Towards Customized Emotional Design: an Explorative Study of User Personality and User Interface Skin Preferences Willem-Paul Brinkman Nick Fine.

Towards Customized Emotional Design: an Explorative Study of User Personality and User Interface Skin Preferences

Willem-Paul Brinkman Nick Fine

Topics Research Motivation

Study 1: Explorative study into

personality and skin

preferences

Study 2: Exploring initial findings

with a specific set of skins

Conclusions & Future Research

Research MotivationWhy considering emotion in design?

“First, someone who is relaxed, happy, in a pleasant mood, is more creative, more able to overlook and cope with minor problems with a device - especially if it’s fun to work with….Second, when people are anxious, they are more focused, …, designer must pay special attention to ensure that all the information required to do the task is continually at hand, readily visible, with clear and unambiguous feedback about the operations that the device is performing.”

(Norman, 2004, p.26)

Research Motivation

• Emotion, Aesthetics, Fun, Self-image all Highly Personal

• Segmenting user population?

Research Motivation

Segmentation based on: Personality

traits Gender Age etc

Personality traits: “dimensions of individual differences in tendencies to show consistent patterns of thoughts, feelings, and actions” (McCrae and Costa, 2003)

Example of dimensions of individual differences are Neuroticism, Extraversion, Openness, Agreeableness, Conscientiousness.

Research Motivation

Segmentation for customized user interfaces, e.g. with skinning technology

Research Motivation

Skinning properties

Traits

Neuroticism

Extraversion

Openness

Agreeableness

Conscientiousness

ColourFont ThemeStyle

Study 1: Method Focus on skin

selection Often users scan

through a list of potential skins and select one

Selection is less based on interaction experience, but more on previous knowledge, taste and expectation.

Study 1: Method Material Measures Participants Procedure

Windows Media Player Randomly selected 61 skins from

set of 178 downloaded from the internet

4 screen shots of each skin

Study 1: Method Material Measures Participants Procedure

Intention of selecting a specific type of skin“I would try this skin on my media player”

BIS/BAS scales: Motivation towards something desirable and undesirable (Carver and White , 1994)

IPIP-NEO Big-5 Model: extraversion, agreeableness, conscientiousness, neuroticism, and openness to experience (Buchanan, Johnson, and Goldberg, 2005).

unlikely

likely

Study 1: Method Material Measures Participants Procedure

99 students and staff of School Computing, Information Systems, and Mathematics (Brunel University)

58 males, 35 females, 6 unspecified

Age between 18 and 52 years. Median 20 years

Study 1: Method Material Measures Participants Procedure

Class Room Setting

Slides projected on large screen

Participants were first show complete set (reduce learning effect)

Next, set was shown again and participants rate each skin

Example

Skin 1

Default Radio

Video CD / MP3

SKIN 1

I would try this skin on my media player. Unlikel

yLikely

Please rate the statement for the previous skin (1)

Skin 2

Default Radio

Video CD / MP3

SKIN 2

I would try this skin on my media player. Unlikel

yLikely

Please rate the statement for the previous skin (2)

Study 1: Method Material Measures Participants Procedure

Skins shown in sets of 10

Afterwards participants completed BIS/BAS scale (24 questions) Done in the class room

Participants were asked to complete IPIP-NEO inventory online and return the results afterwards 120 questions Only 17 participants return results

Study 1: Results

Factor Analysis

Did participants use limited number of dimensions to rate the skins?

Standardized Skin rating: z-score transformation per participant

Principal component as factor extraction method

19 components explaining 0.74 of overall variance

Examining varimax rotated component matrix

No clear interpretation of components

Study 1: Results

Correlation with Gender

Females were more in favour of :

Charming, ‘cute’ characters

Study 1: Results

Females were more in favour of :

Blue

Purple

Bright

Colour

Correlation with Gender

Study 1: Results

Males were more in favour of :

Muscular themes

Correlation with Gender

Study 1: Results

Males were more in favour of :

Large or complex

Correlation with Gender

Study 1: Results

Correlation with BIS/BAS

BAS: Drive

Participants more driven by pursuit of desired goals :

Over-powering, dark vs light or bright

Participants less driven by pursuit of desired goals :

Study 1: Results

Correlation with BIS/BAS

BIS

High on motivation of avoiding undesirable situations :

Blue

Low on motivation of avoiding undesirable situations :

Study 1: Results

Correlation with IPIP-NEO

Similarity-Attraction

Participants select skins with represent similar personalities (adoption of Nass and Reeves’ hypothesis)

friendliness

Not-cheerfulness

Not-neuroticism, Not-vulnerability

Study 1: Results

Correlation with IPIP-NEO

Similarity-Attraction

Participants select skins with represent similar personalities (adoption of Nass and Reeves’ hypothesis)

Depression, Self-consciousness (neuroticism)

Immoderation, excitement-seeking

Study 1: Results

Correlation with IPIP-NEO

Colour?

Extraversion, assertiveness

Gregariousness, not-anxiety, not-

anger

Neuroticism, anxiety

Not-assertiveness

Not-activity level

Study 1: Conclusions and Limitations

Potentially links such as gender and skin themes Personality and colour Similarity-attraction hypothesis

Very Preliminary Conclusions Post-hoc analysis (large number of

correlations examined) Small sample size (IPIP-NEO) Interpretation very subjective and lot of

confounding variables

Study 2: Method

Material Measures Participants

Selecting or designing 4 skins for 17 categories

1. Blue skins2. Purple/violet skins3. Grey skins4. Bright skins5. Colourful skins6. Large, dark skins7. Small skins8. Medium size, squared

shape skins9. Complex interaction skins

10. Small and friendly creature

11. Friendly, charming ‘baby’ character

12. Humoristic skins13. Action theme skins14. Predator skins15. Scary skins16. Mildly scary skins17. Standard skins

Study 2: Method

Material Measures Participants

Skins shown in sets of 17

Afterwards participants completed: reduced version of BIS/BAS scale

(15 questions) reduced version IPIP-NEO (48)

questions (only extraversion and neuroticism)

Study 2: Method

Material Measures Participants

130 Undergraduates Computer science and Information System course

83 male, 44 female, and 3 unspecified

Age: between 18 and 49 years, median 21 years.

Study 2: Results

Designing for similarity or diversity

Highest mean rating (4.85)

25% participants gave the skin their highest rating

mean rating (4.84) 35% participants

gave the skin their highest rating

Study 2: Results Analysing

category rating

Internal consistency

Positive correlations between rating within skin categories

1. Blue skins2. Purple/violet skins3. Grey skins4. Bright skins5. Colourful skins6. Large, dark skins7. Small skins8. Medium size, squared

shape skins9. Complex interaction skins

10. Small and friendly creature

11. Friendly, charming ‘baby’ character

12. Humoristic skins13. Action theme skins14. Predator skins15. Scary skins16. Mildly scary skins17. Standard skins

Study 2: Results

Correlation with age

Older participants were more in favour of Grey skins

Younger participants were more in favour of humoristic skins

Study 2: Results

Correlation with gender

Female participants were more in favour of

Friendly, charming ‘baby’ character

Small and friendly creature

Study 2: Results

Correlation with gender

Males were more in favour of

Small skins

Predator skins

Medium size, squared shape skins

Scary skins

Study 2: Results

BIS scale High on motivation of avoiding undesirable situations :

Colourful skins

Friendly, charming ‘baby’ characters

Study 2: Results

Similarity-attraction

Colourful skins negative correlation with gregariousness

Action theme positive correlations cheerfulness

Humoristic theme positive correlation cheerfulness, negative correlation depression

Study 2: Results

Colour Same skin, only differs in colour

Study 2: Results

Colour

Regression lines for skin rating from BAS-Drive scale

MANCOVA with repeated measures

Dependent variable: skin rating

Independent variable

between-subjects : (covariates) age, and personality scales; Gender

Within-subjects: skin colour

Significant two-way interaction effect between colour and BAS-Drive

1234

567

4 8 12 16

BAS- Drive

Ski

n r

atin

g

Study 2: Conclusions and limitations

Again found links such as gender and skin themes Personality and colour Similarity-attraction hypothesis

Still Very Preliminary Conclusions Interpretation skins is still subjective

(observers interpretation)

Future work

Repeat study Participants also rating skins on “skin

dimensions” such “how humoristic would you rate this skin?”

Design new skins specific for categories Systematic manipulation of skins

Extending the model

Future work

User Personalit

y

Selecting a specific type of

skinCurrent model

However, correlations are relatively small!

Future work

Attitude towards

selecting a specific skin

type

Subjective Norm

Relative importance

Adoption of Theory of Planned Control (Ajzen, 1985)

Intention of selecting a

specific type of skin

User TaskUser

Personality

Perceived Control

Questions

Thanks for your attention