More Related Content
Viewers also liked
Viewers also liked (20)
Similar to Trust in IT: Factors, Metrics and Models
Similar to Trust in IT: Factors, Metrics and Models (20)
Recently uploaded
Recently uploaded (20)
Trust in IT: Factors, Metrics and Models
- 1. + Trust in IT: Factors, Metrics and Models Dr. Clare Hooper clare@clarehooper.net www.twitter.com/clarejhooper
- 2. + Hi! I’m Dr. Clare Hooper I enjoy exploring boundaries quantitative and qualitative industry and academia I’m into HCI, web science, user experience, ubicomp, hypertext Recent topics include design, inclusive research, crime, social media and trust
- 3. + TRIFoRM TRust in IT: Factors, metRics, Models Trust Mistrust
- 4. + Multidisciplinary approach Computer science, health science, social science and engineering The University of Southampton… Faculty of Health Sciences (Jane Prichard) Faculty of Business and Law (Melanie Ashleigh) IT Innovation Centre (Clare Hooper, Brian Pickering, Mike Surridge, Stefanie Wiegand)
- 5. + The problem What are the trusting beliefs of users of IT systems? What factors influence trust of systems? How can we model those factors and trust levels?
- 6. + Trust in a system may be too high or low compared with its actual trustworthiness (security) Trigger adequate trust perceptions allowing users to make risk-aware, informed decisions Methodologies to increase trustworthiness www.it- innovation.soton.ac.uk/projects/opte t
- 7. + TRIFoRM How do users develop and apply trust to systems? Models to determine when user trust is out of balance with system trustworthiness Add to set of warnings and reassurances that OPTET can provide
- 8. + The process Literature • State of the art from social sciences • Analyse trust models from OPTET Data gathering • Semi-structured interviews • Service users and a service provider Thematic analysis • Identify and model threats • Identify threat controls and key issues
- 9. + The process Literature • State of the art from social sciences • Analyse trust models from OPTET Data gathering • Semi-structured interviews • Service users and a service provider Thematic analysis • Identify and model threats • Identify threat controls and key issues
- 10. + The process Literature • State of the art from social sciences • Analyse trust models from OPTET Data gathering • Semi-structured interviews • Service users and a service provider Thematic analysis • Identify and model threats • Identify threat controls and key issues
- 11. + Model of trust
- 12. + Rheumatoid Arthritis App (RApp)
- 13. Technology Acceptance (3+1) Demographics HCI Trust Transfer Cognitive Impairment +ve Effects on H2H Interaction Fault Tolerance (2+1) Usefulness (4+1) Ease of Use (5+1) Risk Taking (4+1) Convenience (4+0)
- 14. Technology Acceptance (3+1) Demographics HCI Trust Transfer Cognitive Impairment +ve Effects on H2H Interaction Fault Tolerance (2+1) Usefulness (4+1) Ease of Use (5+1) Risk Taking (4+1) Convenience (4+0)
- 15. Technology Acceptance (3+1) Demographics HCI Trust Transfer Cognitive Impairment +ve Effects on H2H Interaction Fault Tolerance (2+1) Usefulness (4+1) Ease of Use (5+1) Risk Taking (4+1) Convenience (4+0)
- 16. Technology Acceptance (3+1) Demographics HCI Trust Transfer Cognitive Impairment +ve Effects on H2H Interaction Fault Tolerance (2+1) Usefulness (4+1) Ease of Use (5+1) Risk Taking (4+1) Convenience (4+0)
- 17. Technology Acceptance (3+1) Demographics HCI Trust Transfer Cognitive Impairment +ve Effects on H2H Interaction Fault Tolerance (2+1) Usefulness (4+1) Ease of Use (5+1) Risk Taking (4+1) Convenience (4+0)
- 18. Technology Acceptance (3+1) Demographics HCI Trust Transfer Cognitive Impairment +ve Effects on H2H Interaction Fault Tolerance (2+1) Usefulness (4+1) Ease of Use (5+1) Risk Taking (4+1) Convenience (4+0)
- 19. Technology Acceptance (3+1) Demographics HCI Trust Transfer Cognitive Impairment +ve Effects on H2H Interaction Fault Tolerance (2+1) Usefulness (4+1) Ease of Use (5+1) Risk Taking (4+1) Convenience (4+0)
- 20. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 21. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 22. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 23. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 24. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 25. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 26. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 27. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 28. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 29. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 30. Trust Transfer (4+1) Trust in Experts (2+1) +ve Effects on H2H Interaction (5+1) Age (1+0) Physical Health (2+0) -ve Effects on H2H Interaction (4+1) Experience (4+0) Cognitive Impairment (2+0) Trust in Process (4+1) Demo- graphics (4+1)
- 31. + Threat 1: User Disengagement
- 32. + Threat 2: Unusable System
- 33. + Monitoring technology to manage pain: issues Likelihood of user risk-taking and fault-tolerance: users are more vulnerable. We must be cognisant of user motivation and application domain throughout design Relationships are important. Interviewees were concerned that such technology might change their relationship with healthcare providers be trusted by their healthcare providers Users Carers Systems Stake- holders
- 35. + Summary: outputs and findings Usefulness and ease of use are key – of course! Also: demographics, motivation, domain Users managing pain are mistake-tolerant and more likely to take risks Trust transfer and H2H interactions affect H2T trust Technology should enhance, not replace, H2H interaction The right amount of data Displayed in the right way, at the right time OPTET augmented its model Other projects used TRIFoRM outputs Work continued by INTRMS: Individual and Network Trust in Remote Monitoring Systems Trust in the healthcare environment is a complex interaction between technology utility and convenience and its affect on existing human interaction
- 36. + Thank you I’m new to Vancouver; please introduce yourself I’m available for consultancy HCI, web science, user experience, ubicomp, hypertext Design, inclusive research, crime, social media and trust Stay in touch clare@clarehooper.net twitter.com/ClareJHooper
Editor's Notes
- I don’t know many of you folks, so let me take a moment to introduce myself. I was based in Europe ‘til now.
- (Mis)trust of IT utilities has enormous impact on people’s decisions to use or not use those utilities. We set out to identify and model factors that influence individual trust of IT systems. I’m going to take you through the process we used and our results, particularly insights from a thematic analysis.
- Healthcare technologies for monitoring chronic conditions, and interviewed people who may use or provide healthcare monitoring technology to understand what was important to them as individuals.
- Help users trust safe systems Protect users from risk with compromised systems OPTET infers trust with a model: trust change as a result of UX. Supports IT system engineering goals (its for technical system designers & devs), ignores what users feel about systems, or how their feelings affect their behaviour.
- TRIFoRM addresses the gap. A more holistic model of trust, not just “trust as a response to system behaviour”
- SotA: >70 pubs on defining/modelling trust, theoretical foundations of trust in tech, technology acceptance, formation of trust in tech.
- SotA guided semi-structured interviews. 7 interviewees (1 withdrew) connected w/healthcare monitoring tech. What’s important to them as individuals? 6 service users 1 service provider, who treats pain and designs/develops assistive tech.
- Professional transcription, 2 coders analysed for themes, 1 person analysed the themes 2 thematic maps: identified new OPTET threats, & controls. 4 major themes…
- Model of trust formation, 1st and 2nd order constructs. Performance: does it help user achieve goal? Accuracy, reliability, functionality Process: how much does it do do what is expected? Understandable, predictable, confidential, data integrity Purpose: the intentions of the provider: benevolence, authorised data usage The model of trust specifically for IT artifacts from Söllner, Hoffmann, Hoffmann, Wacker & Leimeister, 2012
- Smarpthone app for monitoring RA Objective inputs: monitoring physical movements Subjective inputs: patient reports of pain levels
- Thematic map 1: HCI & technology acceptance (willingness to trust and adopt tech; impact of tech on willingness to do that). Figures in brackets show how many service users mentioned each topic, and whether the service provider did. The arrows link themes with sub-themes and show some assumed relationships between themes or sub-themes associated with a different theme.
- Not so much trust as necessity, “I am a bit more nervous than I use to be but the alternative, which is not to use any of these things, is not an option for me”. The medical context helps, though: “Quite honestly if medical people can’t be trusted who can.” (Quotes from users unless I say from SP, service provider.)
- Fault tolerance: users typically accepted that problems will occur; that doesn’t necessarily undermine trust. “We might come unstuck somewhere along the line but thus far I trust it […] and I don’t see any reason for that changing.”
- Usefulness: users had an appreciation that tech can make things easier, from aide-memoir (“I had this app called Pill Scheduler and there would be alarms and everything and it would tell me what to take”) to even improving treatment (“It provides very new data, in that sense, outside the traditional clinical encounter” –SP).
- Physical aspects (“the iPad is a lot easier when I’m stressed becaues it’s flat and I don’t actually have to bend the wrists in the same way”), responsveness, allowing achievement of complicated tasks, being intuitive and easy to navigate. Technology format, given RApp context: “So I’m wondering how, unless you go on to a smart watch or something and wear it on your wrist, I don’t know, how accurate or how misleading that could be depending on the advanced-ness, as it were, of what it was recording or how it was recording rather.”
- Tech must fit in with everyday life (“Yeah, I really like the iPad I find it easy to use and it doesn’t take up any space. You can take it with you.”) and potentially help not just the individual but others (“bothering receptionists and having to explain who you are, what you want, blah blah […] So it’s just great. I don’t hassle anyone, nobody has to hassle me and it works smoothly”)
- In summary: if motivated and appropriately positioned, technology supporting healthcare is likely to be accepted with maybe initial resistance. It’s really about usefulness.
- Thematic map 2: trust transfer and demographics. How trust in overall healthcare management process & individuals supporting that process might affect trust and acceptance of tech. How culture and characteristics of users may affect that. (Trust in one agent or component can logically transfer to others).
- There’s implicit and explicit acceptance that medical team has best interests of user at heart (“Quite honestly if medical people can’t be trusted then who can? […] I haven’t really thought about it because I always assumed they were tickety boo. And I’m sure they are”)
- Users showed willingness to trust tech as long as they take sensible precautions, responsibility (“I mean, I change my password a lot so they’d have to be pretty clever to get into my stuff. […] I think it’s just common sense really.”). SP sees role as providing tech for support (“this care emerges from interaction between people, clinicians and patients being the central part of that care relationship, where technology adds to, enhances, supports, exemplifies that activity, I think it absolutely will be very, very important to how medicine and the clinical world develop. Where it’s used to substitute it, I think that can be detrimental.”)
- Users see the special nature of interactions between healthcare professionals and those they treat (“the nurses can if you’re not very well by your demeanour and the way you look of course.”). Technology can fill information gaps (“I’m not the best historian really when I go. I’m not the best at describing pain.”). But: “I think it’s imperative that the two-way communication is maintained. I think the technology, not the technology but the data should support the meeting not overtake or subsume it.”
- Importance of companionship and sympathy (“I went for an infusion a few weeks back and the lady came and sat down next to me and when her nurse came up and said how are you she burst into tears because she was so ill. […] Being with somebody else, both professional and another patient, is fantastic. I mean, that’s a great example of how important that is.” Must maintian that conversation, to avoid mechanistic and dehumanised healthcare (“It kind of takes away from the personal aspect of things. It’s all a little robotic, isn’t it? They forget to talk to you because they’re too busy looking at computers instead of you.”). Ensure data is used: “I’d be quite happy to fill something in but there’s no point in doing that unless they were going to look at it.”
- A small pilot. Some users were comfortable with technology (“Despite my age I’m actually quite a geek really. […] I have run a website for a whole group of [people[. I have a very positive attitude towards technology”). The service provider had a negative comment: “We’ve found all sorts of cultural problems in trying to get both clinicians and students to see the worth of these things and to use the technology on a day to day basis.”).
- Existing usage or experience with technology is relevant, i.e. patients using lots of apps or particular computer products (“Yes, I use it for, I mean I’m all Apple based too. I have an iMac and I have an iPad and I have an iPhone”).
- Age and condition may reduce the importance of perceived risk in handling sensitive information: “Well, as far as I’m concerned I’m at retirement age […] I haven’t got any life insurance, it’s not particularly sensitive to me. I don’t really mind who knows, if you see what I mean. ” Stereotypes can mislead: “I’m actually quite a geek really. I do digital artwork as well.”
- “The iPad is a lot easier when I’m very stressed because it’s flat and I don’t actually have to bend the wrists in the same way.”
- “That’s the other thing with rheumatoid, if you’re feeling really tired it’s really easy to get brain fog and do something really stupid.”
- In summary: users trust the whole process & medical team. We must account not just for ease of use and usefulness but background, expectation of potential users. Their condition may lead to physical or cognitive disadvantage that could exacerbate technology issues.
- From the interviews we identified 2 threats not covered by OPTET. Clinician disengages, undermining trust of pain sufferers User Disengagement – if the medical team don’t adopt the technology, then patients will lose interest in that technology. Control: User Training
- An Unusable System causes all users, clinicians as well as patients, to lose interest and disengage. (Possible cause of threat 1) User Disengagement can be mitigated with appropriate training and support; Unusable System can be helped with appropriate design expertise. Control: HCI Analysis (for the Unusable System threat)
- Design, implement, test, maintain systems cognisant of these issues We can speculative that healthcare provider perception of patient trust is also important. Not to mention other stakeholders: family, sysadmins, other types of healthcare provider…
- Interdisciplinary SOTA guided semi-structured interviews w/service users and a service provider Thematic analysis and maps (tech acceptance and trust transfer) Threats and controls (user disengagement; unusable system)
- Human & organisational relationships. Concern that tech may change relationships: nurses focus on data, lose contact w/patients; consultants ignore data, patients don’t trust/use system. Trust depends somewhat on non-system interactions, particularly if it changes expectations or behaviour. Tech must enhance and develop H2H. INTRMS: identify further factors to influence trust, implications for design, evaluate with a human-machine network (Rheumatoid Arthritis App, an app for objective and subjective monitoring of rheumatoid arthritis). eVACUATE: user responses to technology, both evacuees and operational staff HUMANE: how trust relations might affect human-machine networks PRICE: how trust relations might affect social machines in general and TimeBanks in particular ProSocialLearn: how trust relations may translate to co-operation in the virtual world.