Network Correlates of Workgroup Climate

Strategy & Business Policy

Speaker: Brian Rubineau

Associate Professor - MC Gill, Toronto

Conference  Jouy-en-Josas T015

Abstract:

Google's Project Aristotle found that team-level climate factors such as psychological safety are more strongly predictive of team performance than any individual-level predictors. What about network-level predictors? The current study tests for network correlates of workgroup climate. I combine aggregated email data to construct workplace communication networks from a large financial company with HR data provided by the company along with two waves of a variety of workgroup-level climate measures taken from employee survey data. Using these data, I examine associations between workgroup-level climate scores and several types of workgroup-level network measures: structural measures such as density, clustering, and modularity; socio-demographic homophily such as gender and racial/ethnic category homophily; and employment homophily such as tenure and pay level homophily, among others. I test the directionality of these associations to distinguish between leading indicators of workgroup climate and lagging indicators of workgroup climate. Gender category homophily is significantly and consistently a leading negative indicator of workgroup climate. Workgroups where gender category similarity is strongly associated with communication tend to have worse workgroup climate outcomes. The network structure and employment homophily terms that show consistently significant associations do so as lagging indicators. These findings indicate directions for future research regarding both influencing and assessing workgroup climate.