Grant Blank & William H. Dutton Trust on the Internet: 2003- 2009 Summer Doctoral Program, 16 July 2010
Grant Blank & William H. Dutton
Trust on the Internet: 2003-2009 Summer Doctoral Program, 16 July 2010
Paradox of Trust
• Internet simplifies communication – e.g. reduces search costs
• But trust is harder– Lack of face-to-face cues– Greater opportunities for deception and fraud
Research Questions
• To what extent do people extend trust to the Internet?
• Has trust on the Internet changed over time?
Two 2003 Conclusions
• Internet is an experience technology
– “[t]hose exposed to the Internet gain more trust in the technology… Even past users have more confidence in the Internet than do non-users who have no experience” (Dutton & Shepherd 2006)
• Bad experiences reduced confidence and increased perception of risks
OxIS methodology
• Fieldwork: 2003, 2005, 2007 and 2009
• Probability sample of England, Scotland & Wales
• Respondents: 14 year olds and older
• Face-to-face interviews
• Sponsorship: Hefce, British Library, Ofcom, Scottish and Southern Energy, Cisco, Talisma, AOL, BT, Orange
Strengths of OxIS
• Quality– Random sample– High response rate– Face-to-face interviews improve reliability
• Comprehensiveness– 100s of questions cover dozens of topics of interest
• Flexibility– Ability to answer wide variety of research questions
• Consistency– Track trends and changes across years
2003 2005 2007 2009
Fielded in June-July
February-March
March - April
Feb - March
Number of respondents
2,030 2,185 2,350 2,013
Response rate
66% 72% 77% 67%
Sample composition OxIS
0 20 40 60Percentage
Broadband
Narrowband
Past user
Non-user
Note: 2003 N=2,029; 2009 N=1,929
Question: How reliable would you rate information on the Internet?Figure 1. Percent who Answered "Don't Know"
2003 2009
0 10 20 30Percentage
Accidently surf to porn site
Asked about banking details
Stolen credit card details
Misrepresented purchase
Received virus
Obscene, abusive e-mail
Note: 2003 N=1,045; 2009 N=1,296
Figure 2. Bad Experiences on the Internet
2003 2009
Inconsistency
• Fig. 2: Bad experiences common & increasing• Fig. 1: Users become more trusting
• 2003 finding: Bad experiences undermine trust
Two hypotheses
• 1. Bad experiences have no effect– Discounted– Accepted as cost of using Internet
• 2. Experienced users are able to cope with problems– The more time spent online, the more likely bad experiences– If the Internet is an experience technology, most the
experienced users are best able to cope
0 10 20 30 40percent of category
5 Total trust
4
3
2
1 No trust
Note: 2009 data N=1,296
Figure 3. Do Bad Experiences Reduce Trust?
No Bad Experiences Bad Experiences
Four categories of variables in models
• 2003 and 2009 – Experience on the Internet– Demographic variables as controls
• 2009 analysis adds variables– Attitudes toward the Internet and technology– Personal disposition to trust
Four new Indices
• Created using principal components analysis• “Net trust”
– Reflects confidence in the Internet and the people who use it
• “Technology Attitude”– Reflects general faith/suspicion in technology
• “Net Positive”– Reflects general positive attitude toward the Internet
• “Net Negative”– Reflects general negative attitude toward the Internet
• Attitudes are not deep-seated convictions
Table 2‘Net Trust’ Index Results for all Respondents
— 2009 coefficients —2003 Model 1 Model 2
Variable Coeff. Coeff. Coeff.Gender not sig. not sig. not sig.SES not sig. not sig. not sig.Age not sig. -* not sig.Education -* -* -*Broadband +* +* +*Narrowband +* +* not sig.Past user +* not sig. not sig.Trust propensity +*Tech. Attitude +*Net Positive +*Net Negative -*
N 2,026 1,804 1,681R² not reported 15.3% 26.1%
Table 4‘Net Trust’ Index Results for Internet Users
— 2009 coefficients —2003 Model 1 Model 1
Variable Coeff. Coeff. Coeff.Gender (female) not sig. not sig. not sig.SES not sig. not sig. not sig.Age not sig. not sig. not sig.Education -* -* -*Broadband not sig. not sig. not sig.Proximity -* +* not sig.Proximity*SES +* -* not sig.Bad experiences -* -* -*Trust propensity +*Tech. Attitude +*Net Positive +*Net Negative -*
N 1,045 1,214 1,207R² not reported 2.9% 17.5%
Table 62009 Commercial Activity Index
Model 1 Model 2Variable Coefficients CoefficientsGender (female) not sig. not sig.SES +* +*Age not sig. not sig.Education not sig. not sig.Income +* +*Broadband +* +*Proximity +* +*Bad experiences +* +*Net trust not sig. not sig.Net risk -* -*Trust propensity not sig.Positive technology +*Net Positive +*
N 958 952R² 32.7% 34.2%
Conclusions
• 1. Unlike 2003: Bad experiences do not undermine trust– Bad experiences are discounted or accepted as part of the
cost of taking advantage of the Internet
• 2. Experience variables dominate e-commerce model• 3. Attitude variables dominate trust models• 4. Demographic effects are weak in all models• 5. Users perceive valuable benefits: The Internet is
alive and well
Thank you.
Grant [email protected]
Oxford Internet Surveyshttp://www.oii.ox.ac.uk/microsites/oxis/[email protected]
Oxford Internet Institutehttp://www.oii.ox.ac.uk