Transcript User Interface Design -- An Experimental Study John Pourdehnad, et al. Ackoff Center for Advancement of Systems Approaches (ACASA) University of Pennsylvania Philadelphia, Pennsylvania American.

User Interface Design -- An Experimental Study
John Pourdehnad, et al.
Ackoff Center for Advancement of Systems Approaches (ACASA)
University of Pennsylvania
Philadelphia, Pennsylvania
American Society For Cybernetics
Slide 1
Washington DC, 27 - 30 October 2005.
User Interface Design -- An Experimental Study
Barry G. Silverman1
John Pourdehnad1
Gnana Bharathy1
Melanie C. Green2
Joyce A. Salisbury3
1Ackoff
Center for Advancement of Systems Approaches (ACASA), University of
Pennsylvania
2Department of Psychology, University of Pennsylvania
3Market Research, General Motors Corporation -- GM
American Society For Cybernetics
Slide 2
Washington DC, 27 - 30 October 2005.
Introduction



Americans are going online to conduct such day-to-day activities as
business transactions, personal correspondence, research and
information gathering, and shopping.
It is no longer good enough to rely on generalized visual library and
hypermedia principles to support all these activities as if they were
the same.
There has been little time to study these designs and how they
impact consumers.
American Society For Cybernetics
Slide 3
Washington DC, 27 - 30 October 2005.
The Challenge



It is vital to develop a better understanding of how web designs
facilitate consumer needs (or not), and to assess the role of
individual differences and whether designs that reflect such
differences provide improved service.
There are many consumer-oriented websites, yet the science of
website design is relatively immature.
There are few scientific principles upon which to base such designs,
although many designs are used in practice.
American Society For Cybernetics
Slide 4
Washington DC, 27 - 30 October 2005.
The Research


A number of different approaches to user interface design improvements
were considered.
Of the various possible approaches to human computer interface (HCI)
design, the most common practice in the past, and to a great extent today,
is to assume one uniform user group with similar characteristics, needs, and
preferences.
– This approach usually requires an iterative design procedure to minimize the
differences between users and the system.

Another design approach is to assume different user groups with different
characteristics, needs, and performances who will be using the system.
– This approach requires a careful examination of the population in order to
identify such groups, as well as different interface modules for the same
service/product.

The approach taken in this study was to assume a null hypothesis that there
are no differences among the users (although our belief is the opposite) and
to try to disprove that theory. For this purpose, the following research tasks
were embarked on:
American Society For Cybernetics
Slide 5
Washington DC, 27 - 30 October 2005.
Research Tasks






Development of a model to study HCI (this entailed creating a
structured model of the intended user),
Development of instruments for measurement,
Validation of the instruments,
Application of instruments to test the model,
Analysis of the results obtained through the application of the model,
and,
Development of recommendations for the use of the model.
American Society For Cybernetics
Slide 6
Washington DC, 27 - 30 October 2005.
Development Of A Model To Study HCI


We examined individual differences that were relevant to human
computer interaction. Furthermore, we investigated possible
methods of evaluating these differences.
These differences can be categorized as:
– Physiological: Cerebral Hemisphericity, Vision, Hearing, and
Mobility/Dexterity
– Psychological: Intelligence, Cognitive Style, and Personality
– Sociocultural: Language, Culture, and Environmental

From the above list we have determined that the most important
category for the current ACA website challenges is the
Psychological
American Society For Cybernetics
Slide 7
Washington DC, 27 - 30 October 2005.
Development Of A Model To Study HCI (Continued)




Furthermore, we decided that an examination of cognitive
differences was in order, based on initial findings from secondary
research (literature review) that suggest that they are the most
significant with regards to human computer interaction.
However, in our research and consultation with subject matter
experts, we found that Need for Cognition, rather than cognitive
style, can be a vital factor.
Need for cognition (NFC) is an individual difference measuring how
much people like to think. It is a motivational variable, and should
not be confused with cognitive ability or intelligence
The dynamic that we were overlooking was preference (motivation)
to purchases; specifically, the differences between those individuals
characterized as Utilitarian (focusing on practical aspects of a
purchase) vs. those characterized as Lifestyle seeking (wanting to
convey a particular image or impress others through a purchase).
American Society For Cybernetics
Slide 8
Washington DC, 27 - 30 October 2005.
Development Of A Model To Study HCI (Continued)

With these two dimensions, it was possible to hypothesize a two-bytwo matrix with the following four quadrants or cells:
–
–
–
–
Low Need for Cognition/Utilitarian Oriented,
Low Need for Cognition/Lifestyle Oriented,
High Need for Cognition/Utilitarian Oriented, and
High Need for Cognition/Lifestyle Oriented.
Low Need for
Cognition
High Need for
Cognition
Lifestyle
LL
HL
Utilitarian
LU
HU
American Society For Cybernetics
Slide 9
Washington DC, 27 - 30 October 2005.
Formulation Of The Hypothesis To Be Tested


In summary, we would like to confirm whether the web users could
be segmented into one of four cells in the matrix shown above
In order to test this, we have formed the following hypotheses:
Null Hypothesis
(H0):
There is no statistically significant difference in the need for
cognition and preference function scores of the people between
the four classes (quadrants). That is, the variances within the
classes are not significantly smaller than the variance between
the classes
Alternative
Hypothesis
(Ha):
There is a significant difference in the need for cognition and
preference function scores of the people between the four
classes (quadrants). That is, the variances within the classes
are significantly smaller than the variance between the classes.
American Society For Cybernetics
Slide 10
Washington DC, 27 - 30 October 2005.
Development Of Instruments For Measurement

As a measure of Need for Cognition, we used the Need for
Cognition Scale in Petty et.al. (1984):
1. I would prefer complex to simple problems.
2. I like to have the responsibility of handling a situation that requires a lot
of thinking.
3. Thinking is not my idea of fun.
4. I would rather do something that requires little thought than something
that is sure to challenge my thinking abilities.
5. I try to anticipate and avoid situations where there is likely a chance I
will have to think in depth about something.
6. I find satisfaction in deliberating hard and for long hours.
–
For the preference function dimension, there is no existing
instrument. Instead, we had to construct our own questionnaire that
would allow us to identify an individual’s preferences that form the
basis for their purchasing decisions based on their orientation
towards utility or lifestyle.
American Society For Cybernetics
Slide 11
Washington DC, 27 - 30 October 2005.
Development Of Instruments For Measurement

Lifestyle Oriented
need for
Need for Feeling (LF): someone with a high
emotional attachment to a product.
Need for Form (LF): someone with a high need for
aesthetics and style.
Need for External Status (LE): someone with a high
need to feel wanted, loved, and cared for by a peer
group.
Need for Internal Status (LI): someone with a
high need for approval, good opinion, and
regard

Utilitarian Oriented
American Society For Cybernetics
Need for Function (UF): someone who focuses
on the actions and activities assigned to,
required of, or expected of a product.
Slide 12
Washington DC, 27 - 30 October 2005.
Development Of Instruments For Measurement











I purchase things I love (LF)
I want to be liked by others (LI)
I buy things that I like, regardless of current trends (UF)
I often purchase things I don't need (UF)
I compare myself to others (LE)
What you think of yourself is reflected by what you wear (LI)
I choose products based on look and feel (LF)
I'm not as concerned about fashion as I am about wearability (UF)
I like to impress my peers (LE)
I like form over function (UF)
I like belonging to groups or organizations (LE)
American Society For Cybernetics
Slide 13
Washington DC, 27 - 30 October 2005.
Survey

GM then administered the on-line survey to the following groups of
potential customers:
– Group I consisted of 112 participants of 2 general GM online communities
(CCO/CCOC) set up to conduct market research. They are new vehicle
shoppers and are approximately 50% male and 50% female, with half
under the age of 45 and half 45 and older. The sample is all from the Los
Angeles, California area., and
– Group II consisted of 69 consumer, also part of a GM online community
(CEO), who are interested in cutting edge technology, alternative fuels,
hybrid vehicles, or the environment. This group has higher income and
more education that the traditional new car shopper.
– The sample is also skewed approximately 75% male. In addition, the
participants in this group are willing to pay at least $35K for a hydrogenpowered vehicle at the point of recruitment. Half live within 100 miles of
Los Angeles, CA and the other half live within 100 miles of Philadelphia, PA
(with a few outliers)
American Society For Cybernetics
Slide 14
Washington DC, 27 - 30 October 2005.
Results
Results of the Survey for Group I
Low Need for Cognition /
Lifestyle Oriented
64
High Need for Cognition /
Lifestyle Oriented
Preference Function
54
44
34
24
Low Need for Cognition /
Utilitarian Oriented
High Need for Cognition /
Utilitarian Oriented
14
18
28
38
48
58
68
78
88
Need for Cognition
American Society For Cybernetics
Slide 15
Washington DC, 27 - 30 October 2005.
Results
Results of the Survey for Group II
Low Need for Cognition /
Lifestyle Oriented
High Need for Cognition /
Lifestyle Oriented
64
Preference Function
54
44
34
24
Low Need for Cognition /
Utilitarian Oriented
High Need for Cognition /
Utilitarian Oriented
14
18
28
38
48
58
68
78
88
Need for Cognition
American Society For Cybernetics
Slide 16
Washington DC, 27 - 30 October 2005.
Number Of Subjects Categorized Into Each Type
Using Short Survey
# of Subjects
American Society For Cybernetics
Typology
Group I
Group II
LU
20
36
LL
43
10
HL
32
22
HU
9
1
Total # of Subjects
104
69
Slide 17
Washington DC, 27 - 30 October 2005.
Application of the Model to Solve the Problem



Differentiating the Web Interface based on Typology
By studying the unique needs of individuals in each cell, six design
criteria were identified and, based on the secondary research,
requirements for each website interface were developed.
Based on the characteristics developed for each cell, four mockups
were designed whose overall aesthetic and usability should attract,
satisfy, and retain the users in each designated cell, though this
remains to be more fully evaluated by future research.
– For that, as mentioned earlier, we propose to use focus groups to
collect assessments of and refinement ideas for the design mockups
presented in this section.

The merits of the four different mockup designs were preliminarily
examined and as a result three new web user interface options
were developed for possible structures including screenshots and a
brief description.
– It is possible to demonstrate that with any of these options a user can
obtain the necessary assistance in selecting a vehicle while providing
valuable information to the organization.
American Society For Cybernetics
Slide 18
Washington DC, 27 - 30 October 2005.
Design Tailored Interfaces To The Cognition Needs
And Preference To Purchases
High Need for
Cognition (H):
An individual who is highly analytical and enjoys being presented with
a lot of data and the opportunity to process that data. It seems
reasonable to conclude that these people prefer charts and graphs.
Low Need for
Cognition (L):
An individual who focuses less on analytical tasks and prefers not to
be overrun with data. It is likely that these people prefer pictures and
audio representations.
Lifestyle Oriented or
High Preference
Function Score (L):
Someone who believes that the value of a thing depends on how it
reflects the person's values and preferences.
Utilitarian or High
Preference Function
Score (U):
Someone who believes that the value of a thing depends on its
practical usefulness.
American Society For Cybernetics
Slide 19
Washington DC, 27 - 30 October 2005.
Guidelines for User Requirement
Categories
Low Cog/Utilitarian
Low Cog/Life
Style
High Cog/Life Style
High Cog/Utilitarian
Text or pictures
Pictures
Pictures
Charts
Charts
Options presentation
Traditional Options
Collection of
Products
Graph of Value
Graph of Options
Persona preference
Comedy/Wacky Alien
(Alf)
Secret Agent
Butler/Celebrity (Paul
Newman)
Abstract Narrator (Hal 9000)
Complexity
Sorting info/nonlinear
Presented
info/linear
Presented info/linear
Sorting info/nonlinear
Final action
Place car in virtual
“garage”
Drive car
away/get in
and go
Have car delivered
Add car to garage
Preferred imagery
- classic look
- pictures of the product in
use
- art/design look
- customization of product
- prefer to see car in
intended
environment
- non artsy look
- cut-through diagrams of product
American Society For Cybernetics
- extreme look
- pictures of
gaudy and
sexy people
posing with
product
Slide 20
Washington DC, 27 - 30 October 2005.
Differentiating The Web Interface Based On Typology




By studying the unique needs of the individuals in each cell, we
hope to create four website interfaces whose overall aesthetics and
usability will attract, satisfy, and retain the users in each designated
cell.
If this is successful, we will confidently be able to create an
“adaptable” website that will effectively cater to all potential
audiences.
To this point we have mockups of all of the different cells.
Our framework follows the guidelines laid out in Table below:
American Society For Cybernetics
Slide 21
Washington DC, 27 - 30 October 2005.
Introduction Pages For Each Type
American Society For Cybernetics
Slide 22
Washington DC, 27 - 30 October 2005.
The Low Cognition/Lifestyle
American Society For Cybernetics
Slide 23
Washington DC, 27 - 30 October 2005.
The High Cognition/ Utilitarian
American Society For Cybernetics
Slide 24
Washington DC, 27 - 30 October 2005.
The Low Cognition/Utilitarian
American Society For Cybernetics
Slide 25
Washington DC, 27 - 30 October 2005.
The High Cognition/ Lifestyle
American Society For Cybernetics
Slide 26
Washington DC, 27 - 30 October 2005.
Development Of Recommendations For The Use Of Models

Option #1: This option entails the use of a survey technique and the immediate
direction of a visitor to the HCI that corresponds to the appropriate cell based
on the survey. Upon completing the survey, we envision that a visitor would be
directed to the appropriate introduction page and follow along with one of the
enclosed sets included in previous slides.

Option #2: This option would present the visitor with all four options for each
step along the way and would allow the user to select the desired interface for
each step along the way. This method calls for a display much like that seen
above for each step in the process from Introduction to a final analysis of one
particular vehicle.

Option #3: This option is based on an extension of what we have seen recently
on such services as America Online and MSN. This option allows the user to
use the designated interface as prescribed by some survey or self-selection
technique, but also allows the individual to switch to one of the other available
options in the event that they choose to do so, yet does not force a choice at
each step.
American Society For Cybernetics
Slide 27
Washington DC, 27 - 30 October 2005.
Further Research Questions




To undertake additional research to support the premise that
different categories of individuals should be presented with different
interfaces;
To develop an improved way of determining the user’s profile;
To determine an exact design for the interface, using focus groups;
and
To develop an associated decision support system for each
interface.
American Society For Cybernetics
Slide 28
Washington DC, 27 - 30 October 2005.