The relevance of UX models and measures
Download as PPT, PDF2 likes643 views
1. The document discusses the growing number of proposed user experience (UX) components and measures and whether this could lead to a lack of standardization as seen with satisfaction measures. 2. It suggests that UX research could study and use simple UX measures like single rating scales combined with free-text explanations to collect user feedback, and analyze it using ad-hoc models focused on key identified aspects. 3. An example is given of collecting user ratings and comments on conceptual mobile phone designs to identify potential UX issues to address, such as privacy, reliability, and utility.
![A consequence of complex models and measures? [T]he measurement of satisfaction seems in a state of disarray […] The diversity of words used in five-point or seven point semantic differentials or Likert-type rating scales is simply astonishing […] Of the 112 studies that measure satisfaction […] only twelve studies employ standardized questionnaires.](https://image.slidesharecdn.com/i-uxsed2010folstadweb-101018074811-phpapp01/85/The-relevance-of-UX-models-and-measures-4-320.jpg)
The relevance of UX models and measures
- 1. The relevance of UX models and measures I-UxSED 2010 Asbjørn Følstad SINTEF Can simple rating scales and commenting mechanisms improve the relevance of UX measurement?
- 2. Some usability components of the early 90’s Efficiency Helpfulness Affect Control Learnability Satisfaction Memorability Error rate Effectiveness Usability component examples from J. Nielsen (Usbility engineering) and SUMI
- 3. Complex subjective usability measures Example: SUMI Software usability measurement inventory Questionnaire to be filled out by users Covers five usability components Efficiency Affect Helpfulness Control Learnability Respondent to answer 50 (!) items SUMI questionnaire: http://sumi.ucc.ie/en/
- 4. A consequence of complex models and measures? [T]he measurement of satisfaction seems in a state of disarray […] The diversity of words used in five-point or seven point semantic differentials or Likert-type rating scales is simply astonishing […] Of the 112 studies that measure satisfaction […] only twelve studies employ standardized questionnaires.
- 5. Could the same happen to UX models and measures? Flow Engagement Motivation Trust Beauty Goodness Hedonic quality Pragmatic quality Affect (The growing number of proposed UX components may suggest so)
- 6. ” UX-measures” in the real world Data on users’ experience collected through a rating scale and a free text field
- 7. ” UX-measures” in the real world Data on users’ experience collected through a rating scale and a free text field
- 8. ” UX-measures” in the real world Data on users’ experience collected through a rating scale and a free text field
- 9. Suggestion: Maybe HCI researchers could study and use ... Simple UX measures and ad-hoc models Rate (single scale) Users: Usability professional: Explain (free text) Analyze Ad-hoc model with key aspects
- 10. Case example: User feedback on concepts through a simple rating scale and a free text field Design brief: New concepts for textbased communication on mobile phones Concepts by J.O. Eikenes, L.M. Vedeler andT. Tveterås, the Oslo School of Architecture and Design
- 11. Case example: User feedback on concepts through a simple rating scale and a free text field Design brief: New concepts for textbased communication on mobile phones Concepts by J.O. Eikenes, L.M. Vedeler andT. Tveterås, the Oslo School of Architecture and Design MESSAGES AT LOCATIONS
- 12. Case example: User feedback on concepts through a simple rating scale and a free text field User rating and free text feedback indicated potential UX issues Ad-hoc model components in this particular case could include: Privacy Reliability Utility
- 13. Theoretical implications From the field of Marketing Bergkvist & Rossiter (2007) Multi-item measures de-facto standard in research Following Churchill (1979) Practitioners use single-item measures Empirical trial of multi-item vs. single-item measures inducate no difference in predictive validity … when measuring concrete and singular objects Attitude toward ad Attitide toward brand Could UX be seen as concrete and singular? (The spread of single-item measures in the real world may indicate so) And could ad-hoc UX model components be seen as the user’s rationalization of their UX? (Thereby indicating relevant issues for redesign)
- 14. So – what is your UX of this presentation? What do you like / dislike? Submit
Editor's Notes
- #3: Components from Niesens Usability Engineering and SUMI
- #5: Hornbæk2006: Another indication of the disarray is in the limited use of standardized questionnaires and the few studies that use measures that directly build upon earlier work. Of the 112 studies that measure satisfaction, 29 (26%) refer to previous work as a source for the questions. Among those, only twelve studies employ standardized questionnaires for measuring some kind of satisfaction.
- #6: Alle termer unntatt trust hentet fra IwC special issue on UX measurements and models Pragmatic quality, hedonic quality, beatury, goodness (editorial – Law/Shaik) Flow (Nacke et al) Engagement (O’Brien) Affect (mediating variable – Hassenzahl et al) Trust (ikke i special issue, men andre steder)
- #14: Bergkvist and Rossiter (2007) the predictive validity of multiple-item versus single-item measures of the same constructs. Journal of marketing research.