midterm review

Midterm Review

IST 331 – Design and Organization of Information Systems: User and System

Principles

Instructor: Mithu Bhattacharya

Small revisions by Frank Ritter

Spring 2011

Chapter 1: Why and when do we need HCD

• Why study of the user is important?– Understanding the users can save lives– Understanding the user can lead to better products– Understanding the user can save money– Understanding the user does not guarantee success– Not necessary or sufficient for success– Lack of usability is sufficient for failure

• When do you need to study the user?– Early in the design– Lots of users and they are important– When they are different– When lives are at risk

2

Chapter 1: Why and when do we need HCD (Contd.)

• How much do you have to study the user?– Risk driven spiral design model

• Evaluate all risks to success• Study user and their tasks until risk of not knowing

is lower than other risks

3

Chapter 2: History of Human Centered Design

• Component Fields of HCD– Human Factors

• Focus on good ‘fit’ between people and their work environments– Ergonomics

• Focus on ‘fit’ between people and environment by altering the environment

– Human Computer Interaction • Focus on designing for people’s interaction with computer-based

applications– Computer Supported Cooperative Work

• Focus on people’s communication through computer-based applications

4

Chapter 2: History of Human Centered Design (Contd.)• Engineering approaches to studying human

behavior– Motion study (Gillian Gilbreth)

• Task could be broken into individual motions• Nature of motions determine efficiency

– Time study (Frederick Taylor)• Skilled behavior determined by sequencing of

motions made by operator and speed at which they were carried out

• Rationalizing task into most economical sequence of actions

• Maximum rate without overtiring operatives5

Chapter 2: History of Human Centered Design (Contd.)

• Hawthorne Effect– Performance improvements due to

psychological factors – Not physiological

6

Chapter 2: History of Human Centered Design (Contd.)

• What is Usability? • Functionality

• What it does• Learnability

• How easy to learn• Reliability

• Complete, consistent, and robust• Efficiency

• How fast is the system• Maintainability

• How easy is the system to maintain and upgrade• Other characteristics

• Consistency, informative feedback, explicitness, flexibility and control, error prevention and control, user guidance and support

7

Information Seeking in Context

• Article - Effective Information Systems for High-performing Self-managed Teams (Barnes et al., 1996)

• How self-managed teams decide what information is needed‐ High-performing teams communicate better in

team meetings‐ High-performing teams have clear goals‐ High-performing teams do not have domineering

team members

8

Information Seeking in Context (Contd.)

• Factors that enhance or hinder efforts to get information‐ No significant difference in high and low-performing

teams to the effect of openness of information system and amount of information available to team

‐ Team’s tolerance to the 2 factors different‐ Tolerance correlates to team maturity level

9

The Tangled Web We Wove: A Taskonomy of WWW Use (Byrne et al., 1999)

10

The Tangled Web We Wove: A Taskonomy of WWW (Contd.)

11

• Implications for WWW browser design– Users spend more time in reading, visual search, and waiting– Widget design could improve scrolling

• Users spend long time scrolling– Improving performance of caching algorithms

• Users spent long in waiting for page loading• Implications for page design

– Web pages should be designed to improve readability because users read

– Tradeoff between readability and scanability should be carefully evaluated

– Visual search should be supported (Use of color links, underline etc.)

Chapter 3: User Characteristics: Bodies, Behavior, Thinking, and Groups –the ABCS

• ABCS Framework• A: Anthropometric Approach

– Can it be used?– Physical aspects of users and systems– Example: Users defined by their size, muscle strength etc.

• B: Behavioral Approach – How is it used?– Basic behavior users can produce– Users defined by what they can perceive and what they can do– Example: Vision and hearing

12

Chapter 3: User Characteristics: Bodies, Behavior, Thinking, and Groups –the ABCS (Contd.)

• C: Cognitive Approach – How do users’ think they are using it?– Considers how users think about their task and system– Example: Memories, goals, processing

• S: Social Approach – What about others when using it?– Users defined by where they are – the context– Example issues on social level

• Distraction by interruptions thus failing to complete safety checklist

• Communication breakdown among members

13


• Haptic Interface– Touch based interface– Hand main organ of haptic perception– Advantages

• Supports users with poor vision• Supports users with poor sense of touch• Users who need additional input channel or need touch as input

channel

• Why haven’t Haptic Interfaces been used more?– Cost, Power , Safety (Not distal; adds safety requirements),

Resistance to change, Usefulness (Visual and auditory input enough; Not good for communicating large amount of information; Problem being used to transfer information over long time period)

14


15

• Fitt’s Law• Time to point to an object related to distance from object and

inversely related to size of object

Target

x

W

d

Time = 70 ms * Log 2 (Target distance / Target size + 0.5)(Card, Moran, and Newell, 1983)

Chapter 6: Cognitive Capabilities: Memory, Attention, and Learning

The Stages of User Activities When Performing a Task

16

Chapter 6: Cognitive Capabilities: Memory, Attention, and Learning (Contd.)

17

Long-term Memory

Working MemoryVisual Image

StoreAuditory Image

Store

PerceptualProcessor

CognitiveProcessor

MotorProcessor

Eyes

Ears

Fingers, etc.

sensorybuffers

• The Model Human Processor Developed by Card, Moran & Newell (1983)

18


• Memory• Working memory (short term)

– Small capacity– Rapid access (~ 70ms) & decay (~200 ms)

• Pass to LTM after a few seconds• Primacy effect: First things in list easily remembered• Recency effect: The last items in list better remembered

• Long-term memory– Huge (if not “unlimited”)– Slower access time (~100 ms) w/ little decay

19


• Kinds of Memory• Declarative memory

– Facts or statements– What is knowledge– Explicit (Reportable)

• Procedural memory– Performing procedures– How to knowledge– More robust against decay– Implicit (not reportable)

• Prospective memory– To do something at future time– Prone to decay– Calendars, to do list etc

20


• Recognition over Recall• Recall

– Info reproduced from memory• Recognition

– Presentation of info provides knowledge that info has been seen before

• We want to design UIs that rely on:– Recognition

• Implications of Memory for Interface Design– Trying harder does not help– Ordering the presentation of objects to memory helps

21


• Process of Learning– Stage 1 (Cognitive): Acquire domain declarative

information– Stage 2 (Associative): Declarative knowledge

compiled to procedural information– Stage 3 (Skills or Tuning): Tunes knowledge that

is applied

22


• Rasmussen’s Theory of Knowledge– Skill based control

• Operator sees and acts• Little or no effects on other tasks• E.g. following a car in a lane

– Rule based control• Operator applies effort to act• Attention is needed• E.g. lane changing in a car

– Knowledge based control• Performance is very effortful, error prone• Used when other two levels are not applicable• E.g. Driving in an unfamiliar town

23


Power Law of LearningTask time on the nth trial follows a power lawTime of a trial = Constant 1 (Number of trial + PP)˄-alpha + Constant 2

• Constant 1 is base time that decreases with practice; PP is previous practice on the task; alpha is between 0.1 – 0.4; Constant 2 is limiting constant

• You get faster the more times you do it• Applies to skilled behavior (sensory & motor)• Does not apply to knowledge acquisition, scores, or quality

24


• Implications of Learning for Interface Design– Users learn and get faster– Learning curve may provide insights

• How difficult to novices• Steep curve, system often used, initial task time not an issue• System used few times, initial task time needs to be acceptable

Chapter 7: Cognitive Capabilities: Human-Computer Communication

• What Syntactic/Semantic Model Reveals • Mapping between three items is extremely important

– Task semantics to computer semantics to computer syntax• Task semantics: Write letter• Computer semantics: Open a file, use editor, save it to disk• Computer syntax: Select menu items, key strokes for formatting

– Bad mapping: Using LaTex to write letter• Aside from task semantics, must also know semantics/syntax of:

– Text editor– Latex– Unix compiling and printing sequence (to typeset and print)

– Relatively good mapping: Trashcan to throw away files• Must know mouse syntax of selecting and dragging• Computer semantics almost analogous to task semantics

25

Chapter 8: Cognitive Capabilities: Mental Representations and Problem Solving

• What does “Mental Model” mean?• Some sort of mental representation of things in our

environment⁻ Semantic/propositional ⁻ Visual/spatial

• They help us understand how people reason about different phenomena

• Characteristics of MM‐ Incomplete, constantly evolving, not accurate representations,

contains errors‐ Simplistic representation of complex phenomena‐ Set of if-then-else rules

26

Chapter 8: Cognitive Capabilities: Mental Representations and Problem Solving (Contd.)

•Types of MM ‐Structural

•Facts user has about how a certain system works‐Functional

•Procedural knowledge about how to use the system•Can be constructed from existing knowledge about similar domain or system

27


•MM Limitations‐Capturing and validating is hard‐MMs are built on the fly‐Asking about MM often modifies user’s MM

•Post Completion Errors‐This error arises when goals for task is completed but the goals of the subtasks are not completed‐Example: Old-fashioned ATM machine

•Get money before card and leave card•Goal is to withdraw money, not get card•Design suggestion

•Put most important goal last so that all sub-goals are met

28


• Simple Decision Making• Hick-Hyman law

– T=a+b log2(n+1)– T=time to make a decision, n=number of options– a, b are constants that depend on the display, response mode etc.

• Example: What is relative time to select from one menu of eight items vs. two menus of four items?

29

30

Chapter 16: Cognitive Capabilities - Cognitive Dimensions and the Gulfs

• Influences on Decision Making• Confirmation Bias

– We tend to look for or notice evidence that confirms our hypotheses, rather than check for contradictions

– Confirmation bias can be made worse by automation, e.g., ignoring road signs while following GPS instructions

• Regression to the Mean / Sample Sizes– Users tend to over generalize

• Availability Bias– Users typically use memory that are easy to retrieve

• Framing Effect– The way outcomes are presented influence on how users choose

between alternatives– Outcomes noted in positive terms are chosen over negative terms

31

Chapter 16: Cognitive Capabilities - Cognitive Dimensions and the Gulfs (Contd.)

• Influences on Decision Making: Learning and Feedback– Let users make decision without feedback

• Users get more confident

– Give users feedback on decision making• Users get better

• Cognitive Dimensions– Hidden Dependencies

• How visible relationships are between components• Example: Spreadsheets show formula in one direction

– Viscosity• How easy is it to change the system• Example: A word document with figure numbers typed manually• Implications: Designers should make dangerous actions viscous

32


• Cognitive Dimensions– Role expressiveness

• How clear the mapping of objects are to their functions• Example: Buttons on interface being clearly buttons• Banners or logos as buttons could be misleading

– Premature commitment• How soon does the user have to decide something• Example: Some databases require planning record structures

and size limits on them before entering any data


•What the four stages model reveals •The “Gulf of Execution”

– Do actions provided by system correspond to the intentions of the user?

– Gulf: Amount of effort exerted to transform intentions into selected and executed actions

– A good system:• Direct mappings between Intention and selections• e.g. printing a letter:

–put document on printer icon–vs select print from menu

GoalsPhysicalSystem

gulf ofexecution

33


•What the four stages model reveals• The “Gulf of Evaluation”

– Can feedback be interpreted in terms of intentions and expectations?– Gulf: Amount of effort exerted to interpret feedback– A good system: Feedback easily interpreted as task expectations

• e.g. graphical simulation of text page being printed– A bad system: No feedback or difficult to interpret feedback

• e.g. Unix: “$”, “bus error”, “command not found”

GoalsPhysicalSystem

gulf ofevaluation

34


Implications of the Gulfs for Design•Make gulfs narrower where appropriate

–Allow visibility of appropriate information for achieving relevant tasks–Feedback, consistency, and understanding users mental models can reduce the gulfs

•Make gulfs wider where appropriate–Relevant for dangerous and expensive actions–Hide components–Make unavailable action impossible to do–Do not give any feedback

35

Chapter 5: Behavioral: Basic Psychology of the Senses of the User (Contd.)

• Sensation– The experience of sensory information – Determined by stimulus quality and sensory organ– Objective process

• Perception– The process of creating meaningful patterns from raw sensory

information– Influenced by past experiences, expectations, and feelings– Subjective process

• Habituation– Habituation occurs when percept occurs repeatedly without

importance– Over time stimuli seems less important and perceptible

36


37

• Signal Detection Theory• SDT measures how accurate is performance• Gives a way to analyze complex situations

Signal Present

Response Yes No

Yes Hit False Alarm

No MissCorrect Rejection

Types of responses to a signal


38

• Key parameter: Distinguish signal from noise• Signal normally distributed some distance away from 0• Noise is distributed around 0• Threshold (Criterion response) – parameter that the observer adjusts• Area of signal distribution correctly classified as signal(to the right of threshold) – Hits• Area of signal to left of threshold – Miss• Noise classified as signal to the right of threshold – False alarm• Noise to the left of threshold - Correct rejection• Distance and threshold can be computed from their ratio using tables from normal

distribution

39


• Two stages in vision– Physical reception of stimulus– Processing and interpretation of stimulus

• Retina– Center of retina has most of the cones – Allows for high acuity of objects focused at center– Edge of retina is dominated by rods– Allows detecting motion of threats in periphery

• Photo-pigments not distributed evenly– Mainly reds (64%) & very few blues (4%) – Center of retina (high acuity) has no blue cones

• Disappearance of small blue objects you fixate on

40


• Different wavelengths of light focus at different distances behind eye’s lens– Need for constant refocusing

• Causes fatigue– Be careful about color combinations– e.g., no blues at the same time as reds

41


• Color Guidelines• Avoid red & green in the periphery - why?

– Lack of RG cones there -- yellows & blues work in periphery

• Avoid pure blue for text, lines, & small shapes– Blue makes a fine background color

• Avoid single-color distinctions– Mixtures of colors should differ in 2 or 3 colors

• e.g., 2 colors shouldn’t differ only by amount of red

42

Chapter 11: Errors: An Inherent Part of Human-System Performance (Contd.)

• What is error?– Precursors to accidents– Errors trigger a chain of events

• Types of errors‐ Perceptual errors (B8, Z2, I1)‐ Cognitive errors (Memory, Link, Inconsistent)‐ Motor errors (Hand, Eye)

‐ Options very close to each other‐ Slips, errors, erroneous knowlege

• How to gather data to study errors?– Laboratory-based experiments– Field-based observations– Archive data– Combination methods are better


• Analyzing Errors• Event trees

– Bottom up technique– Sequence of events leading to all possible outcomes– Based on binary logic (Each node in tree has 2 possible branches;

Yes / No)– Each event can be assigned a probability– Sum of probabilities corresponding to each node must be 1– Probability of different outcomes can be calculated by multiplying

(ANDing) together all event probabilities along the path that leads from initiating event to outcome

43


• Fault trees– Top down technique– Start with outcome and work backwards to find all causes– Does not need to be binary tree– Outcome can be determined by either ANDing or ORing together

a set of possible causal factors– Can be qualitative or quantitative– Quantitative fault tree – Probability of occurrence is allocated to

each of lowest level leaf nodes in tree

44

midterm review

Documents

user guidance

successstudy user

information available

context contd

hcd contd

hcdwhy study

selfmanaged teams barnes

spiral design modelevaluate