Soarian™ User Interface

Download Report

Transcript Soarian™ User Interface

1
Evaluating Interface Designs

How do you know your design is any good?

When will you know?
2
Evaluating Interface Designs

Determinates of the evaluation plan
 Design Stage (early, middle, late)
 Novelty of the project (well defined vs. exploratory)
 Number of expected users
 Criticality of the interface (e.g., life-critical medical systems vs. museumexhibit support)
 Costs of product and finances allocated for testing (range of 5% to 20% of the
total project budget)
 Time available
 Experience of the design and evaluation team

Failure to perform and document testing can result in
 Failed contract proposals
 Malpractice lawsuits
3
Evaluating Interface Designs

Expert Reviews
 Ask colleagues or customers for their feedback
 Expert reviews can be conducted on short notice and with little time
commitment
 Can occur early or late in the design phase
 Deliverable can be a formal report with problems identified and
recommendations
 Deliverable can also be an informal presentation with the development team
and managers
 Expert reviews may require training on the task domain
4
Evaluating Interface Designs

Expert Reviews Methods
 Heuristic Evaluation
http://www.youtube.com/watch?v=hWc0Fd2AS3s&feature=related
• Critique of the interface for conformation to a short list of heuristics
– Consistency
– Universal usability
– Informative feedback
– Closure
– Prevent errors
– Easy reversal of actions
– Internal locus of actions (user as initiator)
– Reduce short-term memory load
5
Evaluating Interface Designs

Expert Reviews Methods
 Guidelines Review
• Based on organizational guidelines
6
Evaluating Interface Designs

Expert Reviews Methods
 Consistency Inspection
• Terminology, fonts, colors, layout, input/output formats
7
Evaluating Interface Designs

Expert Reviews Methods
 Cognitive walkthrough
• Experts simulate users walking through the interface to carry out a typical task.
• Start with high-frequency tasks
• Critical tasks should definitely be evaluated
8
Evaluating Interface Designs

Expert Reviews Methods
 Bird’s Eye View
• Study a complete set of UI screens on the floor (or pinned to walls)
• Provides a easy way to see fonts, colors and terminology
9
Evaluating Interface Designs

Expert Reviews Methods
 Expert-Review Report
• Can use the guidelines document to structure the report
• Comment on novice, intermittent and expert features
• Rank recommendations by importance and effort level
Effort Level
Low
High
Low
1, 3, 5
2, 4, 6
High
7, 9, 11
8, 10, 12
User
Importance
10
Evaluating Interface Designs

Usability Testing and Laboratories
 Controlled experiments
• Generally have at least two treatments
• Need to show statistically significant differences
• Goal is validation or rejection of a hypothesis
 Usability tests
• Goal is to find flaws in the interface
• Fewer participants
• Outcome is a report
 Both studies include carefully prepared
set of tasks
11
Evaluating Interface Designs

Usability Testing and Laboratories
 Having a usability lab on sight shows a commitment to customers, users and
employees
 Generally contains two 10 x 10 rooms, divided by a half-silver mirror
 Staffed by one or more people
• Ideally have been involved in early task analysis or design reviews
 Example – Display based phones
12
Evaluating Interface Designs

Usability Testing and Laboratories
 Two to six weeks before the usability test
• Develop the detailed test plan (list of tasks, subjective satisfaction questions,
debriefing questions)
• Identify the number, types and source of the participants
– Sources: Customer sites, personnel agencies, advertisements
• Conduct a pilot test one week ahead of testing
 Participants
• Notify them that it is the software being evaluated, not them
• Inform them of the tasks they will be performing (e.g., ordering a product on a
website)
• Inform them of how long they will be in the session (normally 1 to 3 hours)
• Obtained informed consent
13
Evaluating Interface Designs

Usability Testing and Laboratories
 Informed consent
• I have freely volunteered to participate in this study
• I have been informed in advance of the tasks and procedures
• I have been given the opportunity to ask questions
• I am aware that I have the right to withdraw consent and to discontinue
participation at any time, without prejudice to my future treatment
• My signature below may be taken a affirmation of all above statements; it was
given prior to my participation in this study
 Post tasks
• Participants can make general comments or suggestions, or respond to specific
questions
 Videotaping
• Reviewing can be tedious
• Log and annotate during the test
• Look for critical incidents
14
Evaluating Interface Designs

Usability Testing and Laboratories
 Eye Tracking – Heat Maps
15
Evaluating Interface Designs

Usability Testing and Laboratories
 Paper mockups
• Early is the design phase
• Get user reactions to wording, layout, and sequencing
16
Evaluating Interface Designs

Usability Testing and Laboratories
 Discount usability testing
• Three to six participants (allows prompt revision and repeated testing)
• Formative evaluation – identifies problems that guide re-design
• Summative evaluation – provides evidence for product announcement
– “99% of our 100 testers completed their tasks without assistance
 Competitive usability testing
• Compares the new interface to previous versions or similar products from
competitors
• Within-subjects designs are the most powerful
 Think Aloud
• http://www.youtube.com/watch?v=tbKnFaW69e0&feature=related
17
Evaluating Interface Designs

Usability Testing and Laboratories
 Field tests and portable labs
• Puts new interfaces to work in realistic environments for a fixed trial period
• Need portable labs with videotaping and logging facilities
 Remote usability testing
• Web-based applications tested internationally, on-line
• Can recruit testers via email
• Less control over user behavior, and less chance to observe their reactions
• Usage logs and phone interviews are useful supplements
• UserWorks, Inc.
 Can-you-break-this tests
• Destructive testing approach
• Users attempt to find fatal flaws
18
Evaluating Interface Designs

Usability Testing and Laboratories
 Short comings
• Limited coverage of interface features
• Hard to predict success in long-term usage
• The lab environment is different than the real work environment
19
Evaluating Interface Designs

Survey Instruments
 Often a companion to usability testing and expert reviews
 Specify survey goals
• Ask the users for the subjective impressions about specific aspects of the
interface. E.g., representation of:
– Task domain objects and actions
» E.g., appointments, PAT, treatment series
– Interface domain metaphors
» E.g., shopping cart
– Syntax of inputs and design of displays
» E.g., copy, add
• User specific information
– Background (e.g., age, gender, education, income)
– Experience with computers (e.g., software packages, length of time, depth of
knowledge, TurboTax)
20
Evaluating Interface Designs

Survey Instruments
• User specific information
– Job responsibilities (e.g., trenches, manager)
– Personality type (e.g., introvert/extrovert, risk taking, early adopter)
– Reasons for not using an interface (e.g., too complex, too slow)
– Familiarity with features (e.g., printing, short-cuts, tutorials)
– Feelings about using the interface (e.g., confused vs. clear, frustrated vs. in
control, bored vs. excited)
• Coleman and Williges (1985) – Bipolar Semantically Anchored Items
– Hostile
1234567
Friendly
– Vague
1234567
Specific
– Misleading
1234567
Beneficial
– Discouraging 1234567
Encouraging
21
Evaluating Interface Designs

Survey Instruments
 Questionnaire for User Interaction Satisfaction (QUIS) – Shneiderman
– Readability of characters
– Layout of displays
– Meaningfulness of icons
– Interface actions (e.g., short-cuts)
– Terminology
– Screen sequencing
22
Evaluating Interface Designs

Survey Instruments

QUIS: General Content
• System experience (e.g., time spent on the application)
• Past experience (e.g., operating systems, devices, software)
• Overall reactions (e.g., terrible/wonderful; rigid/flexible)
• Screen objects (e.g., characters, highlighting, layouts, sequence)
• Terminology (e.g., error messages, amount of system feedback)
• Learning (e.g., getting started, time to learn advanced features)
• Exploration of features by trial and error
• Remembering names and use of commands
• Steps to complete a task are in a logical sequence
• System capabilities (e.g., speed, reliability)
• User manuals, online help, and tutorials
• Multimedia (quality of picture and sound)
• Teleconferencing (e.g., set-up, image quality, connector indicators)
• Software installation
23
Evaluating Interface Designs

Acceptance Tests
 Used for software acceptance today
• Specific cases with possible response time requirements
 Applied to usability acceptance
• Time to learn specific functions
• Speed of task completion
• Rates of errors
• User retention of commands
• Subjective user satisfaction
 The goal is not to detect flaws, but to verify adherence to requirements
24
Evaluating Interface Designs

Evaluation During Active Use
 Major changes should be announced semi-annually or annually
 Interviews and focus-groups
• One-on-one interviews and yield comments that can be discussed with a larger
audience
 Continuous user performance data logging
• The software support the collection of:
– Patterns of usage (e.g., new vs. existing patient)
– Speed of user performance
– Rate of errors
– Frequency of errors
» Can be a candidate for a feature to receive specific attention
– Access to help or support on an issue
– Simplify access to frequently access features
– Rarely accessed features (why are they being avoided)
– Potential privacy issues
25
Evaluating Interface Designs

Evaluation During Active Use
 Online or telephone consultants
• Excellent source of information about problems users are having
• Source of suggested improvements
 Blogs to discuss user problems
 On-line suggestion box and email trouble reporting
26
Evaluating Interface Designs

Goal of an index similar to miles-per-gallon, energy efficiency ratings
 Learning time estimates
 User satisfaction index
27
Evaluating Interface Designs

Simple Designs?

INFOBAR C01 Japan’s Newest Android Phone
28
Evaluating Interface Designs

Controlled Experiments
 The scientific method and HCI
• Deal with practical problems
• State a testable hypothesis
• Identify a small number of independent variables
• Identify the key dependent variables
• Judicially select participants
• Control for biasing factors (participants, tasks)
• Apply appropriate statistical methods
• Resolve practical problems
 Fractions of users can be given improvements for a limited amount of time,
and compared to a control group. Dependent measures may include:
• Performance times
• User satisfaction
• Error rates
• User retention over time