[email protected] 781.891.2500 | www.bentley.edu/usability Bill Albert, Ph.D. Executive Director, Design and Usability Center Bentley University June 6, 2012 How Quick Are We to Judge? A Case Study of Trust and Web Site Design
May 06, 2015
[email protected] 781.891.2500 | www.bentley.edu/usability
Bill Albert, Ph.D. Executive Director, Design and Usability CenterBentley UniversityJune 6, 2012
How Quick Are We to Judge? A Case Study of Trust and Web Site Design
2
• Background
• Trust and web design
• Pre-attentive processing
• Our research
• What it means for you
• Discussion
Agenda
Background
4
• I love the science behind the user experience• I am data driven!• I create connections between research and practice
About Me
5
• Established in 1999
• Independent center within the Human Factors in Information Design graduate program at Bentley’s McCallum School of Business
• Dual mission to support the university and provide an education experience to graduate students
• Solutions delivered for over 200 clients
Bentley DUC
6
Trust is the foundation of the user experience
Trust is not always conscious
Motivation
Trust and Web Design
8
Do you trust this person?
Trustworthiness in motion
http://www.princeton.edu/main/news/archive/S21/79/44O45/
9
Do you trust these websites?
10
What about these sites?
11
What does trust with technology look like?
• The system must act in a reliable and predictable way
• The system must be transparent about intentions and goals
• The system must not take advantage of the user
• The system must be truthful – delivering promises
• There must be an element of risk on behalf of the user
12
10 design elements that impact trust
1. Content• Tone• Relevance• Timeliness• Typos• Clarity
2. Consistency• Visual• Interaction
(navigation and forms)
3. Color
4. Layout• Information density• White space
www.georgehutchins.com
13
10 design elements that impact trust
5. Design elements• Buttons/links/call to
action• Images• Visual treatment
6. Changing behavior• Default selections• Visual prominence
7. Ads• Type (especially
animated)• Location• Content
www.lingscars.com
14
10 design elements that impact trust
8. Contact information• Clarity• Ease of use• Comprehensiveness
9. Community• Photos• Testimonials• Affiliations
10. Logo / Brand / Company
www.bostonbizlab.com
15
Work by BJ Fogg
Presumed credibilityReputed credibilitySurface credibilityEarned credibility
perceived trustworthiness
perceived expertise
perceived credibility
16
Questions about trust and design
• When do we begin to form opinions about trust?
• How dynamic is trust?
• Are their “design primitives” that impact trust?
Pre-Attentive Processing
18
What is pre-attentive processing?
• Precedes “focused attention”
• “Effortless” or “automatic”
• < 200ms
• Basic features that are processed pre-attentively include: colors, closure, contrast, size, flicker, orientation
• Judgments about faces such as attractiveness, trustworthiness, and sexual orientation (men)
19
Lindgaard et al (2006)
20
Lindgaard et al (2006)
• In Experiment 1 participants judged attractiveness of 100 web pages at 500ms in two phases – very highly correlated
• Experiment 2 focused on 50 of the most/least visually appealing pages, and included other dimensions of visual attractiveness – same results as Experiment 1
• In Experiment 3 there were two group who judged 40 pages at either 50ms or 500ms – both groups were consistent in their judgment of visual appeal
• They conclude that people form reliable opinions about visual appeal after only a brief exposure
Lindgaard, Fernandez, Dudek & Brown (2006). Attention web designers: You have 50ms to make a good first impression! Behavior & Information Technology, 25(2), 115-126.
Our Research
22
Experiment 1
• Based on HFES paper with Dr. Bill Gribbons
• Does trust form on a pre-attentive level?
• A total of 72 participants took part in the study• Recruited through the uTest panel, compensated $10• Reside in US, UK, Canada, India, and The Netherlands• 8 participants were removed due to satisficing behavior
• 50 home page screenshots (25 financial, 25 healthcare)
• Within-subjects design• Each participant rated trust on the same screenshot in two trials• Screenshots presented randomly in each trial
23
Experiment 1 Procedure
Screenshot(50 ms)
Mask (150 ms)
Blank Screen
(1000 ms)
Trust Assessment
(Up to 10 seconds)
Blank Screen
(1000 ms)
24
Experiment 1 Results
Trust Assesments for all 50 Web Sites by Trial (Expt 1)
4.0
4.5
5.0
5.5
6.0
6.5
4.0 4.5 5.0 5.5 6.0 6.5
Trial 1 (avg)
Tria
l 2 (a
vg)
• Significant correlation between the average trust assessments in trials 1 and 2 (r = 0.81, p < .001)
25
Experiment 1 Results (contd)
Correlation Coefficient # of Participants % of Participants<.10 11 17.2%.10 - .19 13 20.3%.20 - .29 12 18.8%.30 - .39 13 20.3%.40 - .49 4 6.3%.50 - .59 6 9.4%.60 - .69 3 4.7%.70 - .79 1 1.6%.80 - .89 1 1.6%.90 - 1.0 0 0.0%
Significance # of Participants % of Participantsp<.01 21 32.8%p<.05 10 15.6%p<.10 5 7.8%p>.10 28 43.8%
• 48% of participants (31 out of 64) exhibited a significant correlation in their trust assessments
26
Experiment 2
• Concerns about satisficing behavior in Experiment 1
• Experiment 2 was conducted in a classroom setting
• Same procedure as Experiment 1
• Eleven participants • Graduate students in the Human Factors & Information
Design Program at Bentley University
27
Experiment 2 Results
• Significant correlation between trust assessments in trials 1 and 2 (r = 0.76, p<.001)
• 64% of participants (7 out of 11) exhibited a significant correlation in their trust assessments
Trust Asesments for all 50 Web Sites (Expt 2)
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5
Trial 1 (avg)
Tria
l 2 (a
vg)
Significance # of Participants % of Participants
p<.01 7 63.6%
p<.05 0 0.0%
p<.10 1 9.1%
p>.10 3 27.3%
28
Summary of our research
• Individuals are capable of processing trust on a pre-attentive level
• Approximately one-half of participants in Experiment 1 and two-thirds of participants in Experiment 2 demonstrated a consistent level of trust assessments
• Significant correlation between trust assessments in the first and second trials when averaging across participants
• Preconscious mind may play a larger role in how websites are judged than previously believed
29
Next Steps
• What is the relationship between trust assessments on a pre-attentive and conscious level? (Stage 2)
• What are the specific design attributes that impact pre-attentive trust assessments? (Stage 3)
• What are the effects of trust/distrust primes on subsequent cognitive acts? (Stage 4)
• Are there other aspects of the user experience that may be developed pre-attentively?
30
Stages of trust in web design
• <200ms• Based on
design primitives (color, density, layout)
Exposure
• < 3 seconds• Based on brief
inspection of specific elements
Inspection• > 3 seconds• Based on an
interaction with specific elements
Interaction
• Based on multiple interactions
• Takes into account the entire design
Opinion
What It Means For You
32
Why it matters
• Trust is a key element in web design that impacts short-term and long-term behaviors and perceptions
• Trust directly impacts:• Sales• Customer loyalty• Support costs• Session duration• Satisfaction• Ease of use• Efficiency of use
33
What you should do
• Review designs prior to launch and on a periodic basis• Brief exposures “blink tests” to evaluate “pre-attentive” trust• Deep dive usability evaluations to identify specific elements• Quantitative-based user research to validate the design
• Look at the design as a dialogue with the user
34
Wrap up
Sweat the little things…
It makes a big difference
Questions?
Bill Albert, PhDExecutive DirectorDesign and Usability Center, Bentley [email protected]@UXMetrics | @BentleyDUChttp://usability.bentley.edu