Mediation - School of Communication and Information

Download Report

Transcript Mediation - School of Communication and Information

Mediation in
librarianship &
information retrieval
Reference interview
Human-human interaction
Question negotiation
© Tefko Saracevic, Rutgers University
1
Processes
 In
reference:
reference interview
– long standing concern
– a basic & major professional
skill of librarians
– mostly prescriptive, some
theories from communication
 In
information retrieval (IR):
question analysis
user-intermediary interaction
– connected with humancomputer interaction (HCI)
– also prescriptive, theories from
HCI & cognitive science
© Tefko Saracevic, Rutgers University
2
Reference interview
 Broader
context:
Interview and interviewing as
treated in a number of fields
– theories from communication
 interpersonal,
social interaction
– theories and practices from
sociology - among main
methods
– theories and practices from
journalism
– ethical concerns
© Tefko Saracevic, Rutgers University
3
Reference interview ...
 Dyadic
& (usually) face-toface or by phone
 Reference interview
characteristics:
– purposive by both participants
– restricted subject(s)
– relies on questions - answers
for diagnosis
– situation bound; social
– possible counseling aspects
– connected to informational
outcome - level, quantity, type ...
© Tefko Saracevic, Rutgers University
4
Elements to worry about
User
 Problem, task
 Inf. need
 Knowledge
 Intent
 Demographics
Librarian
 Comm. skills
 Knowledge
– subject
– inf. resources
Affective
 Intent

Library
• Inf. resources
• Situation, set-up
• Policies, rules
Results
• Effectiveness, validity, reliability
© Tefko Saracevic, Rutgers University
5
Diagnosis
 Taylor’s
classification of
information needs:
– Visceral - unexpressed
– Conscious - within mind
– Formalized - statement
– Compromised - as presented
 Types
of questions asked:
– Closed
 ‘yes
- no’; ‘this-that’ answers
– Open
 ‘tell
me more about project ...’
– Neutral
 assessing
situation, gaps, uses
© Tefko Saracevic, Rutgers University
6
Counseling, enabling
 Users
often do not have
– well defined problem
– well expressed or specific
question
– ideas what inf. or resources
exist or may be useful
– what to do next, as to
information or sources
 Counseling:
– help in definition, focusing
– advice on action
 Enabling:
– instruction on use, technology,
structure of resources ...
© Tefko Saracevic, Rutgers University
7
Studies of reference
 Process
– what goes on? Stages?
– reformulations, refinements
– still an art, not fully understood
 Types
of questions
– most asked are closed
 Communication
aspects
– verbal & non-verbal
– progress in discourse
– affective, attentive behavior
 Results
– validity, reliability, satisfaction
– statistics, costs
© Tefko Saracevic, Rutgers University
8
IR - problems addressed
Provide the users with effective
access to & interaction with
information resources.
1. How to organize inf.
intellectually?
2. How to specify search &
interaction intellectually?
3. What systems &
techniques to use for those
processes?
© Tefko Saracevic, Rutgers University
9
Mediation in IR
 Dyadic
model
Interface
User
 Triadic
Computer
model
User
(human interface)
© Tefko Saracevic, Rutgers University
Interface
Intermediary
Computer
10
Elements to worry about
 ‘Computer:’
– stands for a number of things
 hardware,
software
 inf. resources; meta information
 Interface:
– inf. to & from computer
– commands, display, navigation
 User:
– factors as in previous slide
– visualization
 Intermediary:
– acts as additional interface
– factors as in previous slide
© Tefko Saracevic, Rutgers University
11
Roles of intermediaries
Traditional
 mastering
– interfaces
– databases
– technology
 searching
for
users
 diagnosis
– question
analysis
 counseling
 packaging
&
delivery of
results
© Tefko Saracevic, Rutgers University
Evolving
(due to rise in
end-users)
 mastering
– networking
instruction
 assisting
 guiding
 enabling
 inf. resources

– selection etc.

system administration
12
Stages in user intermediary discourse
 Presearch
interview
– opening gambit; socialization
– modeling of user; db selection
– explanations by intermediary
 Online
search interaction
(if user present during searching)
– tactical maneuvering; changes
– terms, search tactics; db
– feedback; reiteration - dynamic
– explanations by both parties
 Closure
– closing downdrift
– focusing on output; evaluation
– delivery; advising - next steps
© Tefko Saracevic, Rutgers University
13
Discourse categories
 Context
 Terminology
 System
explanations
 Search
tactics
© Tefko Saracevic, Rutgers University
User problem, task
Request, inf. need
Expectations
Concepts & terms
Boundaries
Restrictions
How, what, when ..
Features, databases
Selection, variation
– terms, logic,
databases
Mistake correction
14
Discourse categories
(cont.)
 Review
&
relevance
 Actions
 Backchannels
Review, evaluation
– tactics, terms,
sources ...
Relevance judging
Feedback
Description of
activities
Explanations
Communication
prompts, fillers,
acknowledging ..
 Social/
extraneous
© Tefko Saracevic, Rutgers University
Social discourse
Formalities
15
Art of interviewing
purposive social interaction
Situational factors
– setting, physical environment
– rules, regulations, ethics
– appearance, demeanor
Communication skills
– semantics; language
– expression, delivery
– nonverbal communication
– turn taking
– encouragement; backchannels
Social factors
– establishing confidence
– rapport
© Tefko Saracevic, Rutgers University
16
Interviewing ...
Strategies
– progression in stages
– opening moves
 setting
an agenda, stage
– exploration, guidance
– maintaining focus. re-focusing
– feedback, re-iteration
– closure
Content, questions
– from categories in other slides
– role of explanation of choices
© Tefko Saracevic, Rutgers University
17
Examples from a study
 40
users; question each
 4 intermediaries; triadic HCI
 regular setting
 videotaped, logged
 48 hrs of tape (72 min. avrg)
– presearch: 16 min avrg.
– online: 56 min avrg.
 User
judgments: 6225 items
– 3565 relevant or part. relevant
– 2660 not relevant
 Many
variables, measures &
analyses
© Tefko Saracevic, Rutgers University
18
What do they talk about ?
(number of utterances)
Categories
Context
Intermediary
565
5%
User
Total
1102 11%
1667 8%
Terminology
1265 10%
904
9%
2169 10%
System expl.
1399 12%
625 6%
2024 10%
Tactics
3360 30%
1680 17%
5040 24%
Rev./Rel/
1996 18%
1825 19%
3821 18%
Action
1554 14%
399 4%
1953 9%
Backchan.
1179 10%
3179 33%
4358 21%
Total
11318 100%
9714 100%
21032 100%
% of Total
54%
46%
100%
© Tefko Saracevic, Rutgers University
19
Type of questions asked
Categories
User
Intermediary
Total
Terminology
Output/
evaluation
Search proc.
tech. asst.
266 31%
817 53%
1083 45%
67 8%
344 22%
411 17%
176 21%
32 2%
208 9%
Databases
143 17%
57 4%
200 8%
Context
0
172 11%
Social
79 9%
71 5%
150 6%
Backchan.
Other inf.
services
40 5%
50 3%
90 4%
28 3%
12 1%
40 2%
Action
32 4%
0
32 1%
Online issues
21 2%
2 .1%
23 1%
Total
852 100%
1557 100%
2409 100%
% of Total
35%
65%
100%
© Tefko Saracevic, Rutgers University
172
7%
20
Conclusions
 Interaction:
complex process
 Requires varied knowledge &
skills of intermediaries:
– communication, interviewing
– diagnosis, counseling
– inf. resources, meta inf.
– systems, networks
 Intermediaries
role changing
 In IR: terminological
imperative
– most talked & asked
 But:
GREAT FUN &
SATISFACTION
© Tefko Saracevic, Rutgers University
21
Thank you
Gracias
Danke
Merci
Hvala
© Tefko Saracevic, Rutgers University
22