Transcript Usability & sociability in online communities

Usability & sociability in online
communities: A framework for research &
practice
Jenny Preece
Prof. & Chair of Information Systems
UMBC, Baltimore, MD 21250, USA
[email protected]
www.ifsm.umbc.edu/onlinecommunities
Overview
• Definitions
• Sociability & usability
• Research example
• Conclusions & future research
Definitions of online community
• Technologists
• Sociologists and anthropologists
• Business entrepreneurs (e-commerce)
• Any virtual space where people come together to get or
give information or support, to learn, to discuss, to be
with others online.
• Online communities support communication between
patients, professionals, students, citizens and nations
• Small or large, local, national, or international, virtual
or physi-virtual.
My definition (Preece, 2000)
• People –make the community. Group dynamics,
needs and roles shape the community.
• Purposes – people come together for a
purpose(s).
• Policies – behavior is governed by group norms,
rules and sometimes formal policies.
Software – supports and influences community
activity.
Some numbers (10/2001)
•
•
•
•
•
•
•
52m US Internet users, 55% check health sites
230m unique MSN users per month
29m AOL users, 1 million more per month
Over 104m ICQ users, millions now ‘texting’
Over 91,500 UseNet groups
50,000 IBM employees, World Jam, June ‘01
100 -150 immersive CAVE environments
Overview
• Definitions
• Sociability & usability
• Research example
• Conclusions & future research
Sociability and Usability
• Sociability is concerned with social interaction.
Communities with good sociability have
unambiguous, supportive, social structures.
• Usability is concerned with human-computer
interaction. Systems with good usability are
consistent, controllable and predictable.
Sociability
• Purpose – provide a clear statement of purpose,
brand name, symbol
• People – support different types of participants
and participation, show presence when
appropriate, keep participants interested
• Policies – guide behavior by providing and
encouraging conventions, moderate with
policies, support trust and security
Usability
 Dialog & social interaction support –provide
support for communication – icons, reduce
typing, visualizations
 Information design – distinguish between new
& old content, different types of content
 Navigation – support moving around the
community, searching messages, moving
between modules
 Access – consider speed of connection, not
everyone has most recent technology
Pillars of participatory communitycentered development
Sociability
 Purpose
 People
 Policies
Usability
 Dialog & social
interaction support
 Information design
 Navigation
 Access
Support sociability, design usability
• Should there be a registration policy?
- Who can join?
- What effect will it have on membership?
• Write message, design form
- Interaction design
- Layout - e.g. position & size of boxes etc.
- Relationship with database
Community Framework
People
Communication:
2
Community
Purpose
i.e. Type of activity
How much
By whom
Satisfaction
i.e. goals
Policies
KEY
(Lewin, 1930s)
i.e. Authoritarian
Democratic
Laissez-faire
Anarchic
1-3 scaffolds suggested
Signals termination
Many CSCW issues
(Bales, 1950s)
i.e. Informational
Social-emotional
Operations:
3
Sociability
i.e. On-topic
Functions
i.e. roles
Identity
1
(McGrath, 1984)
i.e. Generate
Choose
Negotiate
Execute
Norms & rules
Policies
S
c
a
f
f
o
l
d
S
Community
Reciprocity
Empathy
Trust
Identifiability
Com. ground
Privacy
i.e. Type
Stage
Size
Culture
Infrastructure
3 type
i.e. Media
Network capacity
Computer capacity
Usability:
Individual
Software
i.e. Navigation
Community
Information
Usability:
Community
i.e. Conviviality
Efficiency
Effectiveness
Scaffolds
Examples
1 People - roles
Visibility: individuals, groups,
community
Search: people with certain
characteristics.
Tools to support different roles.
Babble social translucent
(Erickson et al.) Donath
(2002) flower gardens.
Pictures, caricatures, icons,
web pages to support identity.
2 Purposes –
communication –
Informational,
social-emotional
Meaningful name & description
Identify: topics, expertise,
communication type, who speaks
& to whom.
Support: dictionary, thesaurus,
translation, etiquette, FAQs,
common ground, empathy &
trust support, to reduce typing.
Sack social network diagrams
(2000). Phrases to support
common ground (Zimmer).
Palette of communicative
symbols. Tools and spaces for
conflict resolution.
3 Policies –
authoritarian,
democratic, laissezfaire, anarchic –
norm/rules, policies
Make explicit: norms & rules
Support: facilitation, moderating
Decision making: discussion
support, voting
Scaling: large groups, private
discussion
Some large systems such as
Delphi, voting software (e.g.
id-book.com) and governance
(e.g. in Diversity University).
Tools for moderators.
Overview
• Definitions
• Sociability & usability
• Research example
• Conclusions & future research
Research: Silent participants or lurkers
(Blair Nonnecke, 2000)
12 indepth interviews - Reasons for not posting
•
•
•
•
•
•
•
•
Uncomfortable in public
Learning about the group
Building identify
Fear of persistent messages
Information overload
Not necessary to post
Personal characteristics (e.g., shyness)
Group influences
Lurkers often feel part of a community
From a lurker ...
“Maybe it's a sign of my own mild discomfort
around being a lurker, but I found it reassuring
to recognize myself and my behavior within the
continuum you describe, and to see lurking
treated seriously, with both acceptance and
respect. As a lurker, I'm used to observing from
the sidelines and participating vicariously, and
it's strangely gratifying to read an article that
speaks directly to that experience. It's almost
like suddenly feeling part of an (until-now)
invisible community of lurkers.”
Lurking online – data logging
•
•
•
•
•
12 weeks
Started with 135 original subscriptions
Ended with 109 DLs
Health 77, software 21
147,946 messages were transcribed into records
and imported into an SQL database.
• 60,000 members
• 19,000 posters.
(Nonnecke, 2000; Nonnecke & Preece, Chi’2000)
Lurking in 77 health and
21 software support lists
Lurking (% of membership)
100
80
60
40
20
0
health
software
DL type
Variation of lurking levels for
cumulative posts over 3 months
Mean lurkers (% of membership)
100
90
80
70
60
50
software
DLs
health
40
DLs
30
0
1
2
3
Posting levels (cumulative posts in 12 weeks)
% lurking in health & software groups
Software
Health
Low lurking when:
- lists are small
- traffic is high
- messages are short
- few single posters
- ‘stars’ are present
All
90
80
70
60
50
40
30
20
10
0
(Nonnecke, 2000)
(Nonnecke & Preece, 2000)
Question
Result from logging study
P3
How many lurkers are there?
Fewer than expected: still high with an
average of over 55% for all DLs (when
defined as 0 posts in 3 months).
R3a
Does lurking in health and software-support DLs
differ?
Yes: health-support groups have lower
levels of lurking (45% vs. 82%).
R3b
If lurking is defined as no posting, what happens
to lurking levels when the definition is broadened
to include minimal levels of posting, e.g., 1
post/month?
Lurking increases rapidly and then levels
off as definition is broadened. Healthsupport groups maintain their lower levels
of lurking (75% vs. 97% for software
when lurking is defined as 3 or fewer
posts/3 months).
R3c
Is there a relationship between lurking and the
number of members in the DL?
Yes: smaller DLs have fewer lurkers.
R3d
Is there a relationship between lurking and the
traffic level of the DL?
Yes: higher traffic means lower lurking.
R3e
If posting is concentrated with a few posters, how
does that affect lurking levels?
The greater the concentration, the less the
lurking.
R3f
Are short messages related to lower levels of
lurking?
Yes: short messages are related to lower
levels of lurking.
R3g
If clumpiness is an indication of interaction, does
it necessarily follow that increased clump size is
related to decreased lurking?
Yes: larger clumps are related to lower
levels of lurking.
R3h
Is there a relationship between the number of
singleton posters and level of lurking?
Yes: as the number of singleton posters
rises (and those who do not receive a
response), so does the lurking.
Table 6.2: Overview of results ordered by question (From Blair Nonnecke’s thesis, 2000, SBU London)
Social presence in Babble
(Erickson et al., Chi’99)
Criteria for success
Sociability
No. participants
No. messages
Reciprocity
On-topic discussion
Empathy
Trust
Social satisfaction
Lurking
Uncivil behavior
Usability
Speed of learning
Productivity
User satisfaction
Retention
Errors
Overview
• Definitions
• Sociability & usability
• Research example
• Conclusions & future research
Research
• Community dynamics and the role of an online patient
support community in everyday life (Diane Maloney-Krichmar)
• Lurking and participation in 1000 online communities
(Dorine Andrews, Blair Nonnecke, Greg Morton)
• Communicating trust using mobile devices – empathy &
predicability (Heidi Feng, Jonathan Lazar)
• What makes online communities successful? Evaluation
heuristics and metrics (Chadia Abras)
• Framework for online community development (Clarisse S. de
Souza)
• Supporting lightweight communication in health support
communities (Clarisse S. de Souza)
‘We shape our buildings,
and afterwards our
buildings shape us’
Winston Churchill
‘My experience of the
world is that things left to
themselves don’t get right’
T. H. Huxley
Web sites
www.ifsm.umbc.edu/onlinecommunities
“Online Communities: Desinging
Usability, supporting sociability”(2000)
Jenny Preece, John Wiley &Sons
www.id-book.com
“Interaction Design: Beyond HCI”(2002)
J. Preece, Y. Rogers, H. Sharp, John
Wiley & Sons
www.ifsm.umbc.edu/~preece
www.ifsm.umbc.edu/onlinecommunities
Id-book.com
Publications
•
•
•
•
•
•
•
•
•
Andrews, D. & Preece, J. (2001) A conceptual framework for demographic groups resistant to
online community interaction. Proc. HICSS-34 IEEE Computer Society, Maui, Hawaii.
Preece, J. & Ghozati, K. (2000) Experiencing empathy online. In R. Rice & J. Katz, The
Internet and Health communication: experience and expectations. Thousand Oaks: Sage
Nonnecke, B. & Preece, J. (2000) Counting the silent. ACM CHI’2000, Hague, 73-80.
Brown, J., van Dam, A., Earnshaw, R., Encarnacao, J., Guedj, R., Preece, J., Shneiderman, B.
& Vince, J. (1999). Human-centered computing, online communities and virtual
environments. ACM Interactions, 6 (5).
Lazar, J., Tsoa, R., & Preece, J. (1999). One foot in cyberspace and the other on the ground: A
case study of analysis and design issues in a hybrid virtual and physical community. WebNet
Journal: Internet Technologies, Applications and Issues, 1(3), 49-57.
Nonnecke, B., & Preece., J. (2000). Persistence and lurkers: A pilot study. Proc. HICSS-33
IEEE Computer Society, Maui, Hawaii.
Preece, J. (1998). Empathic communities: Reaching out across the Web. ACM Interactions 5
(2), 32-43.
Preece, J. (1999). Empathic communities: Balancing emotional and factual communication.
Interacting with Computers, 12, 63-77.
Preece, J., & Ghozati, K. (1998). In search of empathy online: A review of 100 online
communities. Proc. 1998 Association for Information Systems, Americas Conference,
Baltimore, USA.
Additional material if time
Community Framework – Sociability
Community type
Stage
Size
Culture
i.e. local, national
Sociability:
i.e. On-topic
Reciprocity
Empathy
Trust
Identifiability
Common ground
Privacy
Community Framework – Usability
Community context
Conviviality
Efficiency
Effectiveness
Individual context
Infrastructure
Media type
Network capacity
Computer capacity
Software
Navigation design
Community design
Information design
Consistent
Controllable
Predictable
Universal usability
Trustworthiness
• Is evidence of trustworthiness needed?
What are the implications for:
- social interaction?
- privacy and security?
• How can trust be assessed & communicated?
- what are the usability issues?
Social capital
‘A society characterized by general reciprocity is
more efficient than a distrustful society …’
Robert D. Putnam, Bowling Alone, 2000. P.21
Evaluating & measuring sociability
Purpose Number of messages
Amount of on-topic discussion
Level of interactivity
Degree of reciprocity
Quality of contribution
Satisfaction with social interactions
People
Number of participants
Number different types
Policies Flaming and uncivil behavior
Level of trustworthiness
Degree of empathy
Cyber-balkanization
‘Internet enables us to confine our
communication to precisely those people
who share our interests and are like us. …
Fragmentation and group polarization, are
significant risks.’
Cass Sunstein, republic.com, 2001, p. 192
Research: Empathy
‘Knowing what another person is feeling,
feeling what another person is feeling
and responding compassionately to
another person’
Levenson & Reuf, 1992
Analysis of 500 messages
45
40
35
30
25
20
15
10
5
0
Other
Narrative
Empathy
Factual
Evaluating & measuring usability
Dialog &
social
interaction
Information
design
Navigation
Access
Time to learn to read or send, move, etc.
Number of messages. Time to do a task
Satisfaction with dialog & interaction
Amount remembered. Number of errors
Time to read & understand.
Satisfaction with information design.
Amount of information remembered.
Number of misunderstandings
Time to learn to navigate application.
Time to complete navigation task.
Satisfaction with navigation.
Amount remembered. Wrong paths, errors
Can the software be run/down loaded?
Time to download. Response time.
Satisfaction with access