Keynote by David Redmiles @ International Conference on Global Software Engineering (ICGSE 2013), 26-29 Aug., Bari, Italy.

1. Awareness,
Trust,
and
So0ware
Tool
Support
in
Distance
Collabora8ons
David
Redmiles
Ins8tute
for
So0ware
Research
and
Department
of
Informa8cs
University
of
California,
Irvine
1

2. Thank
you
to
the
organizers!
Especially
to
Filippo
Lanubile
and
Marcelo
Cataldo
for
the
invita8on,
and
Teresa
Baldassarre
for
many
emails
organizing
my
trip.
2

3. Acknowledgement
of
Funding
(since
2004)
• Na8onal
Science
Founda8on
under
grants
534775,
0808783,
0943262,
1111446
• Department
of
Informa8cs
and
Donald
Bren
School
of
Informa8on
and
Computer
Sciences,
UC
Irvine
• Ins8tute
for
So0ware
Research,
UC
Irvine
• Center
for
Organiza8onal
Research,
UC
Irvine
• IBM,
Hitachi,
Intel
• Brazilian
Government
under
grant
CAPES
BEX
1312/99-­‐5
3

4. the
Irvine
team
in
2011
Werner
Beuschel
Erik
Trainer
Oliver
Wang
Ma`
Bietz
Hiroko
N.
Wilensky
David
Redmiles
Patrick
Shih
Ben
Koehne
Ban
Al-­‐Ani
Steve
Abrams

5. 5
Colleagues
at
PUCRS
Porto
Alegre
,
Brazil
Rafael
Prikladnicki
Sabrina
Marczak
Colleague
at
Vale
Ins8tute
of
Technology
and
Federal
University
of
Pará,
Brazil
Cleidson
de
Souza

6. Example:
Working
at
a
Distance
6

7. Some
of
the
problems
in
our
Example
• Isola8on
prevents
knowing
what
others
are
doing
• Lack
of
awareness
also
prevents
knowing
why
they
are
doing
or
not
doing
something.
• Distance
prevents
familiarity
–
both
professional
and
personal
7

8. Distance
Ma`ers
–
Common
Ground
/
Effects
of
Isola8on
• Olson,
G.,
Olson,
J.
Distance
Ma+ers,
Human-­‐Computer
Interac8on,
V.
15,
N.
2,
September
2000,
pp.
139-­‐178.
– Seminal
and
highly
cited
paper
on
the
research
of
geographically
distributed
teams.
– “four
key
concepts:
common
ground,
coupling
of
work,
collabora8on
readiness,
and
collabora8on
technology
readiness.”
• Koehn,
B.,
Shih,
P.,
Olson,
J.
Remote
and
Alone:
Coping
with
Being
the
Remote
Member
on
the
Team,
ACM
Conference
on
Computer-­‐
Supported
Coopera8ve
Work
(CSCW
2012,
Sea`le,
WA),
February
2012,
pp.
1257-­‐1266.
– Isolated
(remote)
workers
develop
individual
coping
strategies
involving
ICT
and
social
prac8ces.
– E.g.
par8cipants
developed
mentorship
rela8onships
and
communica8on
strategies
to
remain
visible
in
the
team
and
to
leave
visible
trails
for
performance
evalua8ons.
8

9. And
just
about
8me
zones
…
• Tang,
J.,
Zhao,
C.,
Cao,
X.,
Inkpen,
K.
Your
Time
Zone
or
Mine?
A
Study
of
Globally
Time
Zone-­‐Shi?ed
CollaboraAon,
ACM
Conference
on
Computer-­‐Supported
Coopera8ve
Work
(CSCW
2011,
Hangzhou,
China),
March
2011,
pp.
235-­‐244.
– Explores
how
team
members
work
across
global
8me
zone
differences
and
strategize
to
find
8me
for
interac8on.
– E.g.,
selec8ng
a
8me
zone
delegate
and
sharing-­‐the-­‐pain
strategies
• Segalla,
M.
Why
Mumbai
at
1pm
Is
the
Center
of
the
Business
World,
Harvard
Business
Review,
October2010,
pp.
38-­‐39.
– Amazing
sta8s8cs
and
visualiza8ons
about
the
lack
of
overlap
of
working
days
and
8mes
– E.g.,
“only
15
workweeks
(29%)
are
uninterrupted
by
a
holiday”
[p.
38].
9

10. Studies
of
distance
in
so0ware
collabora8ons
• B.
Cur8s,
H.
Krasner
and
N.
Iscoe.
A
Field
Study
of
the
So0ware
Design
Process
for
Large
Systems.
Communica8ons
of
the
ACM,
31(11):1268-­‐1287,
November
1988.
– CommunicaAon
and
CoordinaAon
Breakdowns
• Herbsleb,
J.D.,
Mockus,
A.,
Finholt,
T.A.,
and
Grinter,
R.E.
(2001).
An
empirical
study
of
global
so0ware
development:
Distance
and
speed.
Proceedings
of
the
23rd
Interna8onal
Conference
on
So0ware
Engineering
(ICSE
2001),
81-­‐90.
IEEE.
– Cross-­‐site
communicaAon
may
delay
problem
resoluAon
10

11. Recent
foci
on
so0ware
architecture
and
communica8on
• Cataldo,
M,
Herbsleb,
J.,
Carley,
K.
Socio-­‐Technical
Congruence:
A
Framework
for
Assessing
the
Impact
of
Technical
and
Work
Dependencies
on
So0ware
Development
Produc8vity,
Proceedings
of
the
Second
ACM-­‐IEEE
interna8onal
symposium
on
Empirical
so0ware
engineering
and
measurement
(ESEM'08,
Kaiserslautern,
Germany),
2008,
pp.
2-­‐11.
– When
coordinaAon
needs
and
communicaAon
align,
modificaAon
proceeds
more
efficiently
• de
Souza,
C.R.B.,
Redmiles,
D.F.
The
Awareness
Network,
To
Whom
Should
I
Display
My
Ac8ons?
And,
Whose
Ac8ons
Should
I
Monitor?,
IEEE
Transac8ons
on
So0ware
Engineering,
V.
37,
N.
3,
May/June
2011,
pp.
325-­‐340.
– Many
work
pracAces
are
needed
by
team
members
to
achieve
needed
communicaAon
around
a
so?ware
architecture
(structure)
11

12. Can
we
make
distance
ma`er
a
li`le
less?
• Awareness
• Trust
• So0ware
Tool
Support
12

13. Research
Approach
• Observe
and
collect
data
– Workplace
– Research
literature
• Hypothesize
and
build
systems
• Evaluate
systems
– Controlled
setngs
and
– Not
so
controlled
setngs
–
professionals
• Link
back
to
the
data
13
Observe
Explain
Design
Evalua8on
Theory
Systems

14. Why
this
approach?
• Computer
Science
– From
1976
–
1982
learned
about
the
mechanics
of
doing
things
with
the
computer
• Human-­‐Computer
Interac8on
– Around
1980
onwards
learned
about
the
real
way
people
used
computer
so0ware
– Formal
training
from
1987-­‐1992
in
human-­‐computer
interac8on
• Personally
– Pragma8c
– Open-­‐minded
– Seeking
“bigger”
picture
and
meaning
14

15. Roadmap
to
this
talk
• Themes
– Awareness,
Trust,
and
So0ware
Tool
Support
– Distributed
(Virtual)
Teams
– And,
more
generally,
distance
collabora8on
• For
each
of
Awareness,
Trust,
and
So0ware
Tool
Support
– Selected
literature
cita8ons
and
brief
summaries
– Experiences,
observa8ons,
prototype
so0ware
tools,
and
empirical
work
– Lessons
learned!
• Conclusion
– Immediate
and
long-­‐term
challenges
15

16. Awareness
16

17. Knowing
others’
ac8vi8es
• Dourish,
P.,
Bellot,
V.
Awareness
and
CoordinaAon
in
Shared
Workspaces,
Conference
on
Computer-­‐Supported
Coopera8ve
Work
(CSCW
'92,
Toronto,
Canada),
1992,
pp.
107-­‐114.
– “awareness
is
an
understanding
of
the
ac8vi8es
of
others,
which
provides
a
context
for
your
own
ac8vity”
– “awareness
informa8on
is
always
required
to
coordinate
group
ac8vi8es,
whatever
the
task
domain”
17

18. Work
prac8ces
for
coordina8on
• Schmidt,
K.
The
Problem
with
'Awareness'
-­‐
Introductory
Remarks
on
'Awareness
in
CSCW'.
Journal
of
Computer
Supported
Coopera8ve
Work,
2002.
11(3-­‐4):
p.
285-­‐298.
– Many
defini8ons
of
awareness,
but
…
– Monitoring
others’
and
displaying
your
own
ac8ons
as
part
of
work
18

19. Work
prac8ces
that
maintain
awareness
• de
Souza,
C.R.B.,
Redmiles,
D.F.
The
Awareness
Network,
To
Whom
Should
I
Display
My
Ac8ons?
And,
Whose
Ac8ons
Should
I
Monitor?,
IEEE
Transac8ons
on
So0ware
Engineering,
V.
37,
N.
3,
May/June
2011,
pp.
325-­‐340.
– Following
on
Schmidt
…
who
should
I
be
monitoring
and
to
whom
should
I
be
displaying
ac8ons.
19
[de
Souza
Redmiles
2011]

20. The
Awareness
Network
• How
do
social
actors
know
to
whom
they
should
display
ac8ons
and
whose
ac8ons
should
they
monitor?
• The
awareness
network
is
the
set
of
actors
whose
ac8ons
need
to
be
monitored
and
those
to
whom
one
needs
to
make
one’s
own
ac8ons
visible.
20
[de
Souza
Redmiles
2011]
Monitoring!

21. How
is
it
achieved?
• Read
everything!
– E.g.
emails,
design
documents,
problem
reports,
change
records
• Employ
a
personal
network!
– E.g.,
emailing
friends
who
might
know
etc.
• Ad
hoc
tools
– E.g.,
a
discussion
database
iden8fying
who
can
answer
what
ques8ons
21
[de
Souza
Redmiles
2011]

22. Where
is
our
data
from?
• 3
So0ware
Development
Projects
– Non
modular
legacy
so0ware
– Highly
modular
following
reuse
and
reference
architecture
– Adap8ng
so0ware
for
mobile
devices
• Data
Collec8on
– 51
semi-­‐structured
interviews
– Par8cipant
and
non-­‐par8cipant
observa8on
• Data
analysis
– Grounded
theory
methods
22
[de
Souza
Redmiles
2011]

23. Ariadne 1.0 - Social and Technical
Dependencies among Developers and
Components

24. Progression of Graphs to Brackets (1)
Developers Code

25. Progression of Graphs to Brackets (2)
Developers Code

26. Ariadne
2.0
26

27. Grand Overview – many
variables

29. Comparisons / Filter by Author

30. Comparisons / Filter by Artifact

31. Addi8onal
examples
of
visual
interface
features
to
compensate
for
distance,
especially
for
the
isola8on
…
• Sarma,
A.,
Redmiles,
D.,
van
der
Hoek,
A.
Palanwr:
Early
Detec8on
of
Development
Conflicts
Arising
from
Parallel
Code
Changes,
IEEE
Transac8ons
on
So0ware
Engineering,
V.
38,
N.
4,
June
2011,
pp.
889-­‐908.
– Visual
awareness
cues
(decorators)
could
help
developers
avoid
direct
and
indirect
conflicts
while
otherwise
working
isolaAon.
• Redmiles,
D.,
van
der
Hoek,
A.,
Al-­‐Ani,
B.,
Quirk,
S.,
Sarma,
A.,
Silva
Filho,
R.,
de
Souza,
C.,
Trainer,
E.
Con8nuous
Coordina8on:
A
New
Paradigm
to
Support
Globally
Distributed
So0ware
Development
Projects,
Wirtscha0sinforma8k,
V.
49,
2007,
pp.
S28-­‐S38.
– ConAnuous
coordinaAon
is
required
in
distributed
so?ware
development
[even
when
highly
structured].
– Awareness
can
support
conAnuous
coordinaAon
and
be
greatly
achieved
by
so?ware
tools.
31

32. Awareness
–
Lessons
Learned
-­‐
Tools
• Socio-­‐technical
systems
– Ariadne
(and
other
systems)
integrate
both
the
social
and
technical
elements
• Visual
user
interface
• So0ware
tools
can
help
awareness
– E.g.,
in
iden8fying
colleagues
– E.g.,
in
perceiving
situa8ons
such
as
bo`lenecks
– E.g.
in
avoiding
conflicts
32

33. Awareness
–
Lessons
Learned
-­‐
Behavior
• Awareness
is
key
to
coordinated
work
• Yet
awareness
and
common
ground
are
hard
to
achieve
at
a
distance
• There
are
prac8ces
that
are
a
part
of
work
that
an8cipate
awareness
– Specifically,
to
establish
and
maintain
an
awareness
network
33

34. Some
of
the
problems
in
our
Example
• Isola8on
prevents
knowing
what
others
are
doing
• Lack
of
awareness
also
prevents
knowing
why
they
are
doing
or
not
doing
something.
• Distance
prevents
familiarity
–
both
professional
and
personal
34

35. Trust
35

36. Trust
emerging
as
a
theme
• Al-­‐Ani,
B.,
Redmiles,
D.
In
Strangers
We
Trust?
Findings
of
an
Empirical
Study
of
Distributed
Development,
IEEE
Interna8onal
Conference
on
Global
So0ware
Engineering
(ICGSE,
Limerick,
Ireland),
July
2009,
pp.
121-­‐130.
– Re-­‐examining
data
from
open-­‐ended
interviews
at
a
Fortune
500
company
on
distributed
collabora8on
– Without
asking,
interviewees
stated
that
the
greatest
concern
around
successful
collabora8on
was
“trust”
– The
emergence
of
trust
as
a
theme
36
[Al-­‐Ani,
Redmiles
2009]

37. Defini8ons
of
trust
…
• Jarvenpaa,
S.
L.,
Knoll,
K.,
and
Leidner,
D.
E.
Is
anybody
out
there?
antecedents
of
trust
in
global
virtual
teams,
J.
Manage.
Inf.
Syst.
V.
14,
No.
4,
March,
1998,
pp.
29-­‐64.
– Ra8onal
trust
–
willingness
to
be
less
“self-­‐protec8ve”
and
take
risks
– Social
trust
–
a
duty
or
right
way
to
behave
creates
the
willingness
to
take
risks
• Wilson,
J.M.,
Straus,
S.G.
&
McEvily,
W.J.
All
in
due
Ame:
The
development
of
trust
in
computer-­‐mediated
and
face-­‐to-­‐face
groups,
Organiza8onal
Behavior
and
Human
Decision
Processes,
99,
2006,
pp.
16-­‐33.
– Cogni8ve
trust
–
beliefs
about
others’
competence
and
reliability
– Affec8ve
trust
–
beliefs
about
reciprocated
concern,
emo8onal
8es
and
such
37
[Al-­‐Ani,
Redmiles
2009]

38. The
Role
of
Trust
One
party’s
posi4ve
expecta4ons
of
another
• Trust:
– Enhances
team
produc8vity
– Helps
teams
manage
uncertainty
and
complexity
of
working
remotely
– Promotes
influen8al
informa8on
exchange
– Fosters
innova8on
38
[Al-­‐Ani,
Redmiles
2009]

39. First
Field
Study:
examining
distributed
collabora8on
• Interviews
were
conducted
with
employees
of
a
large
mul8-­‐na8onal
organiza8on.
• USA
with
16
par8cipants.
• Respondents
men8oned
a
total
of
26
different
sites.
• Overall
there
were
an
average
of
4
sites
per
distributed
team.
39
[Al-­‐Ani,
Redmiles
2009]

40. Study
Overview
40
{balani|redmiles}@ics.uci.edu
Seq.
Purpose
Interview
Framework
Par8cipant
Background
(educa8on,
experience…etc)
Project
A:
Collocated
Project
Descrip8ons
and
Team
Structure
Project
B:
Distributed
Project
Decomposi8on
and
Task
Assignment
Communica8on
Leadership
Social
Behavior
and
Tool
Support
Establish the following:
1. Demographics,
2. Participant
terminology,
3. Points of reference,
4. Comparative
evaluation,
5. Problem domain.
Gain understanding
1. How developers
identify tasks,
2. How tasks are
allocated to
developers,
3. Challenges.
Investigate:
1. Models,
2. Types,
3. Efficiency
and
effectiveness
What
impact
does
the
locality
of
the
leader
have
on
team
dynamics?
How
do
developers
exchange
ideas?
[Al-­‐Ani,
Redmiles
2009]

41. Common
thread:
trust
41
Trust
Project
A:
Collocated
Project
Descrip8ons
and
Team
Structure
Project
B:
Distributed
Leadership
Communica8on
Social
Behavior
and
Tool
Support
[Al-­‐Ani,
Redmiles
2009]

42. Lessons
Learned
–
Factors
Influencing
Trust
• The
issue
of
trust
was
raised
by
respondents:
– Team
size:
larger
teams.
– Project
type:
innova8ve
new.
– Team
diversity:
high
diversity.
– Leadership:
strong
leadership.
42
[Al-­‐Ani,
Redmiles
2009]

43. Trust:
Compe8ng
Facets
43
team
diversity
8me
Trust
Threshold
-­‐
+
leadership
team
size
project
type
[Al-­‐Ani,
Redmiles
2009]

44. Imagine
collabora8on
without
trust!
• Double
checking.
• Working
in
isola8on.
• Reluctance
to
share
informa8on.
44
[Al-­‐Ani,
Redmiles
2009]

45. An
example
45
Y
have
a
tendency
to
talk
longer
X
are
very
impaAent
to
leave
when
it
is
the
end
of
the
working
day
[in
their
country].
“engineers
in
X
feel
they
are
superior
and
a
level
of
arrogance.
With
this
comes
a
level
of
mistrust
of
us”
“you
don’t
need
to
know
this
part
of
the
code
you
wouldn’t
understand
it”
[Al-­‐Ani,
Redmiles
2009]

46. Some
of
the
problems
in
our
Example
• Isola8on
prevents
knowing
what
others
are
doing
• Lack
of
awareness
also
prevents
knowing
why
they
are
doing
or
not
doing
something.
• Distance
prevents
familiarity
–
both
professional
and
personal
46

47. Second
Field
Study:
examining
trust
in
par8cular
• What
are
the
antecedents
of
trust
in
distributed
teams?
• What
are
the
behaviors
and
ac8ons
that
team
members
engage
in
that
most
frequently
engender
trust?
• What
would
help
developers
trust
others
on
their
teams?
47

48. 48
Interview
protocol
• Direct
but
open
ended
ques8ons
– Background
and
project
• Scenarios
(contextualized
to
interview)
– You
are
working
on
…
you
need
…
who
would
you
ask?
• Storytelling
– Can
you
tell
me
an
instance
when
…
tell
me
a
story
…

49. Degree
of
trust
“Game”
in
Protocol
49

50. Sought
out
interna8onal
collaborators
for
addi8onal
data!
• Thanks
to
…
– Drs.
Rafael
Prikladnicki
and
Sabrina
Marczak,
both
at
the
Pon8‚cia
Universidade
Católica
do
Rio
Grande
do
Sul
–
PUCRS
in
Porto
Alegre.
50

51. 51
Field
sites
• 5
mul8-­‐site
and
mul8-­‐na8onal
organiza8ons.
• Each
organiza8on
is
considered
one
of
the
leaders
in
the
development
of
computer-­‐based
systems.
• Interview
subjects
were
recruited
through
e-­‐
mails
sent
to
a
cross-­‐sec8on
of
the
organiza8ons,
as
well
as
word
of
mouth
(snowball).
[Al-­‐Ani,
Wang,
Marczak
et
al.,
2012]

52. Par8cipants
• 18
female
and
43
male
employees.
• On
average,
11
years’
experience
working
in
distributed
teams
and
12
years’
experience
in
the
organiza8on.
• Roles
in
one
of
3
broad
categories:
– managers
-­‐
21
(e.g.
project
manager,
por„olio
manager),
– developers
-­‐
35
(e.g.
tester,
so0ware
designer,
system
architect,
business
analyst)
and
– support
staff
-­‐
5
(e.g.
lawyer,
quality
assurance).
• Located
in
the
USA
(34),
Brazil
(18),
Mexico
(2),
and
Costa
Rica,
Ireland,
Israel,
Poland,
China,
Taiwan,
and
Malaysia
(1
each)
[Al-­‐Ani,
Wang,
Marczak
et
al.,
2012]

53. Example
Analysis
and
Result
• Al-­‐Ani,
B.,
Wang,
Y.,
Marczak,
S.,
Trainer,
E.,
Redmiles,
D.
Distributed
Developers
and
the
Non-­‐
Use
of
Web
2.0
Technologies:
A
Proclivity
Model,
The
7th
Interna8onal
Conference
on
Global
So0ware
Engineering
(ICGSE
2012,
Porto
Alegre,
Brazil),
August
2012,
pp.
104-­‐113.
– Web
2.0
technologies
allow
employees
to
build
a
familiarity
with
one
another
and
share
informa8on
and
should
improve
trust.
– However,
less
than
25%
of
our
study
par8cipants
adopted
these
technologies
and
most
have
a
nega8ve
view
of
these
technologies
– Why?
53
[Al-­‐Ani,
Wang,
Marczak
et
al.,
2012]

54. Analysis
• Interviews
were
transcribed
and
coded
using
Atlas.8
(h`p://www.atlas8.com/index.html)
• Qualita8ve
analysis
– Examining
interviewees
comments
– Iden8fying
themes
• Quan8ta8ve
analysis
– Variables
derived
from
coded
interviews,
including
self-­‐reported
demographics
– Various
sta8s8cal
techniques
but
in
this
instance,
logis8c
regression
54
[Al-­‐Ani,
Wang,
Marczak
et
al.,
2012]

55. Variables
Examined
55
Variable
Meaning
Usage
The
usage
of
Web
2.0
technologies
Language
Whether
an
interviewee
can
speak
more
than
one
language.
EducaAon
Whether
an
interviewee
holds
a
postgraduate
degree.
Gender
An
interviewee’s
gender.
AGE
An
interviewee’s
age.
Experience
at
Distributed
Development
An
interviewee’s
experience
with
distributed
so0ware
development.
Job
-­‐
Manager
Whether
an
interviewee
is
a
manager
or
not.
Job
-­‐
Technical
Whether
an
interviewee’s
job
is
technical-­‐
oriented
or
not.
Use
of
(non
Web
2.0)
other
technologies
The
number
of
communica8on
technologies
an
interviewee
has
been
used
in
their
work
except
Web
2.0
technologies.
[Al-­‐Ani,
Wang,
Marczak
et
al.,
2012]

56. Results
of
Quan8ta8ve
Analysis
56
[Al-­‐Ani,
Wang,
Marczak
et
al.,
2012]
Variables
Conclusion
Age
An
increase
of
age
will
result
the
lower
probability
of
using
Web
2.0
to
support
distributed
collaboraAon.
Experience
at
Distributed
Development
An
increase
of
experience
of
distributed
development
will
result
the
higher
probability
of
using
Web
2.0
to
support
distributed
collaboraAon.
Use
of
(non
Web
2.0)
other
technologies
An
increase
of
using
other
CommunicaAon
Technology
will
result
the
higher
probability
of
using
Web
2.0
to
support
distributed
collaboraAon.

57. Results
of
Qualita8ve
Analysis
• The
alignment
between
developers’
work
and
their
suppor8ng
technology
is
posi8vely
associated
with
developers’
trust
towards
collabora8on
tools.
• The
experience
of
being
exposed
to
distributed
so0ware
development
is
posi8vely
associated
with
developers’
trust
towards
collabora8on
tools.
• Posi8ve
organiza8on
policies
on
collabora8on
tools
are
posi8vely
associated
with
developers’
usage
of
tradi8onal
collabora8on
tools.
57
[Al-­‐Ani,
Wang,
Marczak
et
al.,
2012]

58. Results
from
the
management
literature
…
• Jarvenpaa,
S.
L.,
Shaw,
T.
R.,
and
Staples,
D.
S.
Toward
contextualized
theories
of
trust:
The
role
of
trust
in
global
virtual
teams.
In
Informa8on
Systems
Research
15,
3
(2004),
250-­‐267.
– Early
trust
is
cri8cal
to
communica8on
and
performance
– Effects
of
structure
• Teams
with
high
structure
are
less
dependent
on
trust
and
communica8on.
• Teams
with
low
structure
…
– Implica8ons
• For
managers,
what
is
the
right
amount
of
trust
[and
communica8on]
to
encourage?
58

59. • Zolin,
R.,
Hinds,
P.,
Fruchter,
R.
and
Levi`,
R.
(2004).
Interpersonal
trust
in
cross-­‐func8onal,
geographically
distributed
work:
A
longitudinal
study.
Informa8on
&
Organiza8ons,
14,
1-­‐26.
– Trust
is
a
willingness
to
accept
vulnerability
to
others
…
– Trust
is
“one
of
the
major
challenges,”
“central
to
teamwork,”
especially
due
to
“many
sub-­‐tasks
are
interdependent,”
etc.
– IniAal
trust
is
criAcal
to
future
percepAons
– Cultural
diversity
negaAvely
affects
trust
– Factors:
cultural
diversity,
perceived
trustworthiness
(trustor’s
propensity
to
trust,
percepAons
of
follow-­‐
through),
risk
and
reward.
59

60. Lessons
Learned
• The
literature
emphasizes
the
importance
of
trust:
– Effec8ve
communica8on
and
team
collabora8on
• Our
first
study
revealed
compe8ng
factors
influencing
trust
such
as
– Team
diversity,
team
size,
project
type,
leadership,
and
8me
• Our
second
study
indicated
paths
to
be`er
tools
/
be`er
adop8on
– Experience
in
tool
usage
increases
everyday
–
in
personal
as
well
as
professional
use.
– Knowing
the
value
of
“Web
2.0”
tools
can
encourage
changed
organiza8onal
policies.
– Support
for
“ver8cal”
integra8on
–
value
for
many
par8cipants
–
can
increase
adop8on.
• Encouragement
for
the
poten8al
value
of
tools!
60

61. Tool
Support
Specific
to
Trust
61

62. Knowing
personal
or
professional
(exper8se)
informa8on?
• Schumann,
J.,
Shih,
P.,
Redmiles,
D.,
Horton,
G.
Suppor8ng
Ini8al
Trust
in
Distributed
Idea
Genera8on
and
Evalua8on,
The
2012
Interna8onal
ACM
SIGGROUP
Conference
on
Suppor8ng
Group
Work
(GROUP
2012,
Sanibel
Island,
FL),
October
2012,
in
press.
– Effects
of
cogni4ve
and
affec4ve
trust
on
collabora8ve
brainstorming
and
evalua8on.
– Open
to
gender
effects
(as
inspired
by
Professor
Margaret
Burne`,
Oregon
State).
62
[Schumann,
Shih,
Redmiles,
Horton,
2012]

63. Innova8on
and
Trust
• Cogni8ve
Trust
– Judgment
of
competence,
reliability,
and
professionalism
– Deliberate
assessment
of
benefits
of
trus8ng
over
risks
• Affec8ve
Trust
– Emo8onal
8es
among
individuals,
beliefs
about
interpersonal
care
and
concerns
– Sincere
concern
for
the
well-­‐being
of
the
others
• Innova8on
Process
– Idea
Genera8on
– Idea
Evalua8on
[Schumann,
Shih,
Redmiles,
Horton,
2012]

64. Trust
Informa8on
Elements
Personal
informa4on # Exper4se
Informa4on #
Hobbies 14 Experience
(projects) 15
Gender 13 Specific
skills 15
Honorary
ac8vi8es 12 Specializa8on/interests 14
Age 11 References
(awards) 14
Na8onality 8 Degree
(years
in
the
program) 12
Taste
of
music 7 Companies 8
TV
shows 6 Department 7
[Schumann,
Shih,
Redmiles,
Horton,
2012]

65. The
Experiment
• Idea
Genera8on
– Par8cipants
work
to
generate
ideas
– Simultaneously,
2
remote
confederates
produced
10
pre-­‐compiled
ideas
in
the
15-­‐min
session.
• Idea
Evalua8on
– Each
par8cipant
rated
6
ideas.
– Originality
and
feasibility
ra8ngs
of
the
confederates
were
pre-­‐compiled.
• 36
Subjects
– 18
Male
– 18
Female
[Schumann,
Shih,
Redmiles,
Horton,
2012]

66. Idea
Genera8on
Screen
[Schumann,
Shih,
Redmiles,
Horton,
2012]

67. Idea
Evalua8on
Screen
[Schumann,
Shih,
Redmiles,
Horton,
2012]

68. Results
–
Support
for
Trust
• Knowing
personal
informa8on
leads
to
higher
affec8ve
trust
and
knowing
exper8se
informa8on
leads
to
higher
cogni8ve
trust
–
expected.
• However,
knowing
either
personal
or
exper8se
informa8on
boosted
both
trust
levels
–
par8cipants
did
not
make
dis8nc8ons.
68
[Schumann,
Shih,
Redmiles,
Horton,
2012]

69. Results
–
Gender
Effects
• Gender
differences
have
li`le
effect
on
trust
in
idea
genera8on
and
idea
evalua8on
sessions.
• However,
female
par8cipants
created
more
feasible
ideas
while
male
par8cipants
created
more
original
ideas
in
the
experiment
69
[Schumann,
Shih,
Redmiles,
Horton,
2012]

70. Lessons
Learned
–
Tool
Support
for
Trust
• Evidence
that
informa8on
provided
by
tools
can
engender
trust.
• Further
encouragement
towards
tool
support.
70

71. Lessons
from
our
colleagues
here
in
Bari!
• Calefato,
F.,
Lanubile,
F.
AugmenAng
Social
Awareness
in
a
CollaboraAve
Development
Environment,
the
5th
Int'l
Workshop
on
Coopera8ve
and
Human
Aspects
of
So0ware
Engineering
(CHASE'12),
Zurich,
Switzerland,
2
Jun.
2012,
pp.
12-­‐14.
– Integra8ng
social
media
with
collabora8ve
so0ware
development
environments
• Calefato,
F.,
Lanubile,
F.
Can
Social
Awareness
Foster
Trust
Building
in
Global
So?ware
Teams?,
the
5th
Interna8onal
Workshop
on
Social
So0ware
Engineering
(SSE'13),
St.
Petersburg,
Russia,
18
Aug.
2013
(to
appear).
• Examining
the
larger,
social
network
dimension
to
collabora8ve
projects,
exploi8ng
socio-­‐technical
informa8on
to
promote
collabora8ve
ac8vi8es:
caring,
browsing,
climbing,
and
campaigning.
71

72. Toward
a
Design
Space
for
Collabora8on
Tools
• Trainer,
E.H.,
Redmiles,
D.F.
Founda8ons
for
the
Design
of
Visualiza8ons
that
Support
Trust
in
Distributed
Teams,
Interna8onal
Working
Conference
on
Advanced
Visual
Interfaces
(AVI
2012,
Capri
Island,
Italy),
May
2012,
pp.
34-­‐41.
72
[Trainer,
Redmiles,
2012]

73. A
Connec8on
between
Awareness
and
Trust
in
Tool
Support
A
so?ware
tool
can
usefully
provide
informaAon
that
engenders
perceived
trustworthiness
among
distributed
team
members.
• Ques8ons:
– What
informa8on
affects
distributed
team
members
percep8ons
of
others’
trustworthiness?
– Can
this
informa8on
be
delivered
in
a
so0ware
tool?
73
[Trainer,
Redmiles,
2012]

74. Collabora8ve
Traces
• A
term
that
refers
to
data
visualized
by
“awareness”
tools
• Representa8ons
of
past
and
current
ac8vity
of
a
group
of
developers
manipula8ng
so0ware
development
ar8facts
74
Bug
Tracker
CM
System
E-­‐mail
Server
[Trainer,
Redmiles,
2012]

75. Collabora8ve
Traces
for
Trust
• (RQ)
“What
informa8on…..”
• As
shown
by
a
matrix…..
• Columns:
– Trust
factors,
i.e.
informa8on
that
affects
trust,
from
the
literature
on
trust
• Rows:
– Collabora8ve
traces
+
other
data
(e.g.,
8me
zone,
org.
chart)
<see
figure
on
next
slide>
75
[Trainer,
Redmiles,
2012]

76. Collabora8ve
Traces
for
Trust
76
TRUST
FACTORS
COLLABORATIVE
TRACES
X X X X
X X X
X X X X X X
X X X X X X
X X X X X X
X X X X
X X
X X X X X X X
X X X X X X X
X X X X X X X X X
X X X X X X X
X X X X X X
X X
X
X X
X X X
Ini8a8on
and
response
Reputa8on
Assigned
Work
Item
E-­‐mail
Change
Set
Exper8se
[Trainer,
Redmiles,
2012]

77. Visual
Representa8ons
for
Trust
• Visual
representa8ons
summarize
informa8on
provided
by
CTs
• How
to
choose
appropriate
visualiza8ons?
– Web-­‐based
advice
(e.g.,
ManyEyes,
Swivel,
Google
Chart
Tools)
organized
by
task:
• Show
rela8onships
(node-­‐edge,
matrices)
• Show
hierarchy
(trees,
circle
packing)
• Compare
numerical
values
(bar
charts)
77
[Trainer,
Redmiles,
2012]

78. Visual
Representa8ons
and
Trust
Factors
78
TRUST
FACTORS
VISUAL
RERESENTATIONS
X X X X
X X X
X X X X X
X X X X X X
X X X X X X
X X X X
X X
X X X X X X X
X X X X X X X
X X X X X X X X X
X X X X X X X
X X X X X X
X X
X
X X
X X X
Same
loca8on
Map
Ini8a8ons
and
Response
Bar
Charts
Role
Circle
Packing
[Trainer,
Redmiles,
2012]

79. Visual
Representa8ons
and
Collabora8ve
traces
79
COLLABORATIVE
TRACES
VISUAL
RERESENTATIONS
X X X X
X X X
X X X X X
X X X X X X
X X X X X X
X X X X
X X
X X X X X X X
X X X X X X X
X X X X X X X X X
X X X X X X X
X X X X X X
X X
X
X X
X X X
Bar
Charts
E-­‐mails
Source-­‐code
Node-­‐
edge
Line-­‐
based
Assigned
Work
Items
[Trainer,
Redmiles,
2012]

80. A
Design
Space
• Model
of
Design
Space
=
{
Trust
factors,
Visual
representa8ons,
Collabora8ve
traces
}
80
The
space
is
comprised
of
3
matrices:
1. Trust
Factors
x
Collabora8ve
Traces
2. Collabora8ve
Traces
x
Visual
Representa8ons
3. Visual
Representa8ons
x
Trust
Factors
*
[Trainer,
Redmiles,
2012]

81. 1.
Availability
Radar
• Groups
developers
by
their
proximity
to
the
current
user
81
Further
horizontal
distances
from
center
indicate
greater
physical
distance.
• White
x
=
non-­‐manager
• Black
x
=
manager
Visual
Representa4on:
CirclePacking
(fla`ened)
Collabora4ve
Trace:
Organiza8onal
Charts,
Work
site
loca8on,
8me
zone
[Trainer,
Redmiles,
2012]

82. 2.
Responsiveness
Bars
• “Bins”
developers’
reply
8mes
to
e-­‐mails
based
on
8me
to
reply
observed
in
in
org.
literature
– Same
day
– Next
day
– Within
5
days
82
Visual
Representa4on:
Bar
charts
Collabora4ve
Traces:
E-­‐mail,
(instant
messages,
mailing
list
pos8ngs)
[Trainer,
Redmiles,
2012]

83. 3.
Time
Zone
Overlap
(2)
• Show
overlap
in
8mes
of
the
day
– Green
(8am-­‐5pm)
– Yellow
(6pm-­‐9pm,
7am))
– Red
(10pm-­‐6am)
• Time
on
e-­‐mail
(black
dots)
– “Day
laborers”
– “Email-­‐aholics”
83
[Trainer,
Redmiles,
2012]

84. Lessons
Learned
• The
design
space
presented
here:
– Is
a
first
step
toward
exploring
whether
visual
interfaces
can
engender
perceived
trustworthiness
– Can
be
of
value
to
designers
of
visual
interfaces…and
ul8mately
to
distributed
so0ware
developers
– In
a
next
step,
we
empirically
evaluated
an
interface
conceived
within
this
design
space.
84
[Trainer,
Redmiles,
2012]

85. Empirical
Support
for
a
Visual
Tool:
A
Controlled
Study
• 40
human
subjects
/
par8cipants
– 28
graduate
students
with
at
least
1
year
experience
in
so0ware
development
– 12
professional
so0ware
developers
from
2
so0ware
companies
• Quan8ta8ve
and
Qualita8ve
Analysis
85

86. Scenario:
Consider
a
Remote
Co-­‐
worker’s
Failure
to
Deliver
on
Time
You
have
to
come
into
the
office
this
weekend
to
work
on
the
“MIRTH”
project.
Victor
Ward,
a
so?ware
engineer
on
your
team,
failed
to
check
in
his
source-­‐code
changes
on
Ame,
and
has
not
been
responsive
over
e-­‐mail.
As
a
result,
you
are
not
able
to
integrate
your
new
changes
into
the
build,
and
the
project
has
slipped
a
week
behind
schedule.
(a
scenario
based
on
our
field
interviews)
86

87. THESEUS
87
A
very
busy
colleague!

88. THESEUS
88
A
not
so
busy
colleague!

89. Measuring
A`ribu8on
and
Trust
• A`ribu8on
Ranking
– Given
what
you
know
about
how
people
behave,
which
explanaAon
do
you
think
most
likely
describes
why
Victor
was
unable
to
deliver
on
Ame?
(example)
– Situa4onal
aMribu4ons
reflect
high
perceived
trustworthiness.
Disposi4onal
aMribu4ons
reflect
low
perceived
trustworthiness.
• Standardized
Ques8onnaire
– Standard
specific
interpersonal
trust
(Johnson-­‐
George
&
Swap,
1982),
measures
one’s
perceived
trustworthiness
toward
a
specific
individual
(5-­‐pt.
Likert
items)
89

90. THESEUS and
A`ribu8ons
90
Technique
Result
One-­‐way
repeated
measures
ANOVA
Significant
effect
of
Theseus
on
a`ribu8on
type
[F(3,
117)
=
25.96,
p<0.001,
par8al
=
0.40].
Scores
range
from
-­‐7
(highly
situa8onal)
to
7
(highly
disposi8onal),
with
a
neutral
score
or
midpoint
of
0.
€
η2
€
µ =-­‐1.10
€
µ =-­‐3.36
€
µ =-­‐2.02
€
µ =3.57
Legend
a
Baseline
(w/o
Theseus
tool)
b
Theseus
-­‐
situa8onal
c
Theseus
-­‐
mixed
d
Theseus
-­‐
disposi8onal
Standard Deviation of Attribution Scores.

91. THESEUS
and
Interpersonal
Trust
91
Technique
Result
One-­‐way
repeated
measures
ANOVA
Significant
effect
of
Theseus
on
interpersonal
trust
score
[F(3,
117)
=
27.03,
p<0.001,
par8al
=
0.41].
Scores
range
from
15
(low
trust)
to
75
(highest
trust),
with
a
neutral
score
or
midpoint
of
45.
€
η2
=µ 44.13
€
µ = 54.70
€
µ = 46.12
€
µ =39.60
Legend
a
Baseline
(w/o
Theseus
tool)
b
Theseus
-­‐
situa8onal
c
Theseus
-­‐
mixed
d
Theseus
-­‐
disposi8onal
Standard Deviation of Interpersonal Trust Scores.

92. Lessons
Learned
–
Tool
Support
• Theseus
results
in
higher
perceived
trustworthiness
compared
with
no
Theseus
• Theseus
results
in
more
situa8onal
a`ribu8ons
compared
with
no
Theseus
(marginal
support)
• Based
on
subject
feedback,
the
tool
is
usable
• Subjects
quickly
became
immersed
in
the
data
92

93. Conclusions
93

94. A
progression
in
research
• Awareness
– And
tool
support
for
collabora8on
• But
while
we
studied
teams
in
the
field
– Trust
emerged
as
a
major
concern
• We
suspected
the
awareness
tools
we
previously
research
could
help
…
– But
exactly
how?
94

95. Arriving
at
Support
for
Trust
• We
realized
from
our
field
data
that
– Typical
Web
2.0
tools
should
help
…
– But
in
many
cases
went
unused.
– But
some
team
member
characteris8cs
and
some
teams
using
Web
2.0
showed
promise
95

96. Pursuing
tools
further
…
• What
kinds
of
tools
could
support
trust?
– What
kind
of
informa8on
would
they
need
to
provide?
• Cogni8ve
and
affec8ve
trust
…
but
with
a
revela8on
about
the
impact
of
each.
• Situa8onal
and
disposi8onal
informa8on
for
making
accurate
a`ribu8ons.
96

97. Providing
a
design
space
for
tools

98. Some
of
the
problems
in
our
Example
• Isola8on
prevents
knowing
what
others
are
doing
• Lack
of
awareness
also
prevents
knowing
why
they
are
doing
or
not
doing
something.
• Distance
prevents
familiarity
–
both
professional
and
personal
98

99. Research
Approach
• Observe
and
collect
data
– Workplace
– Research
literature
• Hypothesize
and
build
systems
• Evaluate
systems
– Controlled
setngs
and
– Not
so
controlled
setngs
–
professionals
• Link
back
to
the
data
99
Observe
Explain
Design
Evalua8on
Theory
Systems

100. Finally
• The
problems
and
facets
are
– Bigger
than
one
person,
one
approach,
etc.
• Hope
others
will
join
the
pursuit.
100

101. Workshop
Themes
-­‐
factors
that
engender
and
inhibit
trust.
-­‐
overarching
trust
framework.
-­‐
so0ware
tools
support
trust.
Workshop
on
Trust
in
Virtual
Teams:
Theory
and
Tools
h`p://collab.di.uniba.it/trus`heorytools/
16th
ACM
Conference
on
Computer
Supported
Coopera8ve
Work
and
Social
Compu8ng
(CSCW
2013)
will
be
held
February
23-­‐27
in
San
Antonio,
Texas,
USA

102. Diversity
in
Research
Domains
and
Perspec.ves
• Markus
Rohde:
Trust
in
Electronically-­‐Supported
Networks
of
Poli4cal
Ac4vists
• Rasmus
Eskild
Jensen:
Commitment
manifested
in
ac8vity:
A
non-­‐
instrumental
approach
to
commitment
in
virtual
teams
• Bruno
S.
Nascimento,
Adriana
S.
Vivacqua,
Marcos
R.S.
Borges:
Establishing
Trust
in
Cri8cal
Situa8ons
(e.g.
emergency
response)
• Sabrina
Marczak,
Ban
Al-­‐Ani,
David
Redmiles,
Rafael
Prikladnicki:
Designing
Tools
to
Support
Trust
in
Distributed
SoSware
Teams
• Lionel
P.
Robert
Jr.:
Trust
and
Control
in
Virtual
Teams:
Unraveling
the
impact
of
Team
Awareness
Systems
in
Virtual
Teams
• Fabio
Calefato,
Filippo
Lanubile,
Nicole
Novielli:
Social
Media
and
Trust
Building
in
Virtual
Teams:
The
Design
of
a
Replicated
Experiment
• Yoon
Suk
Lee,
Marie
C.
Paret,
Brian
M.
Kleiner:
Non-­‐equivalent
Communica.on
Technology
Impact
on
Trust
in
Par8ally
Distributed
Conceptual
Design
Teams

103. Hope
this
serves
to
inspire
you!

104. Comments?
Ques8ons?
104
h`p://cradl.ics.uci.edu

105. Addi8onal
Results
105

106. Trust
is
not
just
a
snapshot,
but
a
process!
• Al-­‐Ani,
B.,
Bietz,
M.,
Wang,
Y.,
Trainer,
E.,
Koehne,
B.,
Marczak,
S.,
Redmiles,
D.,
Prikladnicki,
R.
Globally
Distributed
System
Developers:
Their
Trust
Expecta8ons
and
Processes,
The
16th
ACM
Conference
on
Computer
Supported
Coopera8ve
Work
and
Social
Compu8ng
(CSCW
2013,
San
Antonio,
Texas),
February
2013,
pp.
563-­‐573.
– Development
of
trust,
adapta8on,
and
repair.
• Al-­‐Ani,
B.,
Trainer,
E.,
Redmiles,
D.,
Simmons,
E.
Trust
and
Surprise
in
Distributed
Teams:
Towards
an
Understanding
of
Expecta8ons
and
Adapta8ons,
The
4th
ACM
Interna8onal
Conference
on
Intercultural
Collabora8on
(ICIC
2012,
Bengaluru,
India),
March
2012,
pp.
97-­‐106.
– “Cultural
surprises”
and
adapta8on.
106

107. More
evidence
for
the
importance
of
“personal”
interac8ons
to
work
• Wang,
Yi,
Redmiles,
D.
Understanding
Cheap
Talk
and
the
Emergence
of
Trust
in
Global
So?ware
Engineering:
An
EvoluAonary
Game
Theory
PerspecAve,
The
6th
Interna8onal
Workshop
on
Coopera8ve
and
Human
Aspects
of
So0ware
Engineering
(CHASE
2013),
held
in
conjunc8on
with
the
35th
Interna8onal
Conference
on
So0ware
Engineering
(ICSE
2013,
San
Francisco,
California),
May
25,
2013,
(in
press).
– We
observed:
• Cheap
talk
is
prevalent
in
GSE
teams
(32/41
interviewees
use
it).
• Significantly
higher
trust
(t-­‐test)
is
found
among
those
who
engage
in
cheap
talk
in
their
interac8ons
(P-­‐value:
0.013;
Effect
Size:
0.921).
– And,
thus,
we
are
mo8vated
to
inves8gate:
• How
trust
emerges
in
collabora8ons
where
cheap
talk
is
present;
and
• Whether
cheap
talk
increases
the
probability
of
coopera8on
while
promo8ng
trust.
107

108. Going
forward,
further
108

109. Medium-­‐sized
Ques8ons
Asked
Above
–
Trust
and
Collabora8on
• Looking
for
antecedents
of
trust
/
“trust
factors”
and
can
there
be
so0ware
support?
• Which
has
more
effect
on
collabora8ve
tasks
…
cogni8ve
or
affec8ve
trust?
• What
informa8on
can
be
“mined”
to
support
tools
to
engender
trust?
• What
informa8on
needs
to
be
mined,
presented
to
collaborators,
and
can
it
engender
trust?
• Trust
is
not
just
a
snapshot,
but
a
process!
• And
social
compu8ng
“Web
2.0”
tools
might
not
always
be
used
…
• And
how
can
we
look
beyond
the
limits
of
our
data?
109

110. Big
Ques8ons
–
Trust
and
Collabora8on
• Can
we
make
distance
ma`er
less?
– What
is
the
role
of
structure
(process)?
– How
much
trust
should
their
be?
• What
kind
of
trust?
Affec8ve,
Cogni8ve?
Which
when?
– How
much
communica8on?
• What
kind
of
communica8on?
– How
to
engender
trust?
Communica8on?
– What
is
the
role
of
culture?
Culture
interacts
with
structure?
Is
culture
/
structure
a
choice?
• Can
we
make
distance
an
advantage?
– Can
we
make
the
virtual
environment
richer
than
presence?
– What
informa8on
goes
into
such
a
rich
environment
• Trust
factors?
• Collabora8ve
Traces?
• Ac8vity
Traces?
110