Researching Trust in Globally Distributed Development Teams

For the past few years, Prof. David Redmiles’ research group has been studying the topic of trust in globally distributed teams.  The definition of trust they rely upon most is that trust, among collaborators, is the meeting of expectations.  In the context of software development, examples of collaborators meeting expectations may include simple behaviors such checking in modules on time or responding to emails within 24 hours.  Expectations can be more complex such as whether an individual or group tends to overcommit or to what degree collaborators are passive or assertive in problem solving discussions.  There can even be different expectations about body language, such as when it is appropriate to smile or not. 

Redmiles’ group came upon the examples above and the idea of trust as meeting expectations through interviews they conducted in five Fortune 500 companies.  They interviewed over 60 subjects over a three-year period.  Those interviewed had various positions such as software developers, engineers, managers, and lawyers.  There were both female and male participants in the interviews.  The interview protocol was designed so as to approach the topic of trust indirectly, using open-ended questions, and allowing the participants to put forward examples. 

The data Redmiles’ team gleaned from these interviews is rich.  Besides the examples about trust across different cultures, they learned about the roles of what one might call personal and professional trust or, more technically, affective and cognitive trust.  Affective trust refers to feelings and beliefs people have toward one another, beliefs that in turn affect how they set expectations for others’ behavior.  Cognitive trust also involves expectations for behavior but is based on beliefs about technical behaviors such as observed performance or knowledge about expertise.  Both play an important role in creating trust during collaborations. 

In fact as the notions of affective and cognitive trust emerged in their interviews, Redmiles’ group began to study them specifically.  They set up an experiment to see which might have more effect on collaborations.  They gave participants in the experiment multiple brainstorming tasks.  Participants were told they were working with two other participants they never saw.  The set-up mimicked situations where collaborators are put together for a project, but have never met and must develop trust swiftly to proceed.  For some of the tasks, participants saw information about the collaborators’ expertise, such as previous work experience, setting up expectations for cognitive trust.  For other tasks, they saw information about the collaborators’ personal traits including hobbies and various likes and dislikes.  As it turned out, there was no difference in the trust level between knowing cognitive and affective information. 

The observations Redmiles’ team are making from the interviews and the results they have gotten in more classic laboratory experiments, such as the brainstorming tasks, motivate their thinking about developing software tools to engender trust in collaboration.  Redmiles and his team are hypothesizing and evaluating specifically what kinds of information tools can be provided to engender trust. The work is ongoing.

To read more about cultural differences in general, a good starting point is the book by Hofstede et al., Cultures and Organizations: Software of the Mind – Intercultural Cooperation and Its Importance for Survival. 

In general, publications and other information from Redmiles’ research group is available to view at their Collaboration Research in Action, Design & Learning (CRADL) website. 

To read more about their observations of cultural differences in distributed development teams, see the paper on the website entitled Trust and Surprise in Distributed Teams: Towards an Understanding of Expectations and Adaptations.  To read more about their brainstorming experiment that discusses cognitive and affective trust, see the paper Supporting Initial Trust in Distributed Idea Generation and Evaluation.  Finally, to read more about software tools for supporting trust, see Foundations for the Design of Visualizations that Support Trust in Distributed Teams.

For more information, contact Prof. David Redmiles or visit the Cradl research group web site.