Predictive and Cognitive Models

1
Predictive and Cognitive Models
2
Agenda

• Questions
• Overview of modeling
• Model Human Processor (CMN reading)
• Fitts Law (Mackenzie reading)
• KLM/GOMS (DFAB, Ch. 12)

3
Overview of Modeling

• Translating empirical evidence into theories and
  models that influence design.

4
What to Expect from a Model

• Performance measures
• Quantitative
• Time prediction
• Working memory constraints
• Competence measures
• Legal behavior
• Consistency (transfer)

5
Our Plan

• Fitts Law
• Low-level motor model
• Predictive cognitive models
• GOMS, KLM, Production Systems, Grammars
• Higher level qualitative theories
• Situated action, Activity theory, Distributed
  cognition

6
Modeling Low-Level Actions
7
Fitts Law

• Fitts Law
• Models movement times for selection tasks
• Paul Fitts war-time human factors pioneer
• Basic idea Movement time for a well-rehearsed
  selection task
• Increases as the distance to the target increases
• Decreases as the size of the target increases

8
Moving

• Move from START to STOP

Index of Difficulty ID log2 ( 2A/W ) (in
unitless bits)
width of target
distance
9
Movement Time
MT a bID or MT a b log2 (2A/W)
10
How MT is determined

• Empirical measurement establishes constants a and
  b
• Different for different devices and different
  ways the same device is used.

11
Extending to 2D

• Recall Fitts Law established in 1D
• 2D Most relevant to UI design
• What is W?

12
Applications

• When does it apply?
• How used in interface design?

13
Extending Fitts Law

• Buxton 3-state model
• adjust Fitts Law for different selection states
  (0, 1, 2)

14
Cognitive/User Modeling

• Idea Build a model of how a user works, then
  predict how she will interact with the interface.
• Goals (Salvendy, 1997)
• Predict performance consequences of design
  alternatives
• Evaluate suitability of design alternatives to
  support and enhance human abilities and
  limitations
• Generate design guidelines that enhance
  performance and overcome human limitations

15
Differing Approaches

• Many different modeling techniques
• Human as Information Processing machine
• Procedural models
• Many subfamilies and related models
• Predict performance, not truth
• 4 examples
• Situated action
• Activity theory
• Distributed cognition

A later lecture
16
1. Model Human Processor

• Card, Moran, Newell (1983, 1986)
• Procedural models
• People learn to use products by generating rules
  for their use and running their mental model
  while interacting with system
• Components (CMN, p. 26-7)
• Set of memories and processors together
• Set of principles of operation
• Discrete, sequential model (later parallel)
• Each stage has timing characteristics (add the
  stage times to get overall performance times)

17
MHP Three Systems in Model

• Perceptual, cognitive, motor systems
• Timing parameters for each stage/system
• Cycle times (?)
• ?p 100 ms (middle man values)
• ?c 70 ms
• ?m 70 ms
• Perception Cognition have memories
• Parameters code, decay time, capacity

18
MHP Model and Parameters
19
MHP Principles of Operation

• Set up rules for how the components respond.
• Can be based on experimental findings about
  humans.
• Recognize-act cycle, variable perceptual
  processor rate, encoding specificity,
  discrimination, variable cognitive processor
  rate, Fitts law, Power law of practice,
  uncertainty, rationality, problem space

20
Applying the MHP

• Example Designing menu displays
• 16 menu items in total
• Breadth (1x16) vs. Depth (4x4) ?

• Submenu
• Menu item a
• Menu item b
• Menu item c
• Menu item d

• Menu
• Menu item a
• Menu item b
• Menu item c
• Menu item d
• Menu item e
• Menu item f

• Main Menu
• Menu item a
• Menu item b
• Menu item c
• Menu item d

• Submenu
• Menu item a
• Menu item b
• Menu item c
• Menu item d

21
MHP Calculations
Breadth (1x16) ?p perceive item, transfer to
WM ?c retrieve meaning of item, transfer to WM ?c
Match code from displayed to needed item ?c
Decide on match ?m Execute eye mvmt to (a) menu
item number (go to step 6) or (b) to next item
(go to step 1) ?p Perceive menu item number,
transfer to WM ?c Decide on key ?m Execute key
response Time ((161)/2) (?p 3?c ?m)
?p?c?m Time 3470 msec
Depth (4x4) Same as for breadth, but with 4
choices, and done up to four times (twice, on
average) Time 2 x ((41)/2) (?p 3?c ?m)
?p?c?m Time 2380 msec
Therefore, in this case, 4x4 menu is predicted to
be faster than 1x16.
Serial terminating search over 16 items
22
Related Modeling Techniques

• Many techniques fall within this human as info
  processor model
• Common thread - hierarchical decomposition
• Divide behaviors into smaller chunks
• Questions
• What is unit chunk?
• When to start/stop?

23
2. GOMS

• Goals, Operators, Methods, Selection Rules Card,
  Moran, Newell (1983)
• Assumptions
• Interacting with system is problem solving
• Decompose into subproblems
• Determine goals to attack problem
• Know sequence of operations used to achieve the
  goals
• Timing values for each operation

24
GOMS Components

• Goals
• State to be achieved
• Operators
• Elementary perceptual, cognitive, motor acts
• Not so fine-gained as Model Human Processor
• Methods
• Procedures for accomplishing a (sub)goal
• e.g., move cursor via mouse or keys
• Selection Rules
• if-then rules that determine which method to use

25
GOMS Procedure

• Walk through sequence of steps to build the model
• Determine branching tree of operators, methods,
  and selections, and add up the steps needed

26
GOMS Limitations

• GOMS is not so well suited for
• Tasks where steps are not well understood
• Inexperienced users
• Why?

27
GOMS Application

• NYNEX telephone operation system
• GOMS analysis used to determine critical path,
  time to complete typical task
• Determined that new system would actually be
  slower
• Abandoned, saving millions of dollars

28
GOMS Variants

• Keystroke Level Model
• Analyze only observable behaviors
• Keystrokes, mouse movements
• Assume error-free performance
• Operators
• K keystroke, mouse button push
• P point with pointing device
• D move mouse to draw line
• H move hands to keyboard or mouse
• M mental preparation for an operation
• R system response time

29
Other GOMS Variants

• NGOMSL (Kieras, 1988)
• Very similar to GOMS
• Goals expressed as noun-action pair
• e.g., delete word
• Same predictions as other methods
• More sophisticated, incorporates learning,
  consistency (relevant to usability and design)
• Handles expert-novice differences, etc.

30
3. Cognitive Complexity Theory

• Production rules
• IF-THEN decision trees (Kieras Polson, 1985)
• Executable
• Competence
• match against an executable system model
• Performance
• measures of memory requirements

31
4. Grammars

• To describe the interaction, a formalized set of
  productions rules (a language) can be assembled.
• Grammar defines what is a valid or correct
  sequence in the language.
• Reisner (1981) Cognitive grammar describes
  human-computer interaction in Backus-Naur Form
  (BNF) like linguistics
• Used to determine the consistency of a system
  design

32
Task Action Grammars (TAG)

• Payne Green (1986, 1989)
• Concentrates on overall structure of language
  rather than separate rules
• Designed to predict relative complexity of
  designs
• Not for quantitative measures of performance or
  reaction times.
• Consistency learnability determined by
  similarity of rules

33
Summary of Approaches

• Cognitive modeling
• Model Human Processor
• GOMS
• Basic model
• Keystroke-level models
• NGOMSL
• Production systems
• Cognitive Complexity Theory
• Grammars

34
The closure problem

• Another competence measure