Effects of Cohesion-Based Feedback on the Collaborations in Global Software Development Teams

This paper describes a study that examines the effect of cohesion-based feedback on a team member’s behaviors in a global software development project. Chat messages and forum posts were collected from a software development project involving students living in the US and Mexico. Half of the teams in the project received feedback in the form of a graphical representation that displayed the group’s cohesion level, while the other teams received no feedback. The nature of the group interactions as well as the linguistic content of such interactions was then analyzed and compared. Results from this analysis show statistically significant differences between the feedback and non-feedback c onditions. More s pecifically, cohesion-based feedback had a positive relation to a team’s total message count, response rate, and individual cohesion score. In addition, the analysis of linguistic categories showed that the most salient categories observed were related to words about time and work. Furthermore, a comparison between feedback variables and type (i.e., positive and negative feedback) indicates that those individuals exposed to negative feedback had an increase in their communication pacing rates when exposed to positive feedback. Although the feedback system did not appear to affect individual performance, the findings s u ggest t h at t h e c o hesion m e asure d e fined in th is st ud y is positively correlated to the task cohesion construct and is also related to individual and team performance.