IL2007, Day 1: Putting Evidence-based Practice to Work
Posted on October 29, 2007
Filed Under Web 2.0, Web Design | Tags il2007
Putting Evidence-based Practice to Work, Frank Cervone and Amanda Hollister
Frank:
- most librarians haven’t been trained in HCI
- defined evidence-based practice
- data provides primary evidence for decision-making
- it’s not “common sense” - different stuff generally happens than what you “think” will happen
- Ex - doing a usability test, then comparing it with other similar tests to see the larger picture
- similar to user-centered design
- SPICE - setting, population, intervention, comparison, and evaluation
- Northwestern did their first usability test in 2001
- 2002 - did a catalog usability test - they found that the greatest number of searches that failed were title searches - title search was the default search setting, students were typing keywords into the default search box and not finding anything… so they found some great info from this test
- overall, site usability has improved - and they can prove it with statistical measures
- debates about how to proceed are easier - because they have data to fall back to
- easier to develop a strategy for incremental improvements over time - no longer locked into a tight academic schedule - they can prove the change will be an improvement, so have the go-ahead to roll the change out
- remaining issues - jargon and “i can find everything in google” problems
Amanda:
- spoke on making dynamic, page-based breadcrumbs on a website
- did a study of common paths customers took to get to certain pages
- they made something that constantly tells what paths customers are taking - very cool! They can narrow down to a single day if they want to
- future directions - implement predictive track analysis - find out where people are getting lost dynamically, then have something po up that says “were you really looking for this?”
il2007
Comments
One Response to “IL2007, Day 1: Putting Evidence-based Practice to Work”
Leave a Reply
RSS Feed
I wonder if they could develop a suggestion module that more accurately can guess the title of the book they were looking for?