Controversy CornerBridging the gap between awareness and trust in globally distributed software teams
Introduction
Trust is one of the most well researched concepts in the literature on organizational teams (McAllister, 1995, Jarvenpaa and Leidner, 1999, Olson and Olson, 2000, Zolin et al., 2004, Wilson et al., 2006). There are strong motivations to understand trust. Teams with high levels of trust are associated with high productivity and team member satisfaction (Wilson et al., 2006). Teams can be successful without trust, but they tend to bear additional costs such as monitoring teammates and backing up their work (McAllister, 1995). These costs accumulate into productivity losses and low levels of commitment among team members (Wilson et al., 2006). Yet, as some researchers point out, trust is a complex concept involving both cognitive and emotional dimensions and some researchers find it useful to distinguish between cognitive trust and affective trust, as we explain later in this paper (McAllister, 1995, Rusman, 2011). In our previous work, in the context of distributed, collaborative work, we defined trust as the expectations that the trustee will perform a certain action important to the trustor in environments characterized by uncertainty (Al-Ani et al., 2013). At first glance, this definition is mechanistic and devoid of the role of affect, yet emotions are intertwined with the mechanics of trust. Events perceived as met or unmet expectations can trigger positive and negative emotions, as well as judgments in a trustor, just as in general, an individual's perception of events along with their previous experiences trigger a variety of affects (Krystal, 2015, Chapter 1).
Even though in face to face situations, individuals can and do make judgments about trust with little social interaction, in distributed settings individuals have much less and lower quality information about their teammates because of the reduced social cues present in online tools (Cramton, 2002, Cramton et al., 2007). They cannot easily observe their teammates’ efforts and hear what they say to others (Olson and Olson, 2000). When they attempt to explain their collaborators’ behavior, a process known as attribution (Ross, 1977, Harvey et al., 2014), they often do not have sufficient information to do so. When negative things happen, such as receiving late work from a teammate, they may be blind to contextual factors that offer a reasonable explanation, such as a holiday or emergency. These errors in attribution create misleading, usually negative, impressions of that team member that can lead to wasteful behaviors such as withholding information and backing up that team member's work (Al-Ani et al., 2013). The problems are reminiscent of those addressed by research on awareness (Dourish and Bellotti, 1992, Heath and Luff, 1992, Schmidt, 2002). Indeed, in earlier work, we found promise in using software support for awareness to help support trust (Trainer et al., 2011).
Our initial success in supporting trust with greater awareness motivated the extensive research and experimental study that we report on in this present paper. In light of the findings discussed above, we decided to investigate two research questions: i) whether greater awareness could lead to more accurate attribution and ii) in general, whether greater awareness could lead to more accurate perceptions of the trustworthiness of colleagues. By greater awareness, we mean providing collaborators with views of information related to the collaborators’ activities. Finally, with the possibility of more accurate judgments and perceptions, we provide collaborators with the opportunity to have more measured affective responses.
We proceeded by developing a tool, called Theseus, grounded in the literature on awareness, visualization, and trust, which would provide views of collaborators’ activities. We then conducted a controlled experiment with 40 participants: 28 computer science students and 12 professional software developers. We collected data on the kinds of attributions our participants made and, in general, their perceived trustworthiness of four remote teammates with different work situations. Our results indicate that participants exposed to visual summaries of the remote team member's situation tended to make more accurate attributions compared to relying on a judgment with no information available. In addition, participants’ perceived trustworthiness of the remote teammate more accurately reflected actual circumstances.
Our work makes three primary contributions:
- •
We fill a theoretical gap between the concepts of awareness and trust, extending the literature on software engineering awareness tools to include possible impacts on affect.
- •
We provide a proof of concept prototype tool to focus our research on how modern software development environments can in practice have positive impact on trust among distributed global software engineering team members.
- •
We provide a design space of information sources and trustworthiness factors, laying the groundwork for further exploration of software support in this arena.
The rest of the paper is structured as follows. In Section 2 we review three related concepts that are at the heart of our work: interpersonal trust, attribution, and perceived trustworthiness. We also present data from a large qualitative study we previously conducted that illustrate the interplay of these concepts in modern practice and the problems they create for globally distributed team members. We complete that section with a discussion of related background on the concept of awareness. In Section 3 we describe Theseus, our tool environment that provides awareness of information related to a collaborator's activities. In Section 4 we present the design of our experiment along with a description of the information shown by Theseus's visual awareness widgets in each condition. We report results of the experiment in Section 5. Section 6 follows with a discussion of the findings, their implications for theory and practice, and limitations. Finally, we draw conclusions from our work in Section 7.
Section snippets
Interpersonal trust
Interpersonal trust is a complex cognitive and emotional psychological state describing positive or negative feelings of a trustor toward a trustee (Rusman, 2011). The multi-dimensional nature of trust is one of its most researched aspects. Accordingly, some researchers define two categories of trust: cognitive and affective (McAllister, 1995, Rusman, 2011). Cognitive trust refers to the capability of the trustee, as determined by their knowledge, skills, and competencies. Definitions of
Filling the gap between awareness and trust with the Theseus prototype
In response to the issues and challenges discussed above, we developed a software engineering awareness environment called Theseus (Trainer, 2012). Theseus collects information about software development activities from the project's source-code repository and issue tracker. Theseus also generates information that could be, but is not necessarily exposed by, project management tools (e.g., calendars and e-mail records). The information is then aggregated and displayed in the interface using a
Present study
The present study evaluates whether providing visual summaries of availability and responsiveness information via Theseus helps globally distributed software developers make more accurate attributions and, in general, better trustworthiness judgments about remote team members. As we explained previously, providing important contextual information can be helpful in exploring someone's situation. This informs the attribution process, reducing bias in a coordination setting. As such, we
Results
We begin this section by presenting results from a statistical analysis of our entire sample of participants. To provide additional support for our conclusions and enhance the overall validity of our study, we present an additional analysis in Section 5.2 treating students and professional participants as separate sub samples.
Discussion
In the present study we set out to understand whether providing information one might not normally have about a remote team member can yield more accurate attributions, and in general, more accurate perceptions of the trustworthiness of that team member. Specifically, we designed an experiment where participants saw views of four different developers’ availability and responsiveness. We measured the kinds of attributions participants made for each team member's failure to complete their work on
Conclusion
A great deal of research has been done around the concept of trust. However, in practice, distributed, collaborative teams still face challenges. Individuals make errors of attribution about situations involving their remote collaborators and, in general, fail to develop a sense of trust with their teammates. These negative attributions and perceptions can lead to negative emotions and, in general, be disruptive to teamwork and collaborative projects. However, these challenges resemble those
Acknowledgments
This research was supported by the U.S. National Science Foundation under grants 0943262, 0808783 and 1111446. Data collection for this study was performed while the first author was a doctoral student at The University of California, Irvine. We thank our colleagues Drs. Oliver Wang and Matthew Bietz for their feedback on experimental design, as well as Dr. Ban Al-Ani who spearheaded the previous qualitative work that inspired and motivated the present study. We also gratefully acknowledge our
Erik H. Trainer is a Visiting Assistant Professor in the School of Information at the Pratt Institute. He received his PhD in Information Computer Science from the University of California, Irvine in 2012. Before joining the Pratt Institute, he was a postdoctoral researcher at the Institute for Software Research at Carnegie Mellon University. His recent research focuses on creating technologies and practices that support the relationships of people engaged in technical work, especially in the
References (77)
- et al.
A survey of trust in computer science and the semantic web. Web Semantics: Science
Services and Agents on the World Wide Web
(2007) - et al.
Trust, reciprocity, and social history
Games Econ. Behav.
(1995) - et al.
Gender pairing bias in trustworthiness
J. Socio-Econ.
(2009) The intuitive psychologist and his shortcomings: distortions in the attribution process
Adv. Exp. Social Psychol.
(1977)- et al.
All in due time: the development of trust in computer-mediated and face-to-face teams
Org. Behav. Hum. Decis. Process.
(2006) - et al.
Interpersonal trust in cross-functional, geographically distributed work: a longitudinal study
Inf. Org.
(2004) - et al.
In strangers we trust? Findings of an empirical study of distributed development
- et al.
Trust and surprise in distributed teams: towards an understanding of expectations and adaptations
- et al.
Distributed developers and the non-use of Web 2.0 technologies: a proclivity model
- et al.
Globally distributed system developers: their trust expectations and processes
Design of Comparative Experiments
The Logic and Limits of Trust
Effects of four computer-mediated communications channels on trust development
Improving the performance of low self-esteem individuals: an attributional approach
Acad. Manag. J.
Communication factors and trust: an exploratory study
Psychol. Rep.
The role of social media in affective trust building in customer-supplier relationships
Electron. Comm. Res.
SocialCDE: a social awareness tool for global software teams
Can social awareness foster trust building in global software teams?
The mutual knowledge problem and its consequences for dispersed collaboration
Org. Sci.
Attribution in distributed work groups
Situation invisibility and attribution in distributed collaborations
J. Manag.
Coefficient alpha and the internal structure of tests
Psychometrika
Awareness and coordination in shared workspaces
The awareness network, to whom should I display my actions? And, whose actions should I monitor
IEEE Trans. Software Eng.
Socially translucent systems: social proxies, persistent conversation and the design of "Babble
Beyond attribution theory: cognitive processes in performance appraisal
J. Applied Psychol.
‘Can I trust you to trust me? A theory of trust, monitoring, and cooperation in interpersonal and intergroup relationships’
Group Org. Manag.
Stanford guidelines for web credibility. a research summary from the stanford persuasive technology lab
Dynamic trust processes: trust dissolution and restoration
23rd IACM
Virtual team effectiveness: a proposed research agenda
Inf. Syst. J.
Group awareness in distributed software development
Impression formation in computer-mediated environments revisited: an analysis of the breadth and intensity of impressions
Commun. Res.
Attribution theory in the organizational sciences: the road traveled and the path ahead
Acad. Manag. Perspect.
Collaboration and control: crisis management and multimedia technology in London underground control rooms
J. Comput. Supp. Cooperative Work
Understanding conflict in geographically distributed teams: the moderating effects of shared identity, shared context, and spontaneous communication
Org. Sci.
Situated coworker familiarity: how site visits transform relationships among distributed workers
Org. Sci.
Developing trust in virtual teams
Homophily and differential returns: sex differences in network structure and access in an advertising firm
Administrative Sci. Q.
Cited by (0)
Erik H. Trainer is a Visiting Assistant Professor in the School of Information at the Pratt Institute. He received his PhD in Information Computer Science from the University of California, Irvine in 2012. Before joining the Pratt Institute, he was a postdoctoral researcher at the Institute for Software Research at Carnegie Mellon University. His recent research focuses on creating technologies and practices that support the relationships of people engaged in technical work, especially in the areas of open-source software development and educational technology.
David F. Redmiles is a Professor in the Department of Informatics at the University of California, Irvine (UCI) in the Donald Bren School of Information and Computer Sciences (ICS). He earned his PhD in Computer Science in 1992 from the University of Colorado, Boulder. He is the author of nearly 200 publications integrating the areas of software engineering, human-computer interaction, and computer-supported cooperative work. His current research focuses on distributed and collaborative software engineering, especially the aspects of awareness and trust among collaborators. He is a Fellow of Automated Software Engineering since 2009 and an ACM Distinguished Scientist since 2011.