Web Credibility Research

1
Web Credibility Research

• A Method for Online Experiments
• Some Early Study Results

B.J. Fogg, Jonathan Marshall, Tami Kameda, Joshua
Solomon, Akshay Rangnekar, John Boyd, Bonny
Brown Stanford Persuasive Technology
Lab www.captology.org
2
Just last week, on March 26th, we saw a
provocative Web page . . .
3
(No Transcript)
4
Why Study Web Credibility?

• Many Web sites contain low-quality information.
• User experience is affected by the credibility of
  Web sites.
• Research is still scarce.
• What causes people to believe--or not
  believe--what they find on the Web?

5
Our Research Goals

• Develop a method for running controlled
  experiments online.
• Study how specific Web site factors affect
  perception of Web credibility.

6
The Overall Method

• Design the online experiment.
• Find participants with help from Vividence.
• Randomly assign participants to an experimental
  condition when they arrive at www.webresearch.org

7
Step 1
Informed Consent
8
Step 2
View Page
September 30, 1997
9
Step 3
AnswerQuestions
10
Results from early online studies

• 1 - Banner ads credibility
• 2 - Attributions credibility
• A. Author photographs
• B. Author names
• 3 - Timeliness of info credibility
• We pilot tested an early version with a CHI 2001
  audience
• We got good results --gt the pilot study served
  its purpose.
• The method wasnt rigorous enough to give
  publishable results.

11
Study 1 Banner Ads

• How do banner ads change perception of Web site
  credibility?
• Low reputability ad vs. high reputability ad
• 164 participants

12
The Banner Ads
Low Reputability Ad
High Reputability Ad
13
Study 1 Some results
14
Study 1 Conclusions

• Banner ads influence how people perceive Web
  information.
• They are not always transparent to the user.
• Some banner ads can reduce perceived credibility
  of web page.

15
Study 2 Attribution

• How do attribution elements affect the
  credibility of Web content?
• Study 2a Author photographs
• Formal photo, casual photo, no photo.
• Study 2b Author names
• Formal name, casual name, no name.
• 484 participants

16
Sample attribution pages
17
Study 2a Results
18
Author Photo Results
casual photo
no photo
formal photo
believable
trustworthy
competent
credible
unbiased
expert
COMPOSITE
19
Study 2a Conclusions

• Author photographs have significant effect on the
  credibility of content.
• Specifically, given this content, formal author
  photograph increased content credibility compared
  to casual author photograph.

20
Study 2b Results
21
Study 2b Conclusions

• Authors name
• limited effect on credibility rating of content.
• Oddly enough . . .
• The casual byline perceived as more believable
  than formal byline
• No obvious explanation for this result.

22
Wrap up

• Online Research Method
• Fast, focused, wide reach.
• Early Findings on Content
• Certain banner ads can reduce Web content
  credibility.
• Use reputable ads or no ads at all.
• Author photographs are more influential than
  author names.
• Make sure photographs dont hurt you.
• Future Online Studies
• Timeliness of content and credibility
• Tailoring and credibility