Using sentiment analysis to measure the effects of leaders in global software development

Abstract

This study examines the effect of a leader's communication activities on global software development learning teams. Data collected from students' chat histories and forum discussions from two global software development projects were processed and compared. Several quantitative analysis methods were used to determine the differences between leader and leaderless teams. In addition, Linguistic Inquiry and Word Count (LIWC) measures were used to investigate the linguistic features of distributed software teams. The results show that teams with leaders had more communication, performed better, and used different word categories than teams with no leaders. These findings are discussed and their implications for research are described in the paper.