Are We Getting Warmer? Query Clarification in Virtual Reference (ALA Redux Part 2)
At the Library Research Round Table at ALA in Washington, DC, on June 23, 2007, Lynn Silipigni Connaway of OCLC and I presented our latest findings on Query Clarification in VR (aka question negotiation or the reference interview).
We examined 600 chat transcripts randomly selected from QuestionPoint bank of almost 500,000 transcripts. Here are some of our findings (and an invitation below to the Seeking Synchronicity web site to see the PowerPoint slides and handouts from this presentation).
Do librarians clarify?
75% (in 434 of 581 usable transcripts) librarians did ask clarifying questions.
Did they ask the highly recommended follow-up question? (some version of “Does this completely answer your question?”)
50% (217) of the 434 librarians who clarified did ask the follow up question.
What types of questions were asked?
66% (554 of 838 questions asked by the librarians) were closed questions.
34% (282 of 838) were open.
What did librarians ask about?
Librarians asked users questions about: topic, background, search history, type of resource needed, extent/depth of information needed, if the user wanted a referral and more.
How about the virtual reference users?
Users offered information about: topic, background, extent/depth, and to correct the librarian’s misunderstanding.
Surprising finding! 2 different patterns of clarification!
Librarians clarified more often in the beginning of the interaction
Users clarified in the middle more often.
Most important finding! How to improve accuracy in chat reference?
For the 180 ready reference questions in our sample, we looked at accuracy (see my blog posting of July 10, 2007 for more on ready reference in chat).
Clarifying the query and asking the recommended follow-up question both boosted accuracy.
Bottom Line
Always ask clarifying questions, even if you think you understand the question (one user asked for diving instructions, but had made a typo and wanted driving instructions, early clarification would have saved the librarian much searching time!)
Always ask a version of the recommended follow-up question: “Does this completely answer your question?”
Interested in more detail on the above findings? Please click on the above links to see the PowerPoint slides and handouts.
We examined 600 chat transcripts randomly selected from QuestionPoint bank of almost 500,000 transcripts. Here are some of our findings (and an invitation below to the Seeking Synchronicity web site to see the PowerPoint slides and handouts from this presentation).
Do librarians clarify?
75% (in 434 of 581 usable transcripts) librarians did ask clarifying questions.
Did they ask the highly recommended follow-up question? (some version of “Does this completely answer your question?”)
50% (217) of the 434 librarians who clarified did ask the follow up question.
What types of questions were asked?
66% (554 of 838 questions asked by the librarians) were closed questions.
34% (282 of 838) were open.
What did librarians ask about?
Librarians asked users questions about: topic, background, search history, type of resource needed, extent/depth of information needed, if the user wanted a referral and more.
How about the virtual reference users?
Users offered information about: topic, background, extent/depth, and to correct the librarian’s misunderstanding.
Surprising finding! 2 different patterns of clarification!
Librarians clarified more often in the beginning of the interaction
Users clarified in the middle more often.
Most important finding! How to improve accuracy in chat reference?
For the 180 ready reference questions in our sample, we looked at accuracy (see my blog posting of July 10, 2007 for more on ready reference in chat).
Clarifying the query and asking the recommended follow-up question both boosted accuracy.
Bottom Line
Always ask clarifying questions, even if you think you understand the question (one user asked for diving instructions, but had made a typo and wanted driving instructions, early clarification would have saved the librarian much searching time!)
Always ask a version of the recommended follow-up question: “Does this completely answer your question?”
Interested in more detail on the above findings? Please click on the above links to see the PowerPoint slides and handouts.
Special thanks to Susanna Sabolcsi-Boros of Rutgers, SCILS, and Patrick Confer of OCLC, for their help with this part of the project.
Labels: accuracy, chat reference, Reference, virtual reference
posted by Marie L. Radford @ 7:52 AM
0 comments
0 Comments:
Post a Comment
<< Home