350 likes | 453 Vues
Study comparing end-user perceptions of hi-fi and lo-fi web site prototypes, revealing differences in design attributes and usability problems. No correlation found between visuals and prototype fidelity.
E N D
End-User Perceptions of Formal and Informal Representations of Web Sites Jason HongFrancis LiJames LinJames Landay Group for User Interface ResearchUC Berkeley CHI 2001 April 5, 2001
Design Representations • Differing levels of detail High-fidelity (Hi-fi) Low-fidelity (Lo-fi)
Static Images • Hi-fi representations • encourage discussion of details • fonts • colors • graphics • pixel-level layout • Lo-fi representations • encourage discussion of high-level issues • information architecture • navigation flow • overall page layout
Interactive Prototypes • Hi-fi prototypes • use hi-fi representations • Lo-fi prototypes • use lo-firepresentations • often on paper • faster to create than hi-fiprototypes
End-User Perceptions • How would end-users judge a web site design, depending on the fidelity of the site’s prototype?
Hypotheses • Between hi-fi and lo-fi prototypes • people will find many of the same usability problems
Hypotheses (cont’d) • Between hi-fi and lo-fi prototypes • people will judge the following attributes differently: • professional • functional • finished • likely to change
Hypotheses (cont’d) • Between hi-fi and lo-fi prototypes • people will rank the importance of visual vs. non-visual design issues differently • visual issues: font, color, graphics • non-visual issues: general location of navigation bars, overall page layout, organization of information
Experiment Design • Each participant evaluated one site with hi-fi, other site with lo-fi Between-subjects
Experiment Design • For hi-fi prototypes • downloaded two actual web sites • pared down product list • greeked some text out • hardwired shopping cart data
Experiment Design • For lo-fi prototypes • made electronic representations in DENIM instead of using paper • exported DENIM pages to GIF + image maps • touched up sketches to improve handwriting • added HTML text fields by hand
Experiment Design • Participants asked to do 5 tasks per web site • find a specific piece of information • find another specific piece of information • find information deeper in site • add two particular items to shopping cart • purchase items in shopping cart
Experiment Design • After each web site, participants: • gave overall impressions of site design(scale: 0–10) • professional, finished, functional, likely to change • ranked ten possible ways of improving web site • 6 were non-visual improvements • 4 were visual improvements • made comments, suggested improvements to web site
between hi-fi sites attractiveness quality security likelihood of changing between lo-fi sites ease of use efficiency complexity attractiveness usefulness Are the two web sites the same? • Significant differences found(p < .05, 2 tailed independent t-test) • Evaluated each site separately • between-subjects analysis
Usability Problems • Many of the same usability problems were found in both hi-fi and lo-fi • e.g., Add to Shopping Cart / Registration problem • Lo-fi prototypes • sufficient for task-based usability evaluation • faster and easier to create
Perceptions of Design Attributes • People did judge attributes differently depending on fidelity • Lo-fi representations look: • less professional • less functional • less finished • more likely to change
Rankings of Visual and Non-visual Issues • Importance of visual vs. non-visual issues did not correlate with fidelity of prototype • electronic medium may raise expectations • quality of sketches may affect perception
Future Work • Study earlier stages of design cycle • Task-based vs. non-task-based evaluation • Investigate effect of medium on perception • paper sketches • scanned paper sketches • electronic sketches • printed electronic sketches
Conclusions • Many of the same usability problems found in hi-fi and lo-fi representations • Lo-fi representations look: • less professional • less functional • less finished • more likely to change • No correlation between importance of visual vs. non-visual improvements and fidelity of representation
For More Information • http://guir.berkeley.edu/pubs/chi2001/formal-informal.pdf
Experiment Design • After each task, participants were asked how well site supported task (scale: 0–10) • easy • efficient • clear • helpful
Non-Visual Visual
Non-Visual Visual
7.25 Non-Visual Visual 2.875