Human Computer Interaction Research in the MIS Discipline

1
Human Computer Interaction Research in the MIS
Discipline

• HCI Panel AMCIS 02
• Dr. Jane M. Carey

2
Definition of HCI

• Human Factors in IS (HCI) is the scientific
  study of the interaction between people,
  computers, and the work environment. The
  knowledge gained from this study is used to
  create information systems and work environments
  which help to make people more productive and
  more satisfied with their work life. (Beard
  Peterson, 1988)
• Consensus definition from the HFIS Symposia series

3
Mason Mitroff (1983)- MIS Relational Triad
Organizational context Problem type Leadership
style of org Hierarchical level of DM Uncertainty
risk
Cognitive style Psychological type Communication
skills Informational needs User experience User
attitude
Scope ,Timeliness Mode of presentation, Horizon,
Accuracy Format, Value
4
Huber, Management Science 1983

• Cognitive Style as a basis for MIS and DSS
  designs Much ado about nothing?
• This article had a dampening effect on studying
  human computer interaction due to the prestige of
  Management Science and the limited means
  available to measure cognitive style at the time
• At the time, most MIS researchers studying
  human/computer interaction used Myers/Briggs type
  indicator to measure CS

5
Referent Disciplines for HCI
MIS/Computer Science
Human Factors Engineering Ergonomics
Management Theory
Psychology
Organizational Behavior
6
Models of HCI
HFIS Symposia Series (1986-1997)
Organization
Interface Specification Tools
IS Professional
Computer
System/User Communication (Documentation)
Human/Computer Interaction
End User Involvement
Information Presentation
USER
7
A Model of HCILong Whitfield (1989)
Science Support Representations
Real World
Intermediary Representations
Human Science
Generalize
Analyze
Acquisition Representation
People

Computer Science
Computers
Synthesize
Particularize

Acquisition Representation
Other Science
Other
8
Research Agenda for HCILewis (1990)

• 1. Understand user goals and preferences
  (interface design theory)
• 2. Broaden applied cognitive theory (problem
  representation)
• 3. Support innovation (technology to support
  interface design)
• 4. Credit assignment (measure the impact of
  interface design on performance)

9
HCI Independent Dependent Variables
Independent Independent Dependent
User/Analyst Interface Characteristics Performance
Indirect A. Aptitude B. Attitude C. Decision Style 1. Content Decision Effectiveness A. Accuracy B. Timeliness C. Quality D. Confidence
2. Direct A. Training B. Experience C. Involvement 2. Form 2. Satisfaction
3. Decision Setting A. Task B. Management Level C. Uncertainty D. Timeliness E. Structuredness F. Context 3. Presentation 3. Learning A. Ease B. Time
3. Decision Setting A. Task B. Management Level C. Uncertainty D. Timeliness E. Structuredness F. Context 4. Media 4. System Responsiveness
3. Decision Setting A. Task B. Management Level C. Uncertainty D. Timeliness E. Structuredness F. Context 5. Context 5. Speed of use
3. Decision Setting A. Task B. Management Level C. Uncertainty D. Timeliness E. Structuredness F. Context 5. Context 6. Error rate
10
Measurement Issues in HCINewsted, Todd, Zmud,
1997

• Theories require constructs constructs require
  measurement, therefore no theory without
  measurement
• Definitional concerns (conceptual ambiguity and
  conceptual overlap)
• Concerns with construct-context interaction (such
  as the tight construct-context linkage in task
  performance)
• Zmud Boynton criteria for instrumentation
• Multiple-item scale
• Ability to find and use an instrument
• Description in refereed source
• Assessment of psychometric properties
• Classifying measures
• Outcome versus process measures
• Perceptual versus behavioral measures
• Obtrusive versus non-obtrusive measures
• Qualitative versus Quantitative measures

11
Measurement Issues in HCINewsted, Todd, Zmud,
1997

• Instrument validation for experimentation
• Construct validation
• Manipulation measurement
• Covariate measurement
• Experimenter effects
• Online data capture
• Conclusion HCI measures are improving, but have
  a long way to go