers first dose of radiation. Patient
  screams. Nurse cannot hear that wrong does was
  delivered.
• Nurse delivers second dose.
• 4 weeks later patient dies of radiation sickness.

18
Aspects of HCI

• Hardware design Computer screen? Immersive
  environment? Pointing devices?
• Function design what does the software or thing
  do?
• Interface design information and format provided
  by computers, control mechanisms for providing
  information to computer,
• Interaction design when and why does the user
  use, request or give information to the tool?
  (Relationship of tool and work process)

19
HCI a type of human factors
Motivation Incentives/rewards
Organizational structure
Work Process
Work environment
Tools (DSSs, software, etc.)
20
Bad HCI can result in

• High error rate
• Confusion, panic, boredom, frustration
• Wasted time, mental effort
• Abandonment of tools,
• Misuse of the system
• Inefficient work-arounds, i.e. changes in work
  process.
• Accidents
• Death

21
Software Design Cycle

• Understand
• Design
• Evaluate
• What makes a design process a user centered
  design process?
• Is e-mailing users periodically with design
  updates sufficient?

22
The Design Process for a DSS Human Centered
Design
A typical spiral design process
Prototype Testing
Prototype Construction
Requirements Gathering
Final Performance Evaluation or Comparison
Design Specification
Design Review
23
Participatory Design

• Incorporating users as members of the design team
• Users have some degree of control over design
  decisions, in other words users have more than an
  advisory capacity.
• Caveat a different set of users must be brought
  in for user testing.

24
Understand System and User Characteristics

• What balance of factors are important in a given
  system?
• Ease of learning
• Ease of use, reductions in users time spent,
• Performance in terms of decision and patient
  health outcome,
• Error rate/accuracy
• Functionality, flexibility in range of tasks to
  which system can be applied

25
Factors impacting what system properties will be
important

• Frequency of task performance
• Mandatory vs. discretionary use
• Life-criticality of the task
• Knowledge level of users familiarity with system
• Familiarity with task domain