T.I.M.E: A Method of Detecting The Dynamic Variances of Trust

T.I.M.
E: A Method of Detecting the

Dynamic Variances of Trust
Laurian Vega
Yeong-tay Sun, D. Scott McCrickard, Steve Harrison
Virginia Tech
Tuesday, April 27, 2010

What does
focusing on
the design Hassanein, K.S. and
M.M. Head. Building
Trust Through Socially
elements of a
Rich Web Interfaces
website tell
us about user
trust?
Laurian Vega, Virginia Tech, WICOW April 2010
Increased
Social
Increased Presence
Social
Presence
Hassanein, K.S. and

M.M. Head. Building
Trust Through Socially
A v. B studies
Rich Web Interfaces

Lumsden, J. and L.
MacKay. How Does
Personality Affect Trust
in B2C e-Commerce?
Will ‘trust’ logos affect

user trust?
A v. B studies
Yang, Y., Hu, Y., and
Chen, J., A Web Trust-
Inducing Model for e-
Commerce and Empirical
Research; Jarvenpaa, S.,
Tractinsky, N., and
Saarinen, L., Consumer
Trust in an Internet Store
Will a strong brand name affect

user trust?
Will ‘trust’ logos affect user trust?
A v. B studies
-./01"2341"
Website Webpage
23.561"7851" 9:3;1<"
Information
'!!!"
&!!"
%!!"
$!!"
#!!"
!"
!(!)(!'"
!(''(#'"
!('+(,'"
!('&(!&"
!('*(!,"
!(#!(!'"
!(!!(!!"
!(!'('!"
!(!#($$"
!(!$($%"
!(!%(!!"
!(!*(#%"
!('$(,#"
How does one user’s trust
change when interacting
with a website?
Today’s Talk
Micro-
Trust & Evaluating
interactions
Credibility Trust
with Websites
Method Results Future Work

Trust & Credibility
Credibility & Trust:

Trust
user-held perceptions
based on a history of
interaction with an
environment. Credibility

Dimensions of Trust
Trust
Consistency Persistence
Competence Fiduciary Responsibility
Integrity
Loyalty
Benevolence
Motives Openness
Reliability
Ability Concern
Discreetness Sincerity
Judgement Intentions
Expertise Faith Understanding
Lee, J.D. and K.A. See, Trust in Automation: Designing for Appropriate Reliance. Human Factors, 2004. 46(1): p. 50-80.

Trust in Different Socio-Technical
Systems

A User is One Part of a Socio-
Technical System
Her Experience
User
Website
Webpage
Information
V1
Her V2
Settings V3

A User is One Part of a Socio-
Technical System
Her Experience
User
Website
Webpage
Information
V1
Her V2
Settings V3

Today’s Talk
Micro-
Trust & Evaluating
interactions
Credibility Trust
with Websites

Evaluating Trust
Qualitative Methods
• Open-ended Questionnaires
• Diaries Did you trust the
• Observations information provided on
• Open-ended Questions & the web site?
Interviews
• Website & Email Logs
• In-person Surveys
• Group Interviews &
Discussion Fisher, J., F. Burstein, K. Lynch and K.
Lazarenko, "Usability plus usefulness =
trust": an exploratory study of Australian
health web sites.

Evaluating Trust
Qualitative Methods
Greenhalgh, T., G.W. Wood, T. Bratan, K.
• Open-ended Questionnaires Stramer and S. Hinder, Patients' attitudes to

the summary care record and HealthSpace:
• Diaries Qualitative study.
• Observations
• Open-ended Questions &
Interviews "Why do you or don't you
• Website & Email Logs trust them?... What makes
• In-person Surveys you trust them?"
Discussion

Evaluating Trust
Qualitative Methods
• Open-ended Questionnaires
• Diaries ●Problems●
• Observations •Analytical Bias
• Open-ended Questions & •Researcher dependent
Interviews interpretations
• Website & Email Logs •Large time commitment to
analyze
• In-person Surveys
Discussion

Evaluating Trust
Quantitative Methods
• Surveys
• Interviews "Did you trust the health
• Online Questionnaires information you found?"
• Phone Surveys Possible answers: "Yes;
Somewhat; No"
Semere, W., H.L. Karamanoukian, M. Levitt,

T. Edwards, M. Murero, G. D'Ancona, et al.,
A pediatric surgery study: Parent usage of the
Internet for medical information.

Evaluating Trust
• Surveys 5-point Likert-scale: "I
• Interviews can trust the records
• Online Questionnaires provided in information
• Phone Surveys technology service (the
online appointment system/
electronic medical record
system)."
Chang, H.H. and C.S. Chang, An assessment

of technology-based service encounters &
network security on the e-health care systems
of medical centers in Taiwan.

Evaluating Trust
• Surveys
• Interviews ●Problems●
• Online Questionnaires •Delayed & decoupled from
• Phone Surveys experience
•Fallible user recollection
•A v. B Studies

A New Method:T.I.M.E
•Trust •Measurement
Understanding a user’s trust as Taking repeated measurements
they interact with a website, a of a user’s trust in the web-
webpage, and the information system
within it
•Incremental •Evaluation
Using incremental Assessing how a user’s trust
measurements of trust to assess changes and the factors that
change affect it

within it
change affect it

within it
change affect it

within it
change affect it

Today’s Talk
Micro-
Trust & Evaluating interactions
Credibility Trust with
Websites

Micro-Interactions
“We live in a world where the little things really do
matter. Each encounter no matter how brief is a micro-
interaction which makes a deposit or withdrawal from
our rational and emotional subconscious. The sum of
these interactions and encounters adds up to how we
feel about a particular product, brand or service. Little
things. Feelings.” ~ David Armano
e.g., clicking a link, going to the previous webpage,

scrolling

Sample Micro-Interaction






Measuring Micro-interactions &
Trust
Repeated Measuring:
How Often? Frequency of measurement
v. Annoying participant too
Pairing Actions with much v. Predictability
Trust: What Action? Result: At least 30 seconds

from prior click in interface
Discrete Measures to
Detect Change: Scale?
Factors affecting Trust:

Which ones?

Trust
Repeated Measuring:
How Often?
Every time segment v.
Pairing Actions with Pairing with Action
Trust: What Action?
Result: screen remains
viewable but system
suspends interaction

Which ones?

Trust
Repeated Measuring:
How Often?
Pairing Actions with Detecting change v.

Trust: What Action? Meaningful result
Discrete Measures to Result: No prior measures

displayed & High precision
scale

Which ones?

Trust
Repeated Measuring: Whole socio-tech system v.
How Often? focused part
Pairing Actions with Result: 3 parts of web-system
Trust: What Action?

•How much do you trust this
entire website?
•How much do you trust the
Detect Change: Scale? information on this page?
•How much do you trust that
Factors affecting Trust: this information will guide
Which ones? you to your answer?

Pairing Actions with Trust
30 seconds have
passed

!

Today’s Talk
Micro-
Trust & Evaluating
interactions
Credibility Trust
with Websites

Method
1 hour sessions
20 participants, 12 male; all university students, 9 graduate

students
Undergraduates received extra credit
20 minutes with each website
4 tasks for each website; pre-tested to take at least 20 minutes
eg: “Who was the first female presidential candidate?”

Two Websites to Evaluate: The Usability Case Study Library +
Wikipedia

Today’s Talk
Micro-
Trust & Evaluating
interactions
Credibility Trust
with Websites

Clicking & Responses
Presented ~ 39 seconds
No difference between Usability Case Study Library &
Wikipedia
Presented ~ 59 times
Usability Case Study Library ~26 times
Wikipedia ~25 times
1,018 Responses for each question
~ 10 seconds answering form
Usability Case Study Library ~4min, 27 seconds out of 20min
Wikipedia ~ 3min, 38 seconds out of 20min
Only one participant finished all tasks

Trust Effects
• Gender & Education Effects
Women &
• Website Effect
Undergraduates more
trusting (p<0.01)
• Variation Effect
• Time Effect
Fogel, J. and E. Nehmad, Internet social
network communities: risk taking, trust,
and privacy concerns; Maddux, W.W. and
M.B. Brewer, Gender differences in the
relational and collective bases for trust.

Trust Effects
• Website Effect Wikipedia was

trusted more
UCS trust: 495;
Wikipedia 617 (out
• Time Effect of 1000)

Trust Effects
User Trust for Each Website
850
UCS Wikipedia
• Website Effect 680
• Variation Effect 510
• Time Effect 340

~ 10% increase or
decrease for all 170
three questions
0
Website Webpage Information

Trust Effects
One Standard Deviation of Websites
300
UCS Wikipedia
• Website Effect
225
150
• Time Effect
~ 10% increase or
75
decrease for all
three questions
0
Website Webpage Information

Trust Effects
• Gender & Education Effects User Trust in 5 Minute Segments
700
UCS Wikipedia
• Website Effect
525
350
• Time Effect
Trust is
higher in 175
the first 5
Minutes
0
0-5 5-10 10-15 15-20

-./01"2341"
Website 23.561"7851"
Webpage 9:3;1<"
Information
'!!!"
&!!"
%!!"
$!!"
#!!"
!"
!(!)(!'"
!(''(#'"
!('+(,'"
!('&(!&"
!('*(!,"
!(#!(!'"
!(!!(!!"
!(!'('!"
!(!#($$"
!(!$($%"
!(!%(!!"
!(!*(#%"
!('$(,#"
How does one user’s trust
change when interacting
with a website?
Today’s Talk
Micro-
Trust & Evaluating
interactions
Credibility Trust
with Websites

-./01"2341"
Website 23.561"7851"
Webpage 9:3;1<"
Information
'!!!"
&!!"
%!!"
$!!"
#!!"
!"
!(!)(!'"
!(''(#'"
!('+(,'"
!('&(!&"
!('*(!,"
!(#!(!'"
!(!!(!!"
!(!'('!"
!(!#($$"
!(!$($%"
!(!%(!!"
!(!*(#%"
!('$(,#"
Why is trust in a website v. trust
in a webpage statistically higher?

-./01"2341"
Website 23.561"7851"
Webpage 9:3;1<"
Information
'!!!"
&!!"
%!!"
$!!"
#!!"
!"
!(!)(!'"
!(''(#'"
!('+(,'"
!('&(!&"
!('*(!,"
!(#!(!'"
!(!!(!!"
!(!'('!"
!(!#($$"
!(!$($%"
!(!%(!!"
!(!*(#%"
!('$(,#"
What happened here? What can
website designers do about this?

What does this mean for
understanding a user’s
trust in a website?

Future Work
Recreating the experiment focusing on health websites to

isolate what factors affect trust.
Looking at video to tie decreases and increases in trust to

experiences.
Exploring what trust means in the larger socio-technical

system by looking at childcares and general medical practices.

Conclusions
There are many ways to measure trust, but none that attempt
to evaluate how trust changes.
T.I.M.E. Method collects measurements of trust at a small

enough interval to detect change.
Trust in different parts of the socio-technical system have

different levels of trust, even within a web-system

Thank You
Thank you to all my co-authors: Yeong-tay Sun, Steve

Harrison, Scott McCrickard
Thanks to Enid Montague for her help
And thanks to Jessica Corker for helping on the future work!
www.laurianvega.com
Twitter: @LadyLaurian
Virginia Tech

T.I.M.E: A Method of Detecting The Dynamic Variances of Trust

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

T.I.M.E: A Method of Detecting The Dynamic Variances of Trust

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

T.I.M.E: A Method of Detecting The Dynamic Variances of Trust

Uploaded by

Copyright:

Available Formats

T.I.M.

E: A Method of Detecting the

Tuesday, April 27, 2010

Hassanein, K.S. and

Laurian Vega, Virginia Tech, WICOW April 2010

Will ‘trust’ logos affect

Will a strong brand name affect

Method Results Future Work

Laurian Vega, Virginia Tech, WICOW April 2010

Credibility & Trust:

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Method Results Future Work

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

• Open-ended Questionnaires Stramer and S. Hinder, Patients' attitudes to

• Diaries Qualitative study.

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Semere, W., H.L. Karamanoukian, M. Levitt,

Laurian Vega, Virginia Tech, WICOW April 2010

Chang, H.H. and C.S. Chang, An assessment

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Method Results Future Work

Laurian Vega, Virginia Tech, WICOW April 2010

e.g., clicking a link, going to the previous webpage,

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Laurian Vega, Virginia Tech, WICOW April 2010

Trust: What Action? Result: At least 30 seconds

Factors affecting Trust:

Laurian Vega, Virginia Tech, WICOW April 2010

Factors affecting Trust:

Laurian Vega, Virginia Tech, WICOW April 2010

Pairing Actions with Detecting change v.

Discrete Measures to Result: No prior measures

Factors affecting Trust:

Laurian Vega, Virginia Tech, WICOW April 2010

Pairing Actions with Result: 3 parts of web-system

Trust: What Action?

Laurian Vega, Virginia Tech, WICOW April 2010

Pairing Actions with Trust

Pairing Actions with Trust

Method Results Future Work

Laurian Vega, Virginia Tech, WICOW April 2010

20 participants, 12 male; all university students, 9 graduate