• Save
"Awareness, Trust, and Software Tool Support in Distance Collaborations" by D. Redmiles
Upcoming SlideShare
Loading in...5
×
 

"Awareness, Trust, and Software Tool Support in Distance Collaborations" by D. Redmiles

on

  • 281 views

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

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

Statistics

Views

Total Views
281
Views on SlideShare
277
Embed Views
4

Actions

Likes
2
Downloads
0
Comments
0

1 Embed 4

https://twitter.com 4

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

"Awareness, Trust, and Software Tool Support in Distance Collaborations" by D. Redmiles "Awareness, Trust, and Software Tool Support in Distance Collaborations" by D. Redmiles Presentation Transcript

  • Awareness,  Trust,  and  So0ware   Tool  Support  in  Distance   Collabora8ons   David  Redmiles   Ins8tute  for  So0ware  Research  and   Department  of  Informa8cs   University  of  California,  Irvine   1  
  • 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  
  • 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  
  • 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   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  
  • Example:  Working  at  a  Distance   6  
  • 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  
  • 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  
  • 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  
  • 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  
  • 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  
  • Can  we  make  distance  ma`er  a   li`le  less?   •  Awareness   •  Trust   •  So0ware  Tool  Support   12  
  • 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  
  • 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  
  • 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  
  • Awareness   16  
  • 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  
  • 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  
  • 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]  
  • 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!
  • 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]  
  • 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]  
  • Ariadne 1.0 - Social and Technical Dependencies among Developers and Components
  • Progression of Graphs to Brackets (1) Developers Code
  • Progression of Graphs to Brackets (2) Developers Code
  • Ariadne  2.0 26
  • Grand Overview – many variables
  • Comparisons / Filter by Author
  • Comparisons / Filter by Artifact
  • 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  
  • 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  
  • 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  
  • 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  
  • Trust   35  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • Trust:  Compe8ng  Facets   43   team  diversity   8me   Trust   Threshold   -­‐   +   leadership  team  size   project  type   [Al-­‐Ani,  Redmiles  2009]  
  • Imagine  collabora8on  without  trust!     •  Double  checking.   •  Working  in  isola8on.   •  Reluctance  to  share  informa8on.   44   [Al-­‐Ani,  Redmiles  2009]  
  • 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]  
  • 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  
  • 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 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  …    
  • Degree  of  trust  “Game”  in  Protocol   49  
  • 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 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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.  
  • 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]  
  • 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  
  • •  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  
  • 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  
  • Tool  Support  Specific  to  Trust   61  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • Idea  Genera8on  Screen   [Schumann,  Shih,  Redmiles,  Horton,  2012]  
  • Idea  Evalua8on  Screen   [Schumann,  Shih,  Redmiles,  Horton,  2012]  
  • 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]  
  • 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]  
  • Lessons  Learned  –  Tool  Support  for   Trust   •  Evidence  that  informa8on  provided  by   tools  can  engender  trust.     •  Further  encouragement  towards  tool   support.   70  
  • 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  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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]  
  • 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  
  • 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  
  • THESEUS   87   A  very  busy  colleague!  
  • THESEUS   88   A  not  so  busy  colleague!  
  • 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  
  • 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.
  • 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.
  • 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  
  • Conclusions   93  
  • 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  
  • 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  
  • 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  
  • Providing  a  design  space  for  tools  
  • 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  
  • 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  
  • Finally   •  The  problems  and  facets  are     – Bigger  than  one  person,  one  approach,  etc.   •  Hope  others  will  join  the  pursuit.   100  
  • 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  
  • 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  
  • Hope  this  serves  to  inspire  you!  
  • Comments?  Ques8ons?   104   h`p://cradl.ics.uci.edu  
  • Addi8onal  Results   105  
  • 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  
  • 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  
  • Going  forward,  further   108  
  • 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  
  • 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