Collaborative Information Retrieval

1
Collaborative Information Retrieval

• Raya Fidel
• The Information School
• University of Washington

Supported by a grant from the National Science
Foundation.
2
CIR Research Team

• University of Washington
• Raya Fidel
• Harry Bruce
• Risoe Center for Human-Machine Interaction
• Annelise Mark Pejtersen
• Microsoft Research
• Susan Dumais
• Jonathan Grudin
• Boeing
• Steven Poltrock

3
What is CIR?

• Any activity that is taken by members of a
  work-team to collectively resolve an information
  problem.
  

4
Does CIR exist?

• Evidence in the following areas
• Among engineers in R D organizations
• In patient care in hospitals
• Education
• Problem No study has investigated the process of
  CIR
  

5
The Projects Goals Three Levels

• Empirical Level To discover and analyze some
  manifestations of CIR as they occur in the work
  place.
  
• Conceptual Level To extend an existing
  conceptual framework to address CIR.
  
• Technological Level To suggest technological
  and organizational developments that might
  enhance CIR.
  

6
Research MethodField studies of four teams that
include Sitting in meetings Interviewing team
members and others Observation Critical incident

• All verbal protocols are transcribed for analysis

7
The First TeamCreated to design the interface
for a new product

• Age 6 months
• Nine members, one of which is the manager

8
CIR Manifestation

• CIR takes place when one or more members of a
  team obtains information from outside the team
  that pertains to the teams work.
  
• Information sharing takes place when team members
  give and obtain information among themselves that
  pertains to the teams work.
  
• Information giving takes place when one or more
  members of the team gives information to someone
  outside the team.
  

9
The Conceptual FrameworkFramework for Cognitive
Work Analysis and Evaluation Jens Rasmussen and
Annelise Mark Pejtersen

• First Step Analysis
• Second Step Evaluation

10
Why this Framework?

• In-depth analysis of information behavior
  processes
  
• In-depth analysis of the context
• A bridge from behavior to design recommendations
  
  
• Comparisons among studies
• It has been used effectively in previous studies

11
Framework for Work Analysis
12
User CharacteristicsE.g., formal training, area
of expertise, tasks

• Education from high school to Ph.D.
• Areas such as visual communication, cognitive
  psychology, and kinesiology.
  
• Level of subject expertise between 3 and 20
  years.
  
• Experience in the company seven with less than a
  year, two with seven years.
  
• Tasks interface design, visual design, usability
  testing.
  

13
Task Situation Mental StrategiesE.g.,
preference about information sources, information
seeking style

• Participate in meetings
• Go to office of person who can give the
  information needed
  
• Create the information together
• Give information in order to receive information
  needed
  
• Browse Web pages of similar products
• Sign up to distribution lists on a variety of
  levels

14
Task Situation Decision MakingE.g.,
information need, information use, decisions made

• Most common information need design
  specifications
  
• Examples of decisions
• How to construct the menus for navigation
• How to develop metrics for usability
• What to name the product
• How to negotiate and coordinate with the other
  units involved in the design

15
Task Situation Work DomainE.g., purpose of
task, physical activities involved, priorities

• Examples of tasks involved
• Analysis
• Design
• Reporting

16
Work Domain Analysis StructureE.g., goals of
the team, priorities, work processes
17
Goals

• Building a standard framework to provide help and
  support to customers.
  
• Saving support organizations dollars.
• Improve the users help experience.
• Have users solve their own problems.
• Provide help experience that is interesting and
  useful to both consumer users and business users
  alike.
  
• Make help navigation among different content
  sources and different types as intuitive and
  efficient as possible.

18
Constraints

• The organizations culture (competition among
  units)
  
• Previous design (the amount of change that can be
  introduced with the new product)
  
• The shell that is used for the visual design
• The content to be included in the product
• Timetables
• Goals, constraints, priorities, functions,
  activities, and physical resources of the units
  with which they have to work
  
• Staff resources

19
Priorities

• Produce a useful product
• Finish on time
• Produce consistent design

20
Organizational Analysis Division of WorkE.g.,
how is the work divided among the team, criteria
used

• The manager decides about the goals,
  constraints, and priorities
  
• Anyone may decide about the functions, the
  physical activities, and the physical resources
  to be used
  
• Criteria for dividing the work within the team
• Subject expertise
• The type of information to which one has access
• What team members like to do
• Cognitive styles

21
Organizational Analysis Social
OrganizationE.g., communication among peers

• Hierarchy and negotiation
• Work is coordinated through
• Three regularly scheduled team meetings a week
• Ad hoc meetings
• Conversations with manager
• Conversations with other team members

22
Basic Concepts in Information Behavior

• Most concepts assume that individuals seek
  information by themselves and for themselves.
  
• How are these basic concepts relevant to CIR?

23
Information NeedA possible definitionInformat
ion needs arise whenever individuals find
themselves in a situation requiring knowledge to
deal with the situation as they see fit.

• Other definitions
• Anomalous State of Knowledge (ASK) Belkin
• A gap in knowledge Dervin
• The guiding idea here
• Information is for problem solving or decision
  making.
  

24
Types of Collaborative Information Needs

• Information about the content
• Design issues
• Administrative issues
• Users issues

25
Types of Collaborative Information Needs

• Information about the content
• How to evaluate the quality of the content?
• How is the content about servers broken down?
• Are updates limited to Microsoft products?

26
Types of Collaborative Information Needs

• Design issues
• How to notify users about changed content?
• How long the text in screen tips should be?
• On what level will the product be produced? CD?
  Web?
  
• How will production handle dynamic content?

27
Types of Collaborative Information Needs

• Administrative issues
• Who will help the team to figure out how the
  content is broken down?
  
• What does higher management think about the
  branding of the product?
  
• What does higher management view as the scope of
  the design team?
  

28
Types of Collaborative Information Needs

• Users issues
• What do customers expect in terms of service?

29
Stages of the Development of an Information Need
(Taylor)Visceral A sense of uneasiness

• Conscious Ill-defined area of indecision
• Formalized Describes area in concrete terms,
  making the need as explicit as possible
  
• Compromised Need as translated into the systems
  language
  

30
Recognizing a Collaborative Information Need

• During design work
• During planning for usability tests
• Raised externally by management
• Through a teams discussion about another issue
  (usually related to the need)
  
• During a discussion with other people (both team
  members and others)

31
Sharing the Understanding Collaboratively

• Through previous communications, face-to-face, or
  by email
  
• A team member, the manager, or a Product Manager
  initiates a discussion in a meeting
  
• Individuals meeting with the manager to discuss
• Individuals meeting among themselves to discuss

32
Collaborative Information Need

• Collaborative information need is most often
  formalized, and may even be compromised.
  
• The earlier stages of the development can take
  place within individual minds, but are rarely a
  collaborative state.

33
Aspects of Information BehaviorInformation
seeking How an individual goes about obtaining
information.

• Information use The outcome of information
  seeking.
  
• Information giving The act of disseminating
  messages.
  

34
Information SeekingHow an individual goes about
obtaining information. Types by level of
purpose
Searching Purposely looking for information to
resolve a particular information need.
Surfing Browsing through a source of informati
on, just to see what it has, without a particular
information need. Encountering Bumping into
information that can resolve a particular
information need when doing other things.
35
Collaborative Information Seeking

• The team we observed focused on searching only.
• We observed no cases of information
  encountering.
  

36
Information Seeking SearchingSearch
strategies

• The browsing strategy Intuitive scanning
  following leads by association without much
  planning ahead.
  
• The analytical strategy Explicit consideration
  of attributes of the information need and of the
  search system.
  
• The empirical strategy Based on previous
  experience, using rules and tactics that were
  successful in the past.
  

37
Information Seeking SearchingSearch
strategies (continued)

• The known site strategy Entering a URL to
  retrieve a particular site, or turning to a
  person who has the answer.
  
• The similarity strategy Find information based
  on a previous successful example that is similar
  to the current need.
  

38
Collaborative Information Searching

• Where did they get the information?
• From people outside the team, in a team meeting,
  or elsewhere (e.g., content providers, Product
  Manager, development programmer)
  
• Files shared by the team
• From their own collective knowledge through
  sharing information and speculating
  
• Library performed a search
• Internet

39
Collaborative Information Searching

• How did they know where to find information?
• Through discussion among themselves
• By knowing the organizational structure of
  Microsoft
  
• By the job specification (e.g., a Product Manager
  is supposed to know certain things)
  
• By the structure of the project (e.g., the
  designer knew which developer worked on the
  product)
  
• They asked someone who knew

40
Collaborative Information Searching

• What did they have to do to get the information?
• Discuss at a teams meeting where to get the
  information, and one person takes the
  responsibility to continue the process of finding
  the information
• Listen to a report in a meeting and ask
  questions
  
• Invite the person who has the information to a
  meeting
  
• Ask the library to find the information

41
Collaborative Information Searching

• Who did what?
• Product Manager and/or manager led a discussion
• A team member initiated a discussion
• Team members discussed what information was
  needed, and one member continued the search
  
• Team contributed own knowledge
• A team member asked for library search

42
Collaborative Information Searching

• How did they get the information?
• From personal network of co-workers and
  ex-co-workers
  
• By talking or emailing
• From Managers and members of other teams
• By talking or emailing
• By calling meetings with them
• By giving a written document and asking for
  feedback
  

43
Strategies for Collaborative Information
SearchingThe strategies that were useful were

• The analytical strategy When they encountered a
  new situation and worked together to figure out
  where to find the information, and whom to
  contact.
• The empirical strategy When they contributed
  their knowledge about how to find information in
  their discussions in meetings.
  
• The known site strategy When they contacted
  people whom they knew had the information for
  their problem.
  

44
Information Behavior

• Information seeking
• Information use
• Information giving

45
Information UseThe outcome of information
seeking May take various forms

• Acting on information
• Changing state of knowledge (making new sense)
• Confirm what one already knows

46
Collaborative Information UseActing on
information

• Received a particular spec and can use in design
• Found the person with whom to follow up

47
Collaborative Information UseChanging state of
knowledge

• Have a general idea of the big picture
• Understand the teams responsibilities
• Had an idea about how to involve other
  individuals
  
• Have an idea about how to involve management
• Found out that the information is not available
  yet
  
• A promise to work together later to obtain
  information
  
• Identifying issues that need to be addressed

48
Collaborative Information UseConfirm what one
already knows

• May take place in any of the previous examples

49
Information Giving

• Triggers for giving
• A user asks for information directly
• A user raises a topic about which the giver has
  information
  
• A user describes his/her situation to the giver
  who has information that can help
  
• A user behaves (or shows signs) in a certain way
  that prompts the giver to give information that
  will help
  

50
Collaborative Information Giving

• A person gives information in order to trigger
  the other person to provide needed information.
  
• This is a very common method to obtain
  information.
  

51
ConclusionsWhat do we know now?

• The first two stages of the need development
  (visceral and conscious) are performed
  individually.
  
• The last two stages (formalization and
  compromise) can then be carried out
  collaboratively.
  

52
Conclusions (continued)

• Only three search strategies are possible
• the analytical strategy
• the empirical strategy
• the known site strategy
• Information seeking is highly connected to
  information creation
  
• in meetings
• with documents
• Even if individuals carry out the process of
  information retrieval and creation, the team is
  highly involved in the preparation for that
  process.

53
If you want to follow up

• Please visit
• http//www.ischool.washington.edu/cir/