Research Motivation Repertory Grid Analysis

Developing an instrument to assess the
impact of attitude and social norms on
user selection of an interface design: a
repertory grid approach
Willem-Paul Brinkman
Steve Love
Topics
 Research Motivation
 Repertory Grid Analysis – initial
instrument
 Survey – validation instrument
 Attitude, Social Norm and Design
Selection
 Conclusions & Future Research
Motivation
Ultimate aim is to understand
why people select a specific
design for their device.
Motivation
User
Personality
Intention of
selecting a
specific type
of design
Because of colour or theme
 However, correlations were relatively small
(Brinkman and Fine, 2005)
 A less direct approach is needed as users
might have different criteria to evaluate a
design.
Motivation
Attitude towards
selecting a
specific design
type
User
Personality
Relative
importance
Intention of
selecting a
specific type
of design
Subjective Norm
Adoption of Theory of
Reason Action (Ajzen
and Fishbein, 1980)
Motivation
Research questions
1. What criteria do people use to
evaluate a design?
2. What impact do attitude and social
norm have on the selection of a
design?
Motivation
Different social
context
1. Public application
(mobile phone)?
2. Private application
(PC multimedia
player)?
Motivation
Aim - Develop two evaluation
instruments to assess people’s
attitude and social norm
towards:
1. Mobile phone
2. PC Multimedia Player Skins
First question: what criteria do people
use to evaluate a phone or Multimedia
Player Skin?
Repertory Grid Analysis
 Personal Construct Theory:
Everyone interprets (or
constructs) events and their
universe differently.
 Constructs are bipolar, for
example to describe friends: easygoing versus tense, reliable
versus unreliable. etc
 Repertory Grid Analysis: a
method to elicit these specific
constructs.
George Kelly
Repertory Grid Analysis
General Procedure
 Split up triad into 2 groups
 Label the groups
 Rate the object on
construct
 Create Grid
Repertory Grid Analysis
General Procedure
 Split up triad into 2 groups
 Label the groups
 Rate the object on
construct
 Create Grid
Male
Female
Repertory Grid Analysis
General Procedure
 Split up triad into 2 groups
 Label the groups
 Rate the object on
construct
 Create Grid
Male
Female
Unusable-Usable
Large-Easy to use
Phone
1
2
3
4
5
6
Complex-Simple
 Looking for similarities
between the constructs
 Apply a Factor Analysis
Boy-Girl
General Procedure
 Split up triad into 2 groups
 Label the groups
 Rate the object on
construct
 Create Grid
Male-female
Repertory Grid Analysis
7
5
2
3
1
2
6
5
2
2
2
3
3
7
6
3
7
1
2
7
7
4
6
2
1
6
6
2
5
3
Repertory Grid Analysis
Method
 Participants
 Material
 Procedure
 20 Participants
 Brunel University Students or Staff
members
 10 males, 10 females
 Mean age of 26.5 years (SD = 4.84)
Repertory Grid Analysis
Method
 Participants
 Material
 Procedure
Photo of 15 mobile phones taken
from Mobile Digest news website.
Repertory Grid Analysis
Method
 Participants
 Material
 Procedure
Screen prints from 15 skins taken
from earlier study (Brinkman and
Fine, 2005)
Repertory Grid Analysis
Method
 Participants
 Material
 Procedure
 Two sessions: Mobile phone and
Skins
 Sequence of the sessions was
counterbalanced
 10 randomly drawn triads, with no
repeating triads
 Afterwards rating the phone and
skins on the 10 constructs
 Total interview took around 2 hours
Repertory Grid Analysis
Complex-Simple
Unusable-Usable
Large-Easy to use
Phone
1
2
3
4
5
6
Boy-Girl
 200 mobile phone and 200 skin constructs
 Factor Analysis used principal-component
extraction method and varimax rotation
 Aim -> to identify common constructs
themes used by multiple participants.
Male-female
Analysis
 Data
 Rules
 Mobile Phone
 Skins
7
5
2
3
1
2
6
5
2
2
2
3
3
7
6
3
7
1
2
7
7
4
6
2
1
6
6
2
5
3
Repertory Grid Analysis
Analysis
 Data
 Rules
 Mobile Phone
 Skins
Selection criteria to select
component
1. Factor loadings (correlations)
below 0.69 were ignored.
2. Components should have
constructs loading from at least
5 different participants
3. A clear semantic relationship
between the labels of the
construct should exist.
Repertory Grid Analysis
Analysis
 Data
 Rules
 Mobile Phone
 Skins
Appearance of
the mobile
phone
Maturity
M1 Immature – Mature
Gender
G1 Feminine – Masculine
M2 Childish – Sophisticated
G2 Female – Male
M3 Playful – Serious
G3 Girl - Boy
M4 Silly - Classy
Reliability
R1 High failure rate - Low
failure rate
Professionalism
P1 Unprofessional –
Professional
R2 Unreliable – Reliable
P2
Fun – Technical
R3 Unsound – Robust
P3
Novelty – Business
P4
Amateur - Expert
R4 Easy to break - Not
easy to break
R5 Unprotected - Protected
Ease of use
E1 Complicated - Simple to use
E2
Difficult – Plain
E3
Hard to use - Easy to use
E4
Complex – Simple
E5
Difficult to carry - Easy to carry
3 Components
/ dimensions
Repertory Grid Analysis
Analysis
 Data
 Rules
 Mobile Phone
 Skins
Progressive
PO1 Expected –
Unexpected
Fun of use
FO1 Serious – Humorous
PO2 Standard –
Fashionable
FO3 Dull – Relaxing
PO3 Conventional –
Unconventional
FO5 Bland – Vibrant
PO4 Boring – Appealing
FO2 Sombre – Lively
FO4 Depressing – Uplifting
PO5 Plain – Interesting
Futuristic
FU1 Outdated –
Ultramodern
Ease
EA1 Complex – Simplistic
FU2 Traditional – Futuristic
EA2 Scream – Peaceful
EA3 Complicated – Intuitive
EA4 Cluttered – Simple
FU3 Basic – Creative
FU4 Natural – Techie
FU5 Old - Space age
EA5 Tension - Calm
4 Components / Dimensions
Survey





Aim
Procedure
Participants
Analysis
Results
Validate the scales obtained
from the Repertory Gird
Analysis
Survey





Aim
Procedure
Participants
Analysis
Results
To evaluate the middle
mobile phone/skin with the
scales provided
Survey





Aim
Procedure
Participants
Analysis
Results



156 students of the School
of Information Systems,
Computing and
Mathematics (Brunel, UK)
Average age 23.6 years (SD
= 4.79)
57 female, 75 male (24 did
not report gender)
Survey





Aim
Procedure
Participants
Analysis
Results





Reliability analysis to
examine the internal
consistency of the scales
within dimension
Dimension “Reliability” for
mobile phone removed
Cronbach’s alpha < 0.7
Factor Analysis
Component with Eigen
value > 1
Scales loading > 0.7
Survey





Aim
Procedure
Participants
Analysis
Results
Scales for Mobile Phones
Gender
Sophistication
Feminine - Masculine
Childish - Sophisticated
Female - Male
Silly - Classy
Girl - Boy
Novelty - Business
Ease of Use
Hard to use - Easy to use
Complex - Simple
Complicated - Simple to use
Scales for Multimedia Player skins
Stimulation
Boring – Appealing
Depressing – Uplifting
Bland – Vibrant
Plain – Interesting
Ease
Complicated – Intuitive
Complex – Simplistic
Cluttered – Simple
Attitude, Social Norm and Design
Selection
Second question: What impact do attitude
and social norm have on the selection of a
design?
Attitude towards
selecting a
specific design
type
Additional
information
collecting in Grid
interviews
Subjective Norm
Intention of
selecting a
specific type
of design
Attitude, Social Norm and Design
Selection




Attitude
Social Norm
Intention
Results



The evaluation (e) of a
skin/phone on a scale.
The importance of an
evaluation scale (w).
For me, having a skin that
is “explaining”, or that is
associated with this is?
Bad
Good
A j   wi  eij 
10
i 1
Attitude, Social Norm and Design
Selection




Attitude
Social Norm
Intention
Results


Their belief (b) on how peers,
family members, or authority
figures would their
(participants) phone or skin to
score on a construct.
Their willingness (g) to comply
with the peers, family
members, or authority figure.
10


SN j    g k   bik  eij 
kO 
i 1

whereby O  peers, family, authority
Attitude, Social Norm and Design
Selection




Attitude
Social Norm
Intention
Results



Behavioural intention to select
a skin or mobile phone
I would try this on my media
player?
I intend to obtain the following
mobile phone?
unlikely
likely
Attitude, Social Norm and Design
Selection




Attitude
Social Norm
Intention
Results
Social Norm has an impact on
phone selection, but not skin
selection.
Phone (public)
Skin (private)
Mean (partial) correlation. *p.<.05.**p<.01.
Conclusions
Research questions
1. What criteria do people use to evaluate a
design?
1. Mobile phone: Gender, Ease of Use and
Sophistication
2. Multimedia Player skins: Stimulation and Ease
2. What impact do attitude and social norm
have on the selection of a design?
•
Social Norm has an impact on phone (public device)
selection, but not Multimedia Player (private device)
selection.
Further research
User properties
User
Personality
Gender
Attitude towards
selecting a
specific design
type
Relative
importance
Subjective Norm
Intention of
selecting a
specific type
of design
Questions
Thanks for your attention