Human-Computer Interaction - INF4820

Chapter 1

What is user experience

user involved

Interface

Interested in experience, to be measured

what are metric

Measure or evaluation

task success, satisfaction, errors

Observable, quantifiable

UX measures effectiveness, efficiency or satisfaction

enables informed decisions

Value of UX metrics

key in calculating ROI

Chapter 2

Independent and dependent

independent we manipulate

dependent are measured - success rate, errors, satisfaction, completion

Types of data

nominal

ordinal

interval

ratio

unordered groups / categories

requires coding (assigning codes)

ordered groups/categories

distance between ranks not meaningful

user can rate a website, good, bad, great

continuous such as temperature

distance between 20 and 30 is meaningful

allows descriptive statistics

ask if point halfway between to make sense

same as interval, but has 0

age, height, weight

zero indicates absence

also descriptive statistice

but includes mean

descriptive statistics

central tendency

variability

confidence intervals

Comparing means

Relationships between variables

nonparametric tests

Graphs

describes data makes no inference

mean, median, mode

how data is spread out

range, variance and standard deviation

95% confidence interval

Mean ± 1.96 * [standard deviation/sqrt(sample size)]

independent or paired samples

overlapping means similar

t test to calculate if means are different

independent samples

paired samples

a, b testing

need equal amount of values to pairs match up

multiple samples compared using anova

correlation

negative relation, as one increases the other decreases

analyse nominal and ordinal data

chi square

column or bar chart

line graphs

scatterplot

pie or donut chart

Stacked

Chapter 3

3.1 STUDY GOALS

3.2 USER GOALS

3.3 CHOOSING THE RIGHT METRICS: TEN TYPES OF
USABILITY STUDIES

3.4 EVALUATION METHODS

3.5 OTHER STUDY DETAILS

3.1.1 Formative Usability

3.1.2 Summative Usability

3.2.1 Performance

3.2.2 Satisfaction

3.3.1 Completing a Transaction

3.3.2 Comparing Products

3.3.3 Evaluating Frequent Use of the Same Product

3.3.4 Evaluating Navigation and/or Information Architecture

3.3.5 Increasing Awareness

3.3.6 Problem Discovery

3.3.7 Maximizing Usability for a Critical Product

3.3.8 Creating an Overall Positive User Experience

3.3.9 Evaluating the Impact of Subtle Changes

3.3.10 Comparing Alternative Designs

3.4.1 Traditional (Moderated) Usability Tests

3.4.2 Online (Unmoderated) Usability Tests

3.4.3 Online Surveys

3.5.1 Budgets and Timelines

3.5.2 Participants

3.5.3 Data Collection

3.5.4 Data Cleanup

Chapter 4 - Performance Metrics

4.4 EFFICIENCY

4.2 TIME ON TASK

4.3 ERRORS

4.1 TASK SUCCESS

4.5 LEARNABILITY

intro

4.1.2 Levels of Success

4.1.3 Issues in Measuring Success

4.2.1 Importance of Measuring Time on Task

4.2.2 How to Collect and Measure Time on Task

4.2.3 Analyzing and Presenting Time-on-Task Data

4.2.4 Issues to Consider When Using Time Data

4.3.1 When to Measure Errors

4.3.2 What Constitutes an Error?

4.3.3 Collecting and Measuring Errors

4.3.4 Analyzing and Presenting Errors

4.3.5 Issues to Consider When Using Error Metrics

4.4.1 Collecting and Measuring Efficiency

4.4.2 Analyzing and Presenting Efficiency Data

4.4.3 Efficiency as a Combination of Task Success and Time

4.5.1 Collecting and Measuring Learnability Data

4.5.2 Analyzing and Presenting Learnability Data

4.5.3 Issues to Consider When Measuring Learnability

changes as product is being developed

iterative

questions

what are most significant usability issues

What is good, what is frustrating

Most common mistakes users make

Do iteration result in improvements

What issues will remain after launch

measure how well a product meets expectations

evaluate against criteria

Identify improvements

questions

did we meet goals

Overall usability?

comapre against compitition

improvements from one release to next

what the user does

how successfully a task is completed

how much effort

how many errors

amount of time it takes

what user says or thinks

was it easy or confusing

performance and satisfaction not always correlated

what tech, money, time do we have available

well defined start and end

% of completions

Drop off rate and where users dropoff

self reporting metrics

liklihood of returning

user expectiation`

efficiency is completion rate

compare to competition or previous releases

is production improving

use success measure, efficiency and satisfaction

taks time

learnability

self reporting on awareness and usefulness

makes use of mockups or wireframes

task success best measure

lostness

card sort study

advertising

underused part

number of interactions with element

self reporting

memory test

eye tracking

a/b testing

periodic checkup

more open ended tests

use actual site if possible

unique per user

assign priority

hard to compare

quick improvements

completing task is utmost importance

large number of participants required

user error

success rate

may tie to efficiency measure too

subject but still measurable

satisfaction most important

Many self reports

on satisfaction

and expectation

likelihood of future use

physiological mesaures for engagement

pupil

heart rate

skin conductance

live site metrics from a/b testing

less effective email or surveys

usually early on

ask participant opinion on different designs

they can also rate prototypes

how many participants and what metrics are required

5-10 users

lab test one-on-one

formative studies

record issues, frequency, type and severity

self reporting possible

beware small sample size cause over generalization

collect lots of data in short amount of time

automatically collected data

can collect quantitve and qualitive data

less usefull when deeper insights are required

needs well defined tasks

start and end state

well defined success criteria

in lab user can verbalise how they completed task

or some other structured way to test user afterwards

proxy measures

4.1.1 Binary Success

success or not

bar chart represent success rate

larger sample size more accurate, more confidence

how close to success user came

user experience of success

optimal success of took longer route to success

3 levels, success, partial, failure

can add assistance to break levels down even further

can use 4 point scoring system

can aggregate into binary

remember it is ordinal

represent with stacked bar chart

dealing with unexpected situation

when to stop a participant

frequent task needs less time to complete

be diligent when doing it manually

if recording use timestamp

bar chart with confidence interval

represent number of users per range

record % of users above/below threshold

only success or failtures too

think aloud could add time

retrospective think out loud

should participants know about timing

error is action that leads to failure

useful in determining cause of failure

usefull when

error will result in big loss

result in significant cost

when error results in task failure

sometime lack of action is failure

you need to know what the correct action should be

single or multiple possible errors

capture amount of errors that occurred

or 1/0 per defined error

could present average error rate (with confidence)

capture frequency

average number of errors per user, to reduce bias

avoid double counting

code errors by type

repeated errors

cognitive or physical effort

identify action to be taken

define start and end

count the actions

must be meaningful action

automated capture much easier

actions per task

which task take most effort

lostness

N = different pages

S = total pages (count revisit)

R = optimal number of pages

L = sqrt[(N/S – 1)^2+ (R/N – 1)^2].

score above 0.5 appear to be lost

average lostness, or count of lost users

ratio of task completion rate to mean time per task

alternative total task success count / total time by participant

time and effort required to become proficient

multiple trials over different periods

when learning occurs then efficiency improves

decide on time between trials

show specific metric per trial

can aggregate tasks or show tasks individually

difference between highest and lowest point is amount of learning required

what is a trial

how many trials

idealy 3-4